Update: Elon Musk, SpaceX’s CEO and chief designer, has posted an update on the status of the upcoming Falcon 9 launch on his Twitter account. “Will do another static fire of rocket to make sure all is good & AF [[Air Force]] needs to test ICBMs, so probable launch Sept 29/30,” Musk tweeted.
“The static fire is scheduled for later this week, perhaps Wednesday, sources said. It will retest the Falcon 9 rocket after several problems cropped up during a hotfire of the launcher’s engines Thursday at Vandenberg Air Force Base, Calif.
“The U.S. Air Force Western Range, which controls a network of tracking and communications assets based at Vandenberg, is busy for the next few weeks due to Minuteman ballistic missile testing.”
The Falcon 9 rocket, with CASSIOPE inside its fairing, on the way to the launch pad at Vandenberg Air Force Base. (Photo credit: SpaceX).
A GPS instrument designed by University of New Brunswick scientists is scheduled to be launched into space aboard the SpaceX Falcon 9 rocket on September 15. The rocket will depart Vandenberg Air Force base in California as part of the CASSIOPE (Cascade Smallsat and Ionospheric Polar Explorer) mission.
Dr. Richard Langley, GPS World Innovation editor and professor in geodesy and geomatics engineering at the University of New Brunswick, is a principal investigator behind the scientific portion of the CASSIOPE mission. Langley and his colleagues will monitor data from the GPS instrument, which is part of the Enhanced Polar Outflow Probe (e-POP) payload aboard the spacecraft.
E-POP will continue the sequence of Canada’s orbiting space environment sensors, which began with Canada’s first satellite, Alouette 1, launched in 1962 to study the ionosphere. e-POP is, perhaps, the most extensive suite of sensors for studying the ionosphere/magnetosphere/thermosphere yet to be launched, and will provide Canadian and other scientists with the opportunity to better understand the impact and variability the sun has on the space environment — what we call “space weather.”
The research satellite CASSIOPE on a test platform at the Canadian Space Agency’s David Florida Laboratory. CASSIOPE hosts the GPS Attitude, Positioning, and Profiling instrument designed by GGE researchers. The four white antennas on the left-facing side of the spacecraft will be used to determine the position, velocity, and attitude of the spacecraft while the antenna on the upper side will be used to profile the ionosphere’s electron density. (Photograph courtesy of MacDonald, Dettwiler and Associates Ltd.)
By Richard Langley, Steffen Thoelert, and Michael Meurer
The spectrum of signals from IRNSS-1A, the first satellite in the Indian Regional Navigation Satellite System, as recorded by German Aerospace Center researchers in late July, appears to be consistent with a combination of BPSK(1) and BOC(5,2) modulation.
Figure 1 shows that, centered at 1176.45 MHz, the signal has a single symmetrical main lobe and a number of side lobes characteristic of the signal structure that the Indian Space Research Organization (ISRO) announced would be used for IRNSS transmissions in the L-band. Figure 2 shows the corresponding IQ constellation diagram. Further analysis will be required to sleuth additional signal details as ISRO, so far, has not publicly released an IRNSS interface control document describing the signal structure in detail.
Figure 1. Spectrum of IRNSS-1A L5 signal.Figure 2. IQ constellation diagram of IRNSS-1A L5 signal.
The German scientists caution that “this is a very early snapshot of the current signal transmission and probably both the signal power and the signal quality will change and possibly improve during the in-orbit-testing phase of the satellite’s operation.
Extra Life for IIRs, IIR-Ms
U.S. Air Force engineers are testing on-orbit a technique to extend the life of the 19 GPS IIR and IIR-M satellites on orbit, roughly 60 percent of the current contellation.
A new charging method may reduce the rate of satellite battery degradation, thereby extending satellite operational life. If the technique passes the test, the initiative could add a combined 20 years to the life of the satellites — saving the Air Force tens of millions of dollars in the process.
Gen. William Shelton, commander of Air Force Space Command, credits Capt. Jacob Hempen of the Air Force’s 2nd Space Operations Squadron for the job. Capt. Hempen says in turn that Warren Hwang of the Aerospace Corporation originated the idea.
When satellite solar panels are directly exposed to the Sun, they charge satellite batteries while continuing to power other operations onboard the space vehicle. When the satellite passes into the Sun’s shadow behind the Earth, it runs on batteries. The batteries can be re- charged at variable rates. When some of the batteries are powered above a certain rate threshold, they can overheat, accelerating their natural rate of decay.
Lowering battery charging rates could still enable the satellites to perform well while minimizing the rate of degradation. Hitting the optimum number called for some finely-honed calculations.
The satellites were built by Lockheed Martin Space Systems, and the oldest still in operation was launched in 1997.
They had an intial design life of eight years, which many have now well outlasted. If the technique proves out and is carefully applied across the board, it could conceivably fill in replenishment gaps equivalent more than two additional spacecraft — conceivably as much hundreds of millions of dollars in build and launch costs, postponed. In today’s budget environment, a postponement can be construed as equivalent to outright savings.
System Briefs
GLONASS Partial Make-Good. Russia will launch two GLONASS satellites later this year to make up for the loss of three satellites in the July 2 Proton rocket explosion. The first is scheduled for the beginning of September, and the second at the end of October. Both will rise aboard Soyuz carrier rockets, which have proven more reliable than the Protons. A constellation of 29 GLONASS satellites is now in orbit, with 24 spacecraft in operation, three spares, one in maintenance, and one in test flight phase.
Meanwhile, plans to reduce GLONASS funding have alarmed at least some deputies of the Duma, the Russian state legislative body. Government officials have floated a plan to reduce funding of the space program in 2014 by 11.7 billion rubles ($355 million), by 13.5 billion rubles in 2015, and by 40 billion rubles in 2016. The federal space program of Russia for 2006-2015 already lacks 10.5 billion rubles funding, and this year there has been a 2.3-billion-ruble additional reduction in R&D. A Duma committee chairperson warned that this trend will “lead to the loss of confidence of the international community in the GLONASS system and, consequently, to a reduction in its use globally. Russia will lose a strategic global instrument of political and economic prestige.” The Duma has recommended that the government maintain funding of federal space programs.
Galileo Satellites’ Trial By Noise. The first Galileo Full Operational Capability (FOC) satellite successfully completed acoustic testing in July, part of a full-scale test campaign at ESA’s ESTEC Test Centre in Noordwijk, the Netherlands.
The satellite was placed in the Large European Acoustic Facility (LEAF), effectively the largest sound system in Europe. A quartet of noise horns embedded in a wall of the 11 x 9 x 16.4 meter test chamber generated an acoustic noise level of 140.7 decibels, about the same noise as standing 25 meters from a jet taking off, and intended to simulate the extreme environment experienced by a satellite atop a rocket about to fire itself off the launch pad.
A second FOC satellite arrived at ESTEC on 9 August from manufacturer OHB in Bremen, Germany. It will undergo a similar acoustic testing and then a System Compatibility Test Campaign will linking it with the Galileo Control Centres in Germany and Italy and ground user receivers as if it were already in orbit.
A total of 14 FOC satellites are being produced and then tested at ESTEC as an integral part of their path to orbit. A second work order of eight satellites has been given to OHB.
GPS III Pathfinder. On July 19, Lockheed Martin delivered a full-sized, functional prototype of the next-generation GPS satellite to Cape Canaveral Air Force Station to test facilities and pre-launch processes in advance of the arrival of the first GPS III flight satellite.
The GPS III Non-Flight Satellite Testbed (GNST) paves the way for the first flight GPS III satellite, expected to arrive at the Cape in 2014, ready for launch by in 2015.
An innovative investment by the Air Force under the original GPS III development contract, the GNST has helped to identify and resolve development issues prior to integration and test of the first GPS III flight space vehicle (SV-01).
Following the Air Force’s rigorous “back-to-basics” acquisition approach, the GNST has gone through the development, test and production process for the GPS III program first, significantly reducing risk for the flight vehicles, improving production predictability, increasing mission assurance and lowering overall program costs.
Lockheed Martin is currently under contract for production of the first four GPS III satellites (SV 01–04), and has received advanced procurement funding for long-lead components for the fifth, sixth, seventh and eighth satellites (SV 05–08).
GNSS Industry Survey. Here are the results of two questions asked about government and industry from the 2013 GNSS STATE OF THE INDUSTRY SURVEY.
Is government committed to private industry in a time of drastic budget cuts?Is industry actively making its concerns known to government?
By Steffen Thoelert, Johann Furthner, and Michael Meurer
Future positioning and navigation applications of modernizing and newly established GNSSs will require a higher degree of signal accuracy and precision. Thus, rigorous and detailed analysis of the signal quality of recently launched satellites, including the discovery of any possible imperfections in their performance, will have important implications for future users.
Global navigation satellite systems achieved amazing progress in 2012, with major milestones reached by the various navigation and augmentation systems, bringing new satellites and satellite generations into orbit. Since the complexity of the satellites and also the requirements for a precise and robust navigation increase consistently, all of the newly available signals of the existing or emerging navigation satellite systems must be analyzed in detail to characterize their performance and imperfections, as well as to predict possible consequences for user receivers.
Since the signals are well below the noise floor, we use a specifically developed GNSS monitoring facility to characterize the signals. The core element of this monitoring facility is a 30-meter high-gain antenna at the German Aerospace Center (DLR) in Weilheim that raises GNSS signals well above the noise floor, permitting detailed analysis. In the course of this analysis, we found differences in the signal quality in the various generations of the Chinese navigation satellite system BeiDou, differences which influence the navigation performance.
This article gives an overview of new navigation satellites in orbit. For selected satellites, a first signal analysis reveals important characteristics of these signals. The data acquisition of these space vehicles was performed shortly after the start of their signal transmission to get a first hint about the quality and behavior of the satellites.
For more detailed analysis, these measurements should be repeated after the satellites become operational. Then the acquired high-gain antenna raw data in combination with a precise calibration could be used for a wider range of analyses: signal power, spectra, constellation diagrams, sample analysis, correlation functions, and codes to detect anomalies and assess the signal quality and consequently the impact at the user performance.
Measurement Facility
In the early 1970s, DLR built a 30-meter dish (Figure 1) for the HELIOS-A/B satellite mission at the DLR site Weilheim. These satellite missions were the first U.S./German interplanetary project. The two German-built space probes, HELIOS 1 (December 1974–March 1986) and HELIOS 2 (January 1976–January 1981), approached the Sun closer than the planet Mercury and closer than any space probe ever. Later, the antenna supported space missions Giotto, AMPTE, Equator-S, and other scientific experiments.
Figure 1. 30-meter high-gain antenna.
In 2005, the Institute of Communications and Navigation of the DLR established an independent monitoring station for analysis of GNSS signals. The 30-meter antenna was adapted with a newly developed broadband circular polarized feed. During preparation for the GIOVE-B in-orbit validation campaign in 2008, a new receiving chain including a new calibration system was installed at the antenna. Based on successful campaigns and new satellite of modernizing GPS and GLONASS, and GNSSs under construction — Galileo and COMPASS — the facility was renewed and updated again in 2011/2012.
This renewal included not only an upgrade of the measurement system itself, but also refurbishment of parts of the high-gain antenna were refurbished.
The antenna is a shaped Cassegrain system with an elevation over azimuth mount. The antenna has a parabolic reflector of 30 meters in diameter and a hyperbolic sub-reflector with a diameter of 4 meters. A significant benefit of this antenna is the direct access to the feed, which is located within an adjacent cabin (Figure 2). The L-band gain of this high-gain antenna is around 50 dB, the beam width is less than 0.5°. The position accuracy in azimuth and elevation direction is 0.001°. The maximum rotational speed of the whole antenna is 1.5°/second in azimuth and 1.0°/second in elevation direction.
Figure 2. The shaped Cassegrain system: (1) parabolic reflector of 30 m diameter; (2) hyperbolic sub- reflector with a diameter of 4 meter; (3) sub-reflector; (4) Cabin with feeder and measurement equipment.
Measurement Set-up
The antenna offers another significant advantage in the possibility to have very short electrical and high-frequency connection between the L-band feeder and the measurement equipment. As mentioned earlier, the challenge for future GNSS applications is the high accuracy of the navigation solution. Therefore, it is necessary to measure and then analyze the signals very accurately and precisely. To achieve an uncertainty of less than 1 dB for the measurement results required a complete redesign of the setup, which consists of two main parts:
paths for signal receiving and acquiring the measurement data;
calibration elements for different calibration issues.
The path for receiving the signal and acquiring the measurement data consists of two signal chains, each equipped with two low-noise amplifiers (LNAs) with a total gain of around 70 dB, a set of filters for the individual GNSS navigation frequency bands, and isolators to suppress reflections in the measurement system. With this setup it is possible to measure right-hand circular polarized (RHCP) and left-hand circular polarized (LHCP) signals in parallel.
This provides the capability to perform axial ratio analysis of the satellite signal, and consequently an assessment of the antenna of the satellite. Using the switches SP01 and SP02, the measurement system is also able to acquire data from two different bands at the same time. This enabless investigations concerning the coherence between the signals in post-processing.
The signals are measured and recorded using two real-time vector signal analyzers with up to 120 MHz signal bandwidth. Both analyzers are connected to a computer capable of post-processing and storing the data. Additional equipment like digitizers or receivers can be connected to the system using the splitter III outputs, where the unfiltered RHCP signals are coupled out after the first LNA. A high-performance rubidium clock is used as reference signal for the whole measurement equipment. In front of the first LNA of each chain, a signal can be coupled in for calibration issues.
Control Software. Due to the distance of the antenna location from the Institute at Oberpfaffenhofen (around 40 kilometers) it was necessary to perform all measurement and calibration procedures during a measurement campaign via remote control. A software tool was developed which can control any component of the setup remotely. In addition, this software can perform a complete autonomous operation of the whole system by a free pre-definable sequence over any period of time. This includes, for example, the selection of the different band-pass filters, the polarization output of the feed, and the control of the calibration routines.
After the measurement sequence, the system automatically copies all data via LAN onto the processing facility, starts basic analysis based on spectral data, and generates a report. Sophisticated analysis based on IQ raw data is performed manually at this time.
Absolute Calibration
To fulfill the challenge of highly accurate measurements, it is necessary to completely characterize all elements of the measurement system, which comprises the antenna itself and the measurement system within the cabin after the feed. An absolutely necessary precondition of the calibration of the high-gain antenna is a very accurate pointing capability. The pointing error should be less than 0.01° concerning antennas of this diameter.
Furthermore, it is important to check long-term stability of these characterizations and the influences of different interference types and other possible error sources. This has to be taken in to account, when it comes to a point where the value of the absolute calibration has the same range as the summed measurement uncertainties of the equipment in use.
Antenna Calibration. High-accuracy measurements require not only the correct antenna alignment but also accurate power calibration of the antenna. To determine the antenna gain, well known reference sources are needed. These could be natural sources like radio stars or artificial sources like geostationary satellites.
Standard reference signal sources for the calibration of high-gain antennas are the radio sources Cassiopeia A, Cygnus, and Taurus. All these radio sources are circumpolar relative to our ground station, and therefore usable for calibrations at all times of the year. A further advantage of these calibration sources is the wide frequency range of the emitted signals. Thus, contrary to other signal sources (like ARTEMIS satellite L band pilot signal) the antenna gain can be calibrated in a wide bandwidth. With the help of the well-known flux density of the celestial radio sources and using the Y-method, the relation between the gain of the antenna and the noise temperature of the receiving system, or G/T, can be measured. Measuring the noise figure of the receiving system, the antenna gain can finally be calculated.
System Calibration. The measurement system calibration behind the feed is performed using wideband chirp signals. The chirp is injected into the signal chains via coupler I and II (Figure 3). The calibration signal is captured by the two vector signal analyzers. In the next step, the signal is linked via the switches directly to the analyzers, and the chirp signals are recorded as reference again. It has to be taken into account that more elements are in the loop during the chirp recordings compared to the receiving chain. These are the link between the signal generator and the couplers and the direct path to the analyzers.
Figure 3. Measurement setup overview.
To separate the receiving chain from the additional elements within the wideband calibration loop, two more measurements are needed. The injection path from the signal generator to the couplers and the direct paths are characterized by network analyzer (NWA) measurements. Based on the chirp and NWA measurements, the transfer function of the system is calculated to derive the gain and phase information. To determine the calibration curve over the frequency range from 1.0 GHz to 1.8 GHz, a set of overlaying chirps with different center frequencies is injected into the signal paths and combined within the analysis. Figure 4 and Figure 5 show the results of the wideband calibration of gain and phase.
Figure 4. Gain of the measurement system after the feed over 14 hours.Figure 5. Phase of measurement system.
Is it enough to determine the gain only once? If we assume that there is no aging effect of the elements, and the ambient conditions like temperature are constant, the gain should not change. In reality the behavior of the system is not constant. Figure 6 shows the temperature within the cabin during a failure of its air conditioning system. Figure 7 shows the corresponding gain of the measurement system during the temperature change in the cabin of about 5° Celsius. Clearly, it can be seen that the gain changed around 0.2 dB.
Figure 6. Cabin temperature increase during outage of the air condition concerning measurements shown in Figure 7.Figure 7. Gain variations of the measurement system based on temperature variations in the cabin (see Figure 6).
This example shows the sensitivity of the system to changes in environmental conditions. Usually the measurement system is temperature-stabilized and controlled, and the system will not change during data acquisition. But every control system can be broken, or an element changes its behavior. For this reason, the calibration is performed at least at the beginning and at the end of a satellite path (maximum 8 hours).
Measurement Results
Here we present selected results from the European Galileo and the Chinese BeiDou navigation systems.
Galileo FM3 and FM4. In October 2012, the third and fourth operational Galileo satellites, FM3 and FM4, were launched into orbit. Signal transmissions started in November and in December, respectively. Both satellites provide fully operational signals on all three frequency bands, E1, E5, and E6. The measurement data of both satellites were captured in December 2012, shortly after the beginning of the signal transmission. Figure 8 shows the spectra of both satellites for El, E5, and E6 bands. The quality of the transmitted signals seems to be good, but for the El signal of FM4 satellite, minor deformations of the spectra are visible.
Figure 8. Measurement results of Galileo IOV FM3 & FM4: El, E5 and E6 spectra.
Figure 9 shows the results of the IQ constellations both for FM3 and FM4 concerning each transmitted signal band. The constellations and consequently the modulation quality of each signal are nearly perfect for the FM3 satellite. The IQ constellation diagrams of FM4 show minor deformations in each band. What impact these imperfections create for future users has yet to be analyzed. Both satellites were at the time of measurement campaign still in the in-orbit test phase and did not transmit the final CBOC signal in the E1 band. It could be expected that especially the signals of the FM4 will be adjusted to become more perfect.
Figure 9 Measurement results of Galileo IOV FM3 & FM4: E1, E5, and E6 – IQ Constellation.
BeiDou M6. BeiDou satellites transmit navigation signals in three different frequency bands, all are located adjacent to or even inside currently employed GPS or Galileo frequency bands. The center frequencies are for the B1 band 1561.1 MHz, B3 band 1268.52 MHz, and B2 band 1207.14 MHz.
In 2012, China launched six satellites: two inclined geostationary space vehicles and four medium-Earth orbit ones, concluding in September (M5 and M6) and October 2012 (IGSO6). There have been further BeiDou launches in 2013, but these satellites’ signals are not analyzed here.
Figure 10 displays calibrated measurement results from the Beidou M6 satellite. The spectra of the B2 and B3 band of the Beidou M6 satellite are clean and show no major deformation. Within the B1 spectra, some spurious results, especially on top of the side lobes, are obvious. This behavior has to be investigated more in detail to determine their origin. The IQ diagrams, which visualize the modulation quality, show also no major deformation. Only within the B3 signal, a marginal compression of the constellation points can be seen, which points to a large-signal operation at the beginning of the saturation of the amplifier of the satellite.
Figure 10. BeiDou M6 satellite signal spectra and IQ constellations at B1, B2 and B3 band
Conclusion
Reviewing the quality of the presented measurements, signal analysis, and verification on GNSS satellites, the use of the 30-meter high-gain antenna offers excellent possibilities and results. Regarding the calibration measurements of the antenna gain and measurement system, the variances are in the range of measurement uncertainty of the equipment.
The sensitivity of the measurement system concerning ambient conditions was exemplarily shown based on the gain drift caused by a temperature drift. But the solution is simple: stabilize the ambient conditions or perform calibration in a short regular cycle to detect changes within the system behavior to be able to correct them.
Based on this absolute calibration, a first impression of the signal quality of Galileo FM3 and FM4 and the BeiDou M6 satellites were presented using spectral plots and IQ diagrams. Only minor distortion could be detected within the Galileo FM4 and Beidou M6 signal; these distortions may be negligible for most users. Concerning FM4 and FM3, both satellites were in the in-orbit test phase during the data acquisition. The signal quality may have been changed during their stabilization process in orbit, or the signals have been adjusted in the meantime. Thus, it would be interesting and worthwhile to repeat the measurements and perform detailed analysis to assess the final satellite quality and consequently the user performance.
Acknowledgments
The authors wish to thank the German Space Operation Centre for the opportunity to use the high-gain antenna. The support of colleagues at the DLR ground station Weilheim for the operational and maintenance service over recent years is highly appreciated. This work was partly performed within the project “Galileo SEIOT (50 NA 1005)” of the German Space Agency, funded by the Federal Ministry of Economics and Technology and based on a resolution by the German Bundestag. Finally, the support of DLR’s Centre of Excellence for Satellite Navigation is highly appreciated.
This article is based on the paper “GNSS Survey – Signal Quality Assessment of the Latest GNSS Satellites” presented at The Institute of Navigation International Technical Meeting 2013, held in San Diego, California, January 28–30, 2013.
Steffen Thoelert received his diploma degree in electrical engineering at the University of Magdeburg. He works in the Department of Navigation at German Aerospace Centre (DLR), on signal quality assessment, calibration, and automation of technical processes.
Johann Furthner received his Ph.D. in laser physics at the University of Regensburg. He works in the DLR Institute of Communication and Navigation on the development of navigation systems in a number of areas (systems simulation, timing aspects, GNSS analysis, signal verification, calibration processes).
Michael Meurer received a Ph.D. in electrical engineering from the University of Kaiserslautern, where he is now an associate professor, as well as director of the Department of Navigation at DLR.
An alternative tool for detecting underground nuclear explosions?
By Dorota A. Grejner-Brzezinska, Jihye Park, Joseph Helmboldt, Ralph R. B. von Frese, Thomas Wilson, and Jade Morton
Well-concealed underground nuclear explosions may go undetected by International Monitoring System sensors. An independent technique of detection and verification may be offered by GPS-based analysis of local traveling ionospheric disturbances excited by an explosion. Most of the work to date has been at the research demonstration stage; however, operational capability is possible, based on the worldwide GPS network of permanently tracking receivers. This article discusses a case study of detecting underground nuclear explosions using observations from GPS tracking stations and the Very Large Array radio telescope in New Mexico.
More than 2,000 nuclear tests were carried out between 1945 and 1996, when the Comprehensive Nuclear Test Ban Treaty was adopted by the United Nations General Assembly. Signatory countries and the number of tests conducted by each country are the United States (1000+), the Soviet Union (700+), France (200+), the United Kingdom, and China (45 each). Three countries have broken the de facto moratorium and tested nuclear weapons since 1996: India and Pakistan in 1998 (two tests each), and the Democratic People’s Republic of Korea (DPRK) in 2006 and 2009, and most recently, in 2013.
To date, 183 countries have signed the treaty. Of those, 159 countries have also ratified the treaty, including three nuclear weapon states: France, the Russian Federation, and the United Kingdom. However, before the treaty can enter into force, 44 specific nuclear-technology-holder countries must sign and ratify. Of these, India, North Korea and Pakistan have yet to sign the CTBT, and China, Egypt, Iran, Israel, and the United States have not ratified it.
The treaty has a unique and comprehensive verification regime to make sure that no nuclear explosion goes undetected. The primary components of the regime are:
The International Monitoring System: The IMS includes 337 facilities (85 percent completed to date) worldwide to monitor for signs of any nuclear explosions.
International Data Center: The IDC processes and analyzes data registered at IMS stations and produces data bulletins.
Global Communications Infrastructure: This transmits IMS data to the IDC, and transmits data bulletins and raw IMS data from IDC to member states.
Consultation and Clarification: If a member state feels that data collected imply a nuclear explosion, this process can be undertaken to resolve and clarify the matter.
On-Site Inspection: OSI is regarded as the final verification measure under the treaty.
Confidence-Building Measures: These are voluntary actions. For example, a member state will notifying CTBTO when there will be large detonations, such as a chemical explosion or a mining blast.
The IMS (see Figure 1) uses the following state-of-the-art technologies. Numbers given reflect the target configuration:
Seismic: Fifty primary and 120 auxiliary seismic stations monitor shockwaves in the Earth. The vast majority of these shockwaves — many thousands every year — are caused by earthquakes. But man-made explosions such as mine explosions or the North Korean nuclear tests in 2006, 2009, and 2013 are also detected.
Hydroacoustic: As sound waves from explosions can travel extremely far underwater, 11 hydroacoustic stations “listen” for sound waves in the Earth oceans.
Infrasound: Sixty stations on the surface of the Earth can detect ultra-low-frequency sound waves that are inaudible to the human ear, which are released by large explosions.
Radionuclide: Eighty stations measure the atmosphere for radioactive particles; 40 of them can also detect the presence of noble gas.
Figure 1. The International Monitoring System (IMS): worldwide facilities grouped by detection technologies used.
Only the radionuclide measurements can give an unquestionable indication as to whether an explosion detected by the other methods was actually nuclear or not. The observing stations are supported by 16 radionuclide laboratories.
Since radionuclide detection method provides the ultimate verification as far as the type of blast goes, it should be mentioned that while the 2006 North Korean event (yield of less than a kiloton) was detected by the IMS stations in more than 20 different sites within two hours of detonation, and both seismic signal and radioactive material were detected, the 2009 event (yield of a few kilotons) was detected by 61 IMS stations; seismic and infrasound signals were detected, but no radioactive material was picked up by the radionuclide stations. Seismic signal was consistent with a nuclear test, but there was no “ultimate” proof by the radionuclide method.
Thus, well-concealed underground nuclear explosions (UNEs) may be undetected by some of the IMS sensors (such as the radionuclide network). This raises a question: Is there any other technology that is readily available that can detect and discriminate various types of blasts, particularly those of nuclear type? Recent experiments have shown that an independent technique of detection and verification may be offered by GPS-based analysis of local traveling ionospheric disturbances (TIDs) excited by an explosion.
GNSS-Based Detection
Atmospheric effects from mostly atmospheric nuclear explosions have been studied since the 1960s.The ionospheric delay in GNSS signals observed by the ground stations can be processed into total electron content (TEC), which is the total number of electrons along the GNSS signal’s path between the satellite and the receiver on the ground. The TEC derived from the slant signal path, referred to as the slant TEC (STEC), can be observed and analyzed to identify disturbances associated with the underground nuclear explosion.
STEC signature (in spectral and/or spatial-temporal domains) can be analyzed to detect local traveling ionospheric disturbances (TID).
TID can be excited by acoustic gravity waves from a point source, such as surface or underground explosions, geomagnetic storms, tsunamis, and tropical storms. TIDs can be classified as Large-Scale TID (LSTID) and Medium-Scale TID (MSTID) based on their periods regardless of the generation mechanism. The periods of LSTIDs generally range between 30–60 minutes to several hours, and those of MSTIDs range from 10 to 40 or even 60 minutes. LSTIDs mostly occur from geophysical events, such as geomagnetic storms, which can be indicated by global Kp indices, while MSTIDs are genrally not related to any high score Kp indices. An underground nuclear explosion can result in an MSTID.
TIDs are generated either by internal gravity wave (IGW) or by acoustic gravity wave (AGW). The collisional interaction between the neutral and charged components cause ionospheric responses. The experimental results indicate IGWs can change the ozone concentration in the atmosphere. In the ionosphere, the motion of the neutral gas in the AGW sets the ionospheric plasma into motion.
The AGW changes the iso-ionic contours, resulting in a traveling ionospheric disturbance.
The past 10–15 years has resulted in a significant body of research, and eventually a practical application, with worldwide coverage, of GPS-based ionosphere monitoring. A significant number of International GNSS Service (IGS) permanent GNSS tracking stations (see Figure 2) form a powerful scientific tool capable of near real-time monitoring and detection of various ionospheric anomalies, such as those originating from the underground nuclear explosions (UNEs).
Figure 2. The IGS global tracking network of 439 stations.
The network is capable of continuously monitoring global ionospheric behavior based on ionospheric delays in the GNSS signals. The GNSS signals are readily accessible anywhere on Earth at a temporal resolution ranging from about 30 seconds up to less than 1 second.
A powerful means to isolate and relate disturbances observed in TEC measurements from different receiver-satellite paths is to analyze the spectral coherence of the disturbances. However, in our algorithms, we emphasize the spatial and temporal relationship among the TEC observations. Spatial and temporal fluctuations in TEC are indicative of the dynamics of the ionosphere, and thus help in mapping TIDs excited by acoustic-gravity waves from point sources, as well as by geomagnetic storms, tropical storms, earthquakes, tsunamis, volcanic explosions, and other effects.
Methodology of UNE Detection
Figure 3 illustrates the concept of the generation of the acoustic gravity wave by a UNE event, and its propagation through the ionosphere that results in a traveling ionospheric disturbance (TID). The primary points of our approach are: (1) STEC is calculated from dual-frequency GPS carrier phase data, (2) after eliminating the main trend in STEC by taking the numerical third order horizontal 3-point derivatives, the TIDs are isolated, (3) we assume an array signature of the TID waves, (4) we assume constant radial propagation velocity, vT, using an apparent velocity, vi, of the TID at the ith observing GNSS station, (5) since the TID’s velocity is strongly affected by the ionospheric wind velocity components, vN and vE, in the north and east directions, respectively, the unknown parameters,vT, vN, and vE, can be estimated relative to the point source epicenter, and (6) if more than six GNSS stations in good geometry observe the TID in GNSS signals, the coordinates of the epicenter can also be estimated.
Figure 3a. Pictorial representation of the scenario describing a GNSS station tracking a satellite and the ionospheric signal (3-point STEC derivative); not to scale.Figure 3b. The scenario describing a GNSS station tracking a satellite and the ionospheric signal and a point source (e.g., UNE) that generates acoustic gravity waves; not to scale.Figure 3c. The scenario describing a GNSS station tracking a satellite and the ionospheric signal, and the propagation of the acoustic gravity waves generated by a point source (e.g., UNE); not to scale.Figure 3d. The scenario describing a GNSS station tracking a satellite and the ionospheric signal, at the epoch when the GNSS signal is affected by the propagation of the acoustic gravity waves generated by a point source (e.g., UNE); not to scale.Figure 3e. Same as 3D, indicating that the geometry between GNSS station, the satellite and the IPP can be recovered and used for locating the point source; multiple GNSS stations are needed to find the point source location and the the velocity components of TID and ionospheric winds; not to scale.Figure 3f. Same as 3D, after the TID wave passed the line of sight between the GNSS stations and the satellite; not to scale.
Figure 4 illustrates the geometry of detection of the point source epicenter. Determination of the epicenter of the point source that induced TIDs can be achieved by trilateration, similarly to GPS positioning concept. The TIDs, generated at the point source, propagate at certain speed, and are detected by multiple GPS stations.
The initial assumption in our work was to use a constant propagation velocity of a TID. By observing the time of TID arrival at the ionospheric pierce point (IPP), the travel distance from the epicenter to the IPP of the GPS station that detected a TID (which is the slant distance from the ith station and the kth satellite) can be derived using a relationship with the propagation velocity. In this study, we defined a thin shell in the ionosphere F layer, 300 kilometers above the surface, and computed the IPP location for each GPS signal at the corresponding time epoch of TID detection.
Figure 4. Geometry of point source detection based on TID signals detected at the IPP of GPS station, i, with GPS satellite k. Unknown: coordinates of the point source, ( ф, λ ); three components of TID velocity vT, vN, and vE ; Observations: coordinates of IPP, (xik, yik, zik) and the corresponding time epoch to TID arrival at IPP, tik; Related terms: slant distance between IPP and UNE, sik; horizontal distance between the point source epicenter and the GPS station coordinates, di; azimuth and the elevation angle of IPP as seen from the UNE, αjk and εjk , respectively.
Very Large Array (VLA)
In addition to GNSS-based method of ionosphere monitoring, there are other more conventional techniques, for example, ground-based ionosondes, high-frequency radars, Doppler radar systems, dual-frequency altimeter, and radio telescopes. In our research, we studied the ionospheric detection of UNEs using GPS and the Very Large Array (VLA) radio telescope.
The VLA is a world-class UHF/VHF interferometer 50 miles west of Socorro, New Mexico. It consists of 27 dishes in a Y-shaped configuration, each one 25 meters in diameter, cycled through four configurations (A, B, C, D) spanning 36, 11, 3.4, and 1 kilometers, respectively. The instrument measures correlations between signals from pairs of antennas, used to reconstruct images of the sky equivalent to using a much larger single telescope. While conducting these observations, the VLA provides 27 parallel lines of sight through the ionosphere toward cosmic sources.
Past studies have shown that interferometric radio telescopes like the VLA can be powerful tools for characterizing ionospheric fluctuations over a wide range of amplitudes and scales. We used these new VLA-based techniques and a GPS-based approach to investigate the signature of a TID originated by a UNE jointly observed by both GPS and the VLA. For this case study, we selected one of the 1992 U.S. UNEs for which simultaneous GPS and VLA data were available.
Table 1. Characteristics of the analyzed events (UNEs).
Experimental Results
We summarize here the test studies performed by the OSU group in collaboration with Miami University and the U.S. Naval Research Laboratory on detection and discrimination of TIDs resulting from UNEs using the GNSS-based and VLA-based techniques. Table 1 lists the UNE events that have been analyzed to date. As of March 2013, the results of the 2013 North Korean UNE were not fully completed, so they are not included here.
In the 2006 and 2009 North Korean UNE experiments, STEC data from six and 11 nearby GNSS stations, respectively, were used. Within about 23 minutes to a few hours since the explosion, the GNSS stations detected the TIDs, whose arrival time for each station formulated the linear model with respect to the distance to the station. TIDs were observed to propagate with speeds of roughly 150–400 m/s at stations about 365 km to 1330 km from the explosion site. Considering the ionospheric wind effect, the wind-adjusted TIDs located the UNE to within about 2.7 km of its seismically determined epicenter (for the 2009 event; no epicenter location was performed for the 2006 event due to insufficient data). The coordinates estimated by our algorithm are comparable to the seismically determined epicenter, with the accuracy close to the seismic method itself. It is important to note that the accuracy of the proposed method is likely to improve if the stations in better geometry are used and more signals affected by a TID can be observed. An example geometry of UNE detection is shown in Figure 5.
Figure 5. Locations of the underground nuclear explosion (UNE) in 2009 and GNSS stations C1 (CHAN), C2 (CHLW), D1 (DAEJ), D2 (DOND), I1 (INJE), S1 (SUWN), S2 (SHAO), S3 (SOUL), U1 (USUD), Y1 (YANP), Y2 (YSSK) on the coastline map around Korea, China, and Japan. The TID waves are highlighted for stations C1, D1, D2, I1. The bold dashed line indicates the ground track for satellite PRN 26 with dots that indicating the arrival times of the TIDs at their IPPs. All time labels in the figure are in UTC.
For the Hunters Trophy and the Divider UNE tests, the array signature of TIDs at the vicinity of GPS stations was observed for each event. By applying the first-order polynomial model to compute the approximate velocity of TID propagation for each UNE, the data points — that is the TID observations — were fit to the model within the 95 percent confidence interval, resulting in the propagation velocities of 570 m/s and 740 m/s for the Hunters Trophy and the Divider, respectively.
The VLA has observing bands between 1 and 50 GHz, and prior to 2008 had a separate VHF system with two bands centered at 74 and 330 MHz. A new wider-band VHF system is currently being commissioned. The VHF bands and L-band (1.4 GHz) are significantly affected by the ionosphere in a similar way as the GPS signal. In this study, we used VLA observations at L-band of ionospheric fluctuations as an independent verification of the earlier developed method based on the GNSS TID detection for UNE location and discrimination from TIDs generated by other types of point sources.
The VLA, operated as an interfer-ometer, measures the correlation of complex voltages from each unique pair of antennas (baselines), to produce what are referred to as visibilities. Each antenna is pointed at the same cosmic source; however, due to spatial separation, each antenna’s line of sight passes through a different part of the ionosphere. Consequently, the measured visibilities include an extra phase term due to the difference in ionospheric delays, which translates to distortions in the image made with the visibilities. This extra phase term is proportional to the difference in STEC along the lines of sight of the two telescopes that form a baseline. Thus, the interferometer is sensitive to the STEC gradient rather than STEC itself, which renders it capable of sensing both temporal and spatial fluctuations in STEC.
The spectral analysis was performed on the STEC gradients recovered from each baseline that observed the Hunters Trophy event. Briefly, a time series of the two-dimensional STEC gradient is computed at each antenna. Then, a three-dimensional Fourier transform is performed, one temporal and two spatial, over the array and within a given time period (here ~15 minutes). The resulting power spectrum then yields information about the size, direction, and speed of any detected wavelike disturbances within the STEC gradient data.
Roughly 20 to 25 minutes after the UNE, total fluctuation power increased dramatically (by a factor of about 5×103). At this time, the signature of waves moving nearly perpendicular to the direction from Hunters Trophy (toward the northeast and southwest) was observed using the three-dimensional spectral analysis technique. These fluctuations had wavelengths of about 2 km and inferred speeds of 2-8 m s-1. This implies that they are likely due to small-scale distortions moving along the wavefront, not visible with GPS. Assuming that these waves are associated with the arrival of disturbances associated with the Hunters Trophy event, a propagation speed of 570–710 m/s was calculated, which is consistent with the GPS results detailed above.
In addition, a TID, possibly induced by the February 12, 2013, North Korean UNE, was also detected using the nearby IGS stations, by the detection algorithm referred to earlier. Eleven TID waves were found from ten IGS stations, which were located in South Korea, Japan, and Russia. Due to the weakness of the geometry, the epicenter and the ionospheric wind velocity were not determined at this point. The apparent velocity of TID was roughly about 330–800 m/s, and was calculated using the arrival time of the TID after the UNE epoch and the slant distance between the corresponding IPP and the epicenter. The reported explosion yield was bigger, compared to the 2009 North Korean UNE, which possibly affected the propagation velocity by releasing a stronger energy. However, more in-depth investigation of this event and the corresponding GPS data is required.
Conclusions
Research shows that UNEs disturb the ionosphere, which results in TIDs that can be detected by GNSS permanent tracking stations as well as the VLA. We have summarized several GNSS-based TID detections induced by various UNEs, and verified the GNSS-based technique independently by a VLA-based method using the 1992 U.S. UNE, Hunters Trophy. It should be noted that VLA observation was not available during the time of the Divider UNE test; hence, only the Hunters Trophy was jointly detected by GPS and the VLA. Our studies performed to date suggest that the global availability of GNSS tracking networks may offer a future UNE detection method, which could complement the International Monitoring System (IMS).
We have also shown that radio-frequency arrays like the VLA may also be a useful asset for not only detecting UNEs, but for obtaining a better understanding of the structure of the ionospheric waves generated by these explosions. The next generation of HV/VHF telescopes being developed (such as the Lower Frequency Array in the Netherlands, the Long Wavelength Array in New Mexico, the Murchison Widefield Array in Australia) utilize arrays of dipole antennas, which are much cheaper to build and operate and are potentially portable.
It is conceivable that a series of relatively economical and relocatable arrays consisting of these types of dipoles could provide another valuable supplement to the current IMS in the future, particularly for low-yield UNEs that may not be detectable with GPS.
Acknowledgment
This article is based on a paper presented at the Institute of Navigation Pacific PNT Conference held April 22–25, 2013, in Honolulu, Hawaii.
Dorota A. Grejner-Brzezinska is a professor and chair, Department of Civil, Environmental and Geodetic Engineering, and director of the Satellite Positioning and Inertial Navigation (SPIN) Laboratory at The Ohio State University.
Jihye Park recently completed her Ph.D. in Geodetic Science program at The Ohio State University. She obtained her B.A. and M.S degrees in Geoinformatics from The University of Seoul, South Korea.
Joseph Helmboldt is a radio astronomer within the Remote Sensing Division of the U.S. Naval Research Laboratory.
Ralph R.B. von Frese is a professor in the Division of Earth and Planetary Sciences of the School of Earth Sciences at Ohio State University.
Thomas Wilson is a radio astronomer within the Remote Sensing Division of the U.S. Naval Research Laboratory.
Yu (Jade) Morton is a professor in the Department of Electrical and Computer Engineering at Miami University.
Emerging GNSS applications in automobiles support regulation, security, safety, and financial transactions, as well as navigation, guidance, traffic information, and entertainment. The GNSS sub-systems and onboard applications must demonstrate robustness under a range of environments and varying threats. A dedicated automotive GNSS test center enables engineers to design their own GNSS test scenarios including urban canyons, tunnels, and jamming sources at a controlled test site.
By Mark Dumville, William Roberts, Dave Lowe, Ben Wales, NSL, Phil Pettitt, Steven Warner, and Catherine Ferris, innovITS
Satellite navigation is a core component within most intelligent transport systems (ITS) applications. However, the performance of GNSS-based systems deteriorates when the direct signals from the satellites are blocked, reflected, and when they are subjected to interference. As a result, the ability to simulate signal blockage via urban canyons and tunnels, and signal interference via jamming and spoofing, has grown fundamental in testing applications.
The UK Center of Excellence for ITS (innovITS), in association with MIRA, Transport Research Laboratory (TRL), and Advantage West Midlands, has constructed Advance, a futuristic automotive research and development, and test and approvals center. It provides a safe, comprehensive, and fully controllable purpose-built road environment, which enables clients to test, validate and demonstrate ITS. The extensive track layout, configurable to represent virtually any urban environment, enables the precise specification of road conditions and access to infrastructure for the development of ITS innovations without the usual constraints of excessive set up costs and development time.
As such, innovITS Advance has the requirement to provide cityscape GNSS reception conditions to its clients; a decidedly nontrivial requirement as the test track has been built in an open sky, green-field environment (Figure 1).
Figure 1. innovITS Advance test circuit (right) and the environment it represents (left).
NSL, a GNSS applications and development company, was commissioned by innovITS to develop Skyclone in response to this need. The Skyclone tool is located between the raw GNSS signals and the in-vehicle system. As the vehicle travels around the Advance track, Skyclone modifies the GNSS signals to simulate their reception characteristics had they been received in a city environment and/or under a jamming attack. Skyclone combines the best parts of real signals, simulated scenarios, and record-and-replay capabilities, all in one box. It provides an advanced GNSS signal-processing tool for automotive testing, and has been specifically developed to be operated and understood by automotive testing engineers rather than GNSS experts.
Skyclone Concept
Simulating and recreating the signal-reception environment is achieved through a mix of software and hardware approaches. Figure 2 illustrates the basic Skyclone concept, in which the following operations are performed.
In the office, the automotive engineer designs a test scenario representative of a real-world test route, using a 3D modelling tool to select building types, and add tunnels/underpasses, and jammer sources. The test scenario is saved onto an SD card for upload onto the Skyclone system.
The 3D model in Skyclone contains all of the required information to condition the received GNSS signals to appear to have been received in the 3D environment.
The Skyclone system is installed in a test vehicle that receives the open-air GNSS signals while it is driven around the Advance track circuit.
The open-air GNSS signals are also received at a mobile GNSS reference receiver, based on commercial off-the-shelf GNSS technology, on the test vehicle. It determines the accurate location of the vehicle using RTK GNSS. The RTK base station is located on the test site.
The vehicle’s location is used to access the 3D model to extract the local reception conditions (surrounding building obstructions, tunnels attenuations, jamming, and interference sources) associated with the test scenario.
Skyclone applies satellite masking, attenuation, and interference models to condition/manipulate raw GNSS signals received at a second software receiver in the onboard system. The software receiver removes any signals that would have been obstructed by buildings and other structures, and adds attenuation and delays to the remaining signals to represent real-world reception conditions. Furthermore, the receiver can apply variable interference and/or jamming signatures to the GNSS signals.
The conditioned signals are then transmitted to the onbaord unit (OBU) under test either via direct antenna cable, or through the air under an antenna hood (acting as an anechoic chamber on top of the test vehicle). Finally, the GNSS signals produced by Skyclone are processed by the OBU, producing a position fix to be fed into the application software.
Figure 2. Skyclone system concept.
The Skyclone output is a commercial OBU application that has been tested using only those GNSS signals that the OBU receiver would have had available if it was operating in a real-world replica environment to that which was simulated within the Skyclone test scenario.
Skyclone Architecture
The Skyclone system architecture (Figure 3) consists of five principal subsystems.
Office Subsystem Denial Scenario Manager. This software has been designed to allow users to readily design a cityscape for use within the Skyclone system. The software allows the users to select different building heights and styles, add GNSS jamming and interference, and select different road areas to be treated as tunnels.
Figure 3. Baseline Skyclone system architecture.
City Buildings. The Advance test site and surrounding area have been divided into 14 separate zones, each of which can be assigned a different city model. Ten of the zones fall inside of the test road circuit and four are external to the test site. Each zone is color-coded for ease of identification (Figure 4).
Figure 4. Skyclone city zones.
The Skyclone system uses the city models to determine GNSS signal blockage and multipath for all positions on the innovITS Advance test site. The following city models, ordered in decreasing building height and density, can be assigned to all zones: high rise, city, semi urban, residential, and parkland.
Interference and Jamming. GNSS jamming and interference can be applied to the received GNSS signals. Jamming is set by specifying a jamming origin, power, and radius. The power is described by the percentage of denied GNSS signal at the jamming origin and can be set in increments of 20 percent. The denied signal then decreases linearly to the jammer perimeter, outside of which there is no denial.
The user can select the location, radius, and strength of the jammer, can select multiple jammers, and can drag and drop the jammers around the site.
Tunnels. Tunnels can be applied to the cityscape to completely deny GNSS signals on sections of road. The user is able to allocate “tunnels” to a pre-defined series of roads within the test site. The effect of a tunnel is to completely mask the sky from all satellites.
Visualization. The visualization display interface (Figure 5) provides a graphical representation of the scenario under development, including track layout, buildings, locations, and effects of interference/jammers and tunnels. Interface/jammer locations are shown as hemispherical objects located and sized according to user definition. Tunnels appear as half-cylinder pipes covering selected roads.
Figure 5. 3D visualisation display.
Reference Subsystem
The reference subsystem obtains the precise location of the test vehicle within the test site. The reference location is used to extract relevant vehicle-location data, which is used to condition the GNSS signals.
The reference subsystem is based on a commercial off-the-shelf real-time kinematic GPS RTK system, capable of computing an accurate trajectory of the vehicle to approximately 10 centimeters. This position fix is used to compute the local environmental parameters that need to be applied to the raw GNSS signals to simulate the city scenario.
A dedicated RTK GNSS static reference system (and UHF communications links) is provided within the Skyclone system. RTK vehicle positions of the vehicles are also communicated to the 4G mesh network on the Advance test site for tracking operational progress from the control center.
Vehicle Subsystem
The vehicle subsystem acquires the GNSS signals, removes those that would be blocked due to the city environment (buildings/tunnels), conditions remaining signals, applies interference/jammer models, and re-transmits resulting the GNSS signals for use by the OBU subsystem.
The solution is based on the use of software GNSS receiver technology developed at NSL. In simple terms, the process involves capturing and digitizing the raw GNSS signals with a hardware RF front end. Figure 6 shows the system architecture, and Figure 7 shows the equipment in the innovITS demonstration vehicle.
Figure 6. Skyclone hardware architecture.
The digitized signals are then processed in NSL’s software receiver running on a standard commercial PC motherboard. The software receiver includes routines for signal acquisition and tracking, data demodulation and position determination.
In the Skyclone system, the raw GNSS signals are captured and digitized using the NSL stereo software receiver. The software receiver determines which signals are to be removed (denied), which signals require conditioning, and which signals can pass through unaffected. The subsystem does this through accurate knowledge of the vehicle’s location (from the reference subsystem), knowledge of the environment (from the office subsystem), and knowledge of the satellite locations (from the vehicle subsystem itself).
The Skyclone vehicle subsystem applies various filters and produces a digital output stream. This stream is converted to analog and upconverted to GNSS L1 frequency, and is sent to the transmitter module located on the same board.
The Skyclone transmitter module feeds the analog RF signal to the OBU subsystem within the confines of a shielded GPS hood, which is attached to the vehicle on a roof rack. An alternative to the hood is to integrate directly with the cable of the OBU antenna or through the use of an external antenna port into the OBU. The vehicle subsystem performs these tasks in near real-time allowing the OBU to continue to incorporate non-GNSS navigation sensors if applicable.
Onboard Unit Subsystem
The OBU subsystem, typically a third-party device to be tested, could be a nomadic device or an OEM fitted device, or a smartphone. It typically includes a GNSS receiver, an interface, and a software application. Examples include:
Navigation system
Intelligent speed adaptation system
eCall
Stolen-vehicle recovery system
Telematics (fleet management) unit
Road-user charging onboard unit
Pay-as-you-drive black-box
Vehicle-control applications
Cooperative active safety applications
Vehicle-to-vehicle and vehicle-to-infrastructure systems.
Tools Subsystem Signal Monitor
The Skyclone Monitor tool provides a continuous monitoring service of GNSS performance at the test site during tests, monitoring the L1 frequency and analyzing the RF singal received at the reference antenna. The tool generates a performance report to provide evidence of the open-sky GNSS conditions. This is necessary in the event of poor GNSS performance that may affect the outcome of the automotive tests. The Skyclone Monitor (Figure 8) is also used to detect any spurious leaked signals which will highlight the need to check the vehicle subsystem. If any spurious signals are detected, the Skyclone system is shut down so as to avoid an impact on other GNSS users at the test site. A visualization tool (Visor) is used for post-test analysis displaying the OBU-determined position alongside the RTK position within the 3D environment.
Figure 8. GNSS signal and positioning monitor.Figure 9. 3D model of city.
Performance
Commissioning of the Skyclone system produced the following initial results. A test vehicle was installed with the Skyclone and RTK equipment and associated antennas.. The antennas were linked to the Skyclone system which was installed in the vehicle and powered from a 12V invertor connected to the car power supply. The output from the RTK GPS reference system was logged alongside the output of a commercial third-party GNSS receiver (acting as the OBU) interfaced to the Skyclone system. Skyclone was tested under three scenarios to provide an initial indication of behavior: city, tunnel, and jammer.
The three test cenarios were generated using the GNSS Denial Scenario Manager tool and the resulting models stored on three SD cards. The SD cards were separately installed in the Skyclone system within the vehicle before driving around the test site.
City Test. The city scenario consisted of setting all of the internal zones to “city” and setting the external zones to “high-rise.”
Figure 10A represents the points as provided by the RTK GPS reference system installed on the test vehicle. Figure 10B includes the positions generated by the COTS GPS OBU receiver after being injected with the Skyclone output. The effect of including the city scenario model is immediately apparent. The effects of the satellite masking and multipath model generate noise within the position tracks.
Figure 10A. City scenario: no Skyclone.Figure 10B. City scenario: withSkyclone.
Tunnel Test. The tunnel scenario consists of setting all zones to open sky. A tunnel is then inserted along the central carriageway (Figure 11). A viewer location (depicted by the red line) has been located inside the tunnel, hence the satellite masking plot in the bottom right of Figure 11 is pure red, indicating complete masking of satellite coverage. The output of the tunnel scenario is presented in Figure 12. Inclusion of the tunnel model has resulted in the removal of all satellite signals in the area of track where the tunnel was located in the city model. The color shading represents signal-to-noise ratio (SNR), an indication of those instances where the output of the test OBU receiver has generated a position fix with zero (black) signal strength, hence the output was a prediction. Thus confirming the tunnel scenario is working correctly.
Figure 11. 3D model of tunnel.Figure 12. Results.
Jammer Test. The jammer test considered the placement of a single jammer at a road intersection (Figure 13). Two tests were performed, covering low-power jammer and a high-power jammer. Figure 14A shows results from the low-power jammer. The color shading relates to the SNR as received within the NMEA output from the OBU, which continued to provide an output regardless of the jammer. However, the shading indicates that the jammer had an impact on signal reception.
Figure 13. Jammer scenario.Figure 14A. Jammer test results: low power interference.Figure 14B. Jammer test results: high-power interference.
In contrast the results of the high-power jammer (Figure 14B) show the effect of a jammer on the OBU output. The jammer denies access to GNSS signals and generates the desired result in denying GNSS signals to the OBU. Furthermore, the results exhibit features that the team witnessed during real GNSS jamming trials, most notably the wavering patterns that are output from GNSS receivers after they have regained tracking following jamming, before their internal filtering stabilizes to nominal behaviors.
The Future
The Advance test site is now available for commercial testing of GNSS based applications. Current activity involves integrating real-world GNSS jammer signatures into the Skyclone design tool and the inclusion of other GNSS threats and vulnerabilities.
Skyclone offers the potential to operate with a range of platforms other than automotive. Unmanned aerial systems platforms are under investigation. NSL is examining the integration of Skyclone features within both GNSS simulators as well as an add-on to record-and-replay tools. This would enable trajectories to be captured in open-sky conditions and then replayed within urban environments.
Having access to GNSS signal-denial capability has an immediate commercial interest within the automotive sector for testing applications without the need to invest in extensive field trials. Other domains can now benefit from such developments. The technology has been developed and validated and is available for other applications and user communities.
Downtown Seattle, a typical test-case environment.
Multi-Constellations Working in a Dense Urban Future
GNSS receivers in cell phones will soon support four or more satellite constellations and derive additional location measurements from other sources: cellular location, MEMS sensors, Wi-Fi, and others. The authors propose test standards covering these sources, meeting industry requirements for repeatable testing while considering the user experience.
By Peter Anderson, Esther Anyaegbu, and Richard Catmur
Cellular location test standards include well-defined and widely used standards for GPS-based systems in both the 3rd Generation Partnership Program cellular technologies of GSM/WCDMA/LTE, typically referenced as the 3GPP standards, and for CDMA technologies in the 3GPP2 standards. These standards provide a reference benchmark for location performance in the laboratory, when the unit under test is directly connected to the test system via a coax connection. In addition, standards are being rolled out, such as the CTIA — The Wireless Association total isotropic sensitivity (TIS) requirement, for over-the-air (OTA) testing and developed further with LTE A-GPS OTA using SUPL 2.0. These tests are typically performed in an anechoic chamber and allow the performance of the antenna to be included.
Recently developed standards such as the 3GPP Technical Specification (TS) 37.571-1 cover multi-constellation systems, typically GPS and GLONASS for a two-constellation system, or GPS, GLONASS and Galileo for a three-constellation system, with options for additionally supporting QZSS and space-based augmentation system (SBAS) satellites. During 2014, the standards will encompass additional constellations such as the BeiDou satellite system.
Figure 1A. GNSS systems available in the 2015-2020 timescale.Figure 1B. GNSS systems available in the 2015-2020 timescale.
Significant change is also happening with the additional technologies such as cellular location, Wi-Fi, and micro-electromechanical systems (MEMS) sensors providing location information. Hybrid solutions using all/any available location information from these multiple technologies present significant challenges to both the test environment and the related test standards.
The acceptance levels required for the platform integrators and their customers are becoming much more stringent, as the use cases of the location become more diverse. These present further challenges to the performance requirements for test standards for cellular location.
Measuring Performance
The rapid growth in the GNSS applications market has driven users to demand improvements in the performance and reliability of GNSS receivers. The test standards currently employed by cellular phone and network manufacturers to evaluate the performance of GNSS receivers are even more stringent than the regulatory mandates for positioning of emergency callers and other location-based services. Emergency-call positioning is an example of a service that must provide a position fix in both outdoor and indoor environments.
A user’s experience with a GNSS receiver begins when he switches on the device. The quality of his experience defines the basic performance criteria used to assess the performance of a GNSS receiver.
How long did it take to get a position fix?
How accurate is the position fix?
When the fix is lost, how long did it take the device to reacquire satellites and re-compute the fix?
These expectations define the performance of the GNSS receiver. Manufacturers use these performance metrics to compare the performance of different GNSS receivers.
The receiver’s time-to-first-fix (TTFF) depends on the initial conditions; that is, the type of acquisition aiding data (almanac data, ephemerides, knowledge of time and frequency, and so on) available to the receiver when it is switched on.
Users now expect location-based applications to work regardless of where they are and whether they are in a fixed location or on the move. They expect the same level of performance when they are indoors at home or at work, as outdoors in a rural or urban environment. This has led to an increased demand for accurate and reliable outdoor and indoor positioning.
Reacquisition time — how quickly a receiver recovers when the user goes through a pedestrian underpass or under a tunnel or a bridge, for instance — is not tested in any of the existing test standards discussed here.
The useable sensitivity of any GNSS receiver is key to its performance. It defines the availability of a GNSS positioning fix. The acquisition sensitivity defines the minimum received power level at which the receiver can acquire satellites and compute a position fix, while the tracking sensitivity of a receiver defines the minimum received power level at which a GNSS receiver is still able to track and maintain a position fix.
Different applications use different criteria to characterize the performance of a GNSS receiver. In an E911 scenario, for instance, position accuracy and response time are critical, whereas for navigation while driving, accuracy and tracking sensitivity are important. The test criteria employed by different manufacturers are intended to verify the suitability of a particular device for the required application.
The initial test conditions are defined by the manufacturers to ensure that the different devices are tested in the same way. These conditions describe how the test sessions are started, and what acquisition aiding data are available at the start of the test session.
The main divisions among performance tests are:
Laboratory-based tests, either conducted versus OTA RF testing, or simulated versus record-and-playback signal testing.
Real-world testing (field testing). This can be difficult because the test conditions are never the same. Fortunately, it is possible to record these scenarios using an RF data recorder. This allows the same real-world scenario (with the same test conditions) to be tested repeatedly in the lab.
Static scenario testing versus moving scenario testing.
Comparison tests — relative testing (comparing one receiver against another): for reported signal-to-noise ratio (SNR), reported accuracy, and repeatability tests.
Current GNSS Test Standards
Varying performance requirements test the TTFF, accuracy, multipath tolerance, acquisition, and tracking sensitivity of the GNSS receiver. The first three following are industry-defined test standards:
3GPP2 CDMA Performance Standards. The 3GPP2 CDMA test standards (C.S0036-A) are similar to the 3GPP test standards. The 3GPP2 is for CDMA cellular systems, which are synchronized to GPS time.
3GPP GNSS Performance Standards. The latest 3GPP TS 37.571-1 test standard describes the tests for the minimum performance requirements for GNSS receivers that support multi-constellations. It is slightly more stringent than the original 3GPP TS 34.171 test standard. In the 3GPP TS 37.571-1 coarse-time sensitivity test case, signals for only six satellites are generated, whereas in the TS 34.171 coarse-time sensitivity scenario, signals for eight satellites are generated.
Table 1 shows the power levels and satellite allocation for a multi-constellation 3GPP TS 37.571-1 coarse-time sensitivity test case. In this scenario, the pilot signal will always be GPS, if GPS is supported. The signal level of the pilot signal for GPS and GLONASS have been set as –142 dBm, while the non-pilot signal level for GPS and GLONASS have been set as –147 dBm.
Table 1. 3GPP TS 37.571-1 Satellite allocation.
For the 3GPP TS 37.571-1 fine-time assistance test case, six satellites are generated. For the dual-constellation fine-time test, the split is 3+3, and for a triple-constellation test case, the split is 2+2+2, as shown in Table 2.
OTA Requirements. Testing standards have been rolled out for OTA testing, where the testing is typically performed in an anechoic chamber, allowing antenna performance to be included, with tests for the receive sensitivity referenced to an isotropic antenna and over partial summations such as the upper hemisphere. They measure the TIS of the final receiver, and operator requirements typically require OTA acquisition sensitivity of –140 dBm and tracking sensitivity of –145 dBm or lower.
Other modified test standards used by manufacturers to assess the performance of the GNSS receiver include:
Nominal Accuracy Margin Test. This test is based on the 3GPP nominal accuracy test case. All signals are reduced in steps of 1 dB till the test fails to achieve a fix in 20 seconds.
Dynamic Range Margin Test. This test is based on the 3GPP dynamic range test case. All signals are reduced in steps of 1 dB till the test fails to achieve a fix in 20 seconds.
Sensitivity Coarse-Time Margin Test. This test is based on the 3GPP sensitivity coarse-time test case. Both the pilot and non-pilot signals are reduced in steps of 1dB till the test fails to achieve a fix in 20 seconds.
Pilot Sensitivity Coarse-Time Margin Test. This test is based on the 3GPP coarse-time sensitivity test case. The non-pilot signals are always kept at –152 dBm while the signal level of the pilot signal is reduced in steps of 1 dB till the test fails to achieve a fix in 20 seconds.
Non-Pilot Sensitivity Coarse-Time Margin Test. This test is based on the 3GPP coarse-time sensitivity test case. In this test, the pilot signal is always kept at –142 dBm while the signal levels of the other seven non-pilot signals are reduced in steps of 1 dB till the test fails to achieve a fix in 20 seconds.
These modified performance tests are used because they map directly to the end-user’s experience in the real world, measuring the position accuracy, response time, and sensitivity of the GNSS receiver.
Current Equipment. The equipment required for the current test standards are all GNSS multi-satellite simulator-based, either using a single constellation (for GPS), or a multi-constellation GNSS simulator as a component of a larger cellular test system.
Limitation of Current Standards
So far, tests for GNSS in cellular devices have been very much customer/manufacturer specific, starting with 3GPP-type tests, but adding to them. Each will have its own preferred type of tests, with different configurations and types of tests. They have included primarily GNSS simulator tests, either directly connected to the device under test or using radiated signals, together with some corner cases. With chips such as the ST-Ericsson CG1960 GNSS IC, this means that different tests need to be performed for each customer.
Typically the tests are focused on cold or hot TTFF type tests, or sensitivity type tests. Live signal tests have typically been used for drive tests, with a receiver being driven around an appropriate test route, normally in an urban environment. More recently RF replays have become much more widely used, but do require truth data to give validity. RF replay tests are typically used for specific difficult routes for urban drive tests or pedestrian tests.
The 3GPP types of test standards were developed to provide a simple set of repeatable tests. However, they are idealistic, and they do not relate closely to any real-world scenario, and the test connection is defined to be at the antenna port of the system. In reality, different manufacturers and network operator standards take these tests as a given, and define margins on the tests to allow for typical losses due to antennas and implementation on a platform. These margins might be as much as 8 or 10 dB. In addition, manufacturers and network operators define their own variants of the 3GPP tests to match typical real-world usage cases, such as deep indoor.
Challenges
Current location test specifications assume that the key input to the location calculation is always the GPS constellation. With the rise of additional constellations and alternative location sources, and the challenges of the urban environment, GPS will be one of many different inputs to the location position. The key for the future will be for standards focused on testing location performance, irrespective of which constellations are visible, and also being able to fully test the system performance. Tests will be suggested that allow the basic functionality of a system to be checked, but can be enhanced to stress-test the performance of a receiver. As future location systems will use all available inputs to produce a location, there will be challenges to the supporting test standards and test equipment to handle all of these in parallel.
The initial challenge for location test standards has been the use of GNSS constellations in addition to GPS. Current leading GNSS receivers in cellular devices make use of GPS, GLONASS, SBAS, and QZSS, and network-aiding information for A-GLONASS is being rolled out in the cellular networks. The 3GPP TS 37.571-1 specification has been derived from the original GPS-only specification TS 34.171, with the addition of GLONASS and Galileo constellation options. These allow single-, dual-, or triple-constellation tests to be performed. If there is GPS in the system, then GPS is viewed as the primary constellation, and tests like the sensitivity coarse-time assistance test would have a satellite from the GPS constellation with the highest signal level. The test standards also accommodate the use of some satellites from SBAS such as WAAS and QZSS. These tests require that the performance shall be met without the use of any data coming from sensors that can aid the positioning.
This is only the first stage in the rollout of new GNSS constellations, and in the near future, GNSS receivers in cellular phones will support four or more constellations, and possibly also on frequencies additional to the L1 band, covering some or all of: GPS, GLONASS, Galileo, BeiDou Phase 2, BeiDou Phase 3, QZSS, SBAS, and IRNSS.
Table 3. Suggested four-constellation mix (Pilot signal to rotate round constellations).
The challenge for the minimum-performance specifications is to accommodate these different constellations as they become fully available. For the new constellations, this will initially be purely simulator-based, but could be extended to use of live data for certain test cases as the constellations are built up. A further challenge for the test specifications is that some of the systems are regionally based, so a performance specification based on a global approach is not applicable.
Further, tests must be severe enough to stress the receiver. With multiple constellations, it can be simple to pass a test without using all available satellites or constellations.
Other Location Sources (Hybrid Solution). Within the cellular platform, location can be provided by a number of different technologies, either separately or compositely, to provide a location to the accuracy required by the user. Technologies currently available include:
Cellular network: cell ID and cell network triangulation
LTE Positioning Protocol
Fine time assistance (for aiding)
Wi-Fi network name (service set identifier, or SSID)
Wi-Fi ranging
MEMS sensors
Near-field communication
Bluetooth
Pseudolites, other beacons, coded LED lights, and so on.
Real-World Environments. Measuring performance in a real environment is becoming much more important, as the user experience becomes much more key. The product must not only pass particular specifications, but must also meet customer expectations. In the age of the blog, negative customer feedback can damage a product’s reputation. But with the various GNSS constellations and other sources of location information, performance testing is growing significantly in complexity, and test standards needed to cover this complexity will also become more complex. The simple user criteria could be stated as “I want the system to provide a rapid, accurate position wherever I am.” But how accurate?
The end-user of a location system does not use a GNSS simulator with clean signals, but a location device with live signals, often in difficult environments. This has been recognized by platform integrators, and live test routes for both urban drive and urban pedestrian routes are now required. The performance required of the receiver in these locations has also changed, from “just need to get a fix of limited accuracy” to getting accurate location information, both from a fix (even from a cold start in a built-up area), to continuous navigation (better than 30-meter accuracy 99 percent of the time) throughout a test run.
Typical environments for these test cases include locales in many major cities, such as the environment in the OPENING PHOTO of Seattle and one shown here of Seoul, Korea.
Seoul, Korea, a typical test-case environment.
Coexistence and Interference. Recent controversies have raised the profile of GNSS interference from other wireless technologies. However, within the cellular platform, significant coexistence and potential interference issues are already present. These can occur due to adjacent channel interference, or from harmonics of cellular frequencies on the platform, for example, the second harmonic of the uplink channel for LTE Band 13 overlays the BeiDou-2 frequency of 1561MHz, and the second harmonics of both Bands 13 and 14 create out-of-band emissions in the GPS band (Figures 2 and 3).
Figure 2. BeiDou and LTE bands 13/14.Figure 3. GPS and LTE bands 13/14.
Test Proliferation. The increase in the number of GNSS constellations together with the use of other location sources to provide a hybrid solution could increase the number of tests to be performed exponentially. When this is then combined with the need to test over a range of simulated and real-world locations, together with customer specific requirements, a set of tests could easily take weeks to run. It is therefore important to ensure that the cellular location test standards are carefully constructed to not significantly proliferate the number and time for tests to be performed.
Future Test Equipment
A new generation of test equipment is emerging to meet the new challenges and requirements of multi-constellation GNSS and hybrid location systems. These include:
GNSS Simulators. Simulators currently provide up to three GNSS constellations, together with augmentation systems. With the roll-out of BeiDou-2, four-constellation simulators will now be required. Currently all GNSS devices integrated in cellular platforms use the L1 band. This will also potentially change to multi-frequency use. The appropriate GNSS simulator will need to be included in the cellular test system.
New Hybrid Test Systems. As the need for testing hybrid positioning systems in cellular devices emerges, hybrid location test systems (HLTS) are becoming available that can simulate and test hybrids of A-GNSS, Wi-Fi, MEMS sensors, and cellular positioning technologies, all in one system.
Today, these test systems use separate simulators for the different individual technologies (like GNSS, Wi-Fi, and so on), but these are now being merged into multi-system simulators that combine a number of different technologies into one device (see Figure 4).
RF Replay. The use of RF replay units for replicating live trials is already widespread. This will extend with further constellations and further frequency bands.
The advantages of using RF recorded data include:
Gives real-world data, which if the location is chosen carefully will stress the device under test;
Allows use of recorded test data from several/many urban locations;
Good for drive and pedestrian test applications;
Will be integrated in the HLTS type of test system.
The disadvantages of using RF recorded data include:
Results not deterministic;
Taken at one point in time, do not allow for future development of satellite constellations;
Proprietary recording devices, difficult to define a standard;
Need to include an inertial measurement unit (IMU) to get accurate truth data.
The difficulties of using RF replays include:
Successfully integrating all the signal environment (cellular, Wi-Fi, MEMS, and so on);
Multiple runs required to give reliable data (for example, 13 runs at different times of day to give a range of satellite geometry and user speed, between rush hour and middle of night);
Multiple locations required to stress the system;
Test time can be up to a day of real-time testing to re-run tests on one location.
Proposal for Hybrid Positioning
Tests should include a mixture of simulator-based tests, RF-replay-based tests, and live tests. This would comprise the following suite:
GNSS Performance Tests. The 3GPP type of tests (TS 37.571-1) are a good starting point for a minimum performance test, but they rely on the person running the test to define the number of constellations. To automate this, there could be a single test at the start of each test sequence to identify which constellations are supported (one to four), and then the formal test run for that mix of constellations. The constellations supported should be reported as part of the test report.
An option should be provided to allow margin tests for specific tests to be run, and these should again be reported in a standard method in the test report, specifying how far the device under test exceeds the 3GPP test. The typical margins expected for a GPS-only test would be between 8 and 10 dB in the 2014 timeframe. For a multi-constellation test, it will depend on the specific constellations used, but could be between 5 and 8 dB margin.
Ideally, a multipath scenario should be created that more closely matches the environment seen in a real urban environment.
Hybrid Location Tests. The main purpose of the hybrid location test is to prove that the different components of a cellular platform providing location are all operating correctly. A basic test would provide a sequence where the different combinations providing location are tested for correct operation separately, and then together. This would not be envisaged as a complete stress test, but each technology should be running in a mode where a location solution is not simple.
A simple example sequence of tests would be:
GNSS performance test;
Cell ID static test;
Wi-Fi SSID static test
Cell ID and Wi-Fi SSID static test
Cell ID and GNSS static test (GNSS –142 dBm)
Wi-Fi SSID and GNSS static test (GNSS –142 dBm)
Cell ID, Wi-Fi SSID, and GNSS static test (GNSS –142 dBm)
Cell ID, Wi-Fi SSID, GNSS, and sensors moving test.
See how easily tests can proliferate!
A more stringent test could then be performed to stress-test the performance if required, and if required a playback test could be performed (see RF Replay test below).
The additional location sources can also aid in providing initial states and information for the position-determination system, in addition to the common assisted-GNSS information provided by the network. This will be particularly important in indoor and other environments where GNSS performance is compromised.
Further developments such as the LTE Positioning Protocol Extensions (LPPe) from the Open Mobile Alliance will also allow the sending of additional information to the device to improve the accuracy of the position. This additional information could include accurate time, altitude information, and other parameters. Future assistance standards should enhance the use of this information, and test standards should verify the correct use of this information.
RF Replay (or Playback) Tests. GNSS performance is statistical, and it is important to ensure that any tests have sufficient breadth and repetition to ensure statistical reliability. This applies to the more normal standard simulator tests, as well as to the uses of tests in the urban environment. For example, performance in the urban environment can vary significantly between two closely spaced runs, and can also be very dependent on the time of the day. A test done in the daytime may hit rush-hour traffic, whereas tests done at night will have relatively free flow, and hence faster average speeds. Additionally, the space-vehicle constellation geometry is constantly changing, which can enhance or degrade the GNSS performance. These factors need to be considered in generating any test routes.
For RF replay tests, a number of specific locations for urban driving and pedestrian routes should be specified. These locations should be based on network-operator test requirements, and include a mixture of suburban and deep urban environments (such as Tehran Street, Seoul). For each location, ten different data sets should be used, captured at different times, including peak rush hour at a specified hour. The data set should also include separate high-performance IMU data to provide truth data. To provide test consistency, a golden-standard data set should be used. But with different suppliers this would be difficult.
For pedestrian tests, a similar number of different routes should be defined, and data captured similarly. Ideally, all data useable for a hybrid solution should be captured, and available for replay. The test criteria analyzed for this could include: yield; horizontal position error, along-track error, across-track error, heading error, and speed error.
Interference Tests with Different Cellular Bands. It is important to have a standard test to demonstrate that the device under test does not have performance degradation due to interference from particular cellular subsystems interfering with the GNSS. For this test, the device should be tested in an OTA environment to ensure that all interference coupling mechanisms are present. Two tests should be performed: first, a tracking test. In this the A-GPS performance is tested by measuring the GNSS carrier-to-noise ratio for each GNSS band, while all the wireless channels on the platform are exercised sequentially. The test result would indicate the maximum number of dBs degradation that occurs.
Second, a cold-start test at –140 dBm should be performed separately while each wireless channel on the platform is exercised. Any extension in cold-start TTFF should be noted.
Conclusions
The challenges for cellular location test standards have increased significantly with the availability of new GNSS constellations, and the use of all available technologies within the cellular platform to provide the best appropriate location for the required use case. For test standards to be relevant, and also able to be run in an appropriate time, they must consider both the requirements to prove that the appropriate technology is operating correctly, and also bear a relationship to the final system performance required. This means, for example, that a multi-constellation GNSS receiver is really using all the constellations appropriately, and also that the end-user performance requirement is considered.
Existing cellular test standards are minimum performance requirements, but future standards should encapsulate the minimum performance requirements while also allowing standard extension to provide a consistent performance description.
Further to this, platform performance must be proved in all standing operating modes, which means, for example, that the cellular system be checked when operating in all supported bands.
Test equipment to support future cellular test standards is in development, but the significant challenges will be in providing equipment to fully support urban drive and pedestrian performance requirements.
In conclusion, the ability to appropriately test a hybrid location system, comprising multi-constellation GNSS and additional location technologies, presents almost as many challenges as generating the hybrid solution in the first place.
Acknowledgments
Many thanks to the GNSS team at ST-Ericsson, and at Spirent, and also to our customers for the challenges that they have presented as the required location performances have changed and increased.
Manufacturers
Figure 4 is taken from a Spirent Hybrid Location Test System (HLTS).
Peter Anderson received master’s degrees in electrical sciences from Cambridge University and in microelectronics from Durham University. Until recently, he was a GPS systems manager and the GNSS Fellow at ST-Ericsson; he is now a consultant with PZA Systems Ltd.
Esther Anyaegbu is a senior systems architect at ST-Ericsson. She earned her Ph.D. in data communications systems from the University of Leeds, where she focused on the processing of GNSS signals in the frequency domain.
Richard Catmur is head of standards development at Spirent Communications. He holds an M.A. in engineering science from Oxford University. He has served as rapporteur, editor, or major contributor to all 3GPP and OMA standards on the testing of positioning in wireless devices.
A New Approach to the Design and Development of Global Navigation Satellite Systems
By Daniele Gianni, Marco Lisi, Pierluigi De Simone, Andrea D’Ambrogio, and Michele Luglio
INNOVATION INSIGHTS by Richard Langley
MY FIRST DEGREE is in applied physics from the University of Waterloo. Founded in 1957, Waterloo was one of the first universities to introduce co-operative education. Co-operative education (or “co-op” as it is commonly known) is a program that uses both classroom study and temporary jobs to provide students with practical experience. Applied Physics was a co-op program and I worked in both industry and research environments including stints at Philips Electronics and the Atomic Energy of Canada Limited’s Chalk River Laboratories.
Both on campus and on the job, I met fellow co-op students from a variety of disciplines including mathematics (computer science) and various branches of engineering. One of those was systems design engineering or systems engineering for short. At that time, I really didn’t know much about systems engineering except that it was an all-encompassing branch of engineering and the most challenging of all of the engineering programs at Waterloo — at least according to the students in the program.
Systems engineering is an interdisciplinary field of engineering focusing on the design and management of complex engineering projects. According to the International Council on Systems Engineering, systems engineers establish processes “to ensure that the customer and stakeholder’s needs are satisfied in a high quality, trustworthy, cost efficient and schedule compliant manner throughout a system’s entire life cycle. This process is usually comprised of the following seven tasks: State the problem, Investigate alternatives, Model the system, Integrate, Launch the system, Assess performance, and Re-evaluate [or, SIMILAR, for short].”
Central to the systems engineering process and the end-product design is the generation of models. Many types of system models are used, including physical analogs, analytical equations, state machines, block diagrams, functional flow diagrams, object-oriented models, computer simulations, and even mental models. (If you want to learn a bit about mental and other kinds of models, including how to fix radios by thinking, you could do no better than to look at some of Richard Feynman’s writings including the eminently readable “Surely You’re Joking, Mr. Feynman!”: Adventures of a Curious Character.)
As aids to the modeling process, systems engineers have developed specialized modeling languages including the Unified Modeling Language (UML) and the Systems Modeling Language (SysML). These are graphical-based languages that can be used to express information or knowledge about systems in a structure that is defined by a consistent set of rules. Both UML and SysML are widely used in systems engineering. However, both are limited when it comes to representing the signal-in-space (SIS) interfaces for global navigation satellite systems.
In this month’s column, a team of authors affiliated with the Galileo project discusses the Interface Communication Modeling Language, an extension of UML that allows engineers to clearly represent SIS interfaces, critical for the design of GNSS receivers.
“Innovation” is a regular feature that discusses advances in GPS technology andits applications as well as the fundamentals of GPS positioning. The column is coordinated by Richard Langley of the Department of Geodesy and Geomatics Engineering, University of New Brunswick. He welcomes comments and topic ideas. To contact him, see the “Contributing Editors” section on page 4.
In this article, we present the results of ongoing research on the use of a modeling language, namely Interface Communication Modeling Language (ICML), for signal-in-space (SIS) interface specification of global navigation satellite systems (GNSS). Specifications based on modeling languages (also known as model-based specifications) have proven to offer a wide range of benefits to systems engineering activities, for supporting system interoperability, reducing design risk, automating software development, and so on. We argue that similar benefits can be obtained for satellite navigation systems and receivers, if a model-based approach is used for defining and expressing the SIS interface specification. In particular, we outline how a model-based SIS interface specification can support the identification of solutions to two key issues: GNSS interoperability and the design of GNSS receivers, particularly Galileo receivers. Both issues are becoming increasingly central to the Galileo program since it entered the In-Orbit-Validation (IOV) phase and is steadily approaching the 2014 milestone, when the first early services — the Open Service (OS) and the Search and Rescue Service — will be provided to users.
GNSS interoperability concerns the integration of different GNSS with the purpose of being used together, along with regional positioning systems, to provide a seamless navigation capability and improved services in terms of availability, continuity, accuracy, and integrity, for example. GNSS interoperability should be addressed in terms of intra-GNSS interoperability and GNSS-receiver interoperability. The intra-GNSS interoperability concerns the data exchanged among the GNSS, including coordination to guarantee data coherence and consistency over time. For example, GNSS may need to share terrestrial reference frames and constantly synchronize their global time references. On the other hand, GNSS-receiver interoperability concerns the capability of the receiver to use independent GNSS signals for the computation of positions globally. This capability implicitly requires that the receiver computations are decoupled from the SIS interface of any particular GNSS. A key condition to achieve this decoupling is that the SIS interface specification is available in a consistent, unambiguous, and possibly standard format, which can support engineers to more effectively design interoperable receivers. A model-based SIS interface specification would considerably facilitate this as it enables designers to use the processing capabilities of a computer system for the verification of the specification consistency and completeness, for example. Moreover, a model-based SIS interface specification would ease the visual and electronic inspection of the data messages, therefore facilitating the automatic identification of different data representations for the same orbital and temporal parameters.
The design of GNSS receivers, and particularly those for Galileo, is increasingly of interest, and a model-based SIS interface specification can similarly support the definition of future solutions. For Galileo, specifically, the receiver design is critical to support the marketing strategies that are promoting the use of Galileo services. Key issues underlying any marketing strategy concern the Galileo receiver market appealing from a cost-to-performance ratio point of view. As Galileo receivers may require new design and adaptation of existing software (SW) or hardware (HW), as well as new production chains, higher costs — in particular non-recurring ones — are likely to occur for the production of the Galileo receivers with respect to the well-established GPS receivers. As a consequence, limitations may be experienced in market penetration and in the growth velocity of Galileo receivers’ share of the receiver market. In turn, this may hinder the estimated economic return for the Galileo project.
Preventing and counteracting this possibility is therefore a critical issue if we aim to achieve the widest possible success of the Galileo project. Market barriers inherently originate from the following needs:
Designing new SW and HW solutions for Galileo receivers;
Reusing existing SW and HW for GPS receivers;
Converting existing production chains to the new Galileo-specific SW and HW solutions.
GNSS receivers often use established mathematical models that can determine the receiver position from a fundamental set of parameters, such as satellite orbit and system time. As a consequence, the intrinsic representation of the parameter set is a major factor in the adaptation of the existing design and implementation of SW and HW solutions.
To reduce the impact of the above needs, a model-based SIS interface specification may play a pivotal role in several ways, such as:
reducing ambiguities in the Galileo SIS interface specification;
enhancing the communication with the involved stakeholders;
linking the SIS interface specification to the design schemas of GNSS receivers — particularly Galileo ones — for tracing the interface elements onto the receiver functional and physical schema, thereby supporting the reuse and adaptation of existing HW and SW solutions;
supporting the model-based design of security solutions for blocking, jamming, and spoofing.
Galileo Project
In October 2012, the final two IOV satellites were launched into orbit, completing the designed configuration for the Galileo IOV phase — the initial stage of the Galileo constellation development. In this phase, preliminary validation tests will be performed and the initial navigation message will be broadcast to the Galileo ground segment for further validation. Shortly after the conclusion of this phase, a series of launches will take place to gradually deploy the remaining 26 satellites that will form the Galileo Full Operational Capability (FOC) configuration. Currently, the Galileo Early Open Service (EOS) is expected to be available by the end of 2014. The EOS will provide ranging capabilities and will enable receiver manufacturers to begin to design and test their technological solutions for Galileo receivers and Galileo overlay services, such as search and rescue.
In the meantime, the European GNSS Agency has been established and assigned the governance of the Galileo sub-systems, including activities such as:
initiating and monitoring the implementation of security procedures and performing system security audits;
system infrastructure management, maintenance, improvement, certification, and standardization, and service provision;
development and deployment of activities for the evolution and future generations of the systems, including procurement activities;
contributing to the exploitation of the systems, including the marketing and promotion of applications and services, including market analysis.
With the now-rapid development of the Galileo project, it becomes increasingly important to support the receiver manufacturers in the design and implementation of global navigation solutions based on the Galileo services. This is necessary to guarantee the widespread use of the Galileo services, particularly in an increasingly crowded GNSS panorama.
Model-Based Systems Engineering
Model-based systems engineering (MBSE) is predicated on the notion that a system is developed by use of a set of system models that evolve throughout the development lifecycle, from abstract models at the early stages down to the operational system. A visual presentation is provided by FIGURE 1, which shows the roles of MBSE approaches within the systems engineering V-shaped process. Specifically, the MBSE approaches enable the designer to effectively trace the requirements and design alternatives on the descending branch of the “V.” For the same characteristics, MBSE facilitates the verification through a model repository that interconnects not only the design products, but also the stakeholders involved in the entire process. In addition, MBSE approaches support the automatic generation of the documentation and of other artifacts, particularly software. All of these capabilities eventually enable the validation of the implementation activities on the ascending branch of the V-process. Also, in this case, MBSE and the model repository play a major role in connecting design to implementation, and users and designers to developers.
FIGURE 1. Systems engineering V-process supported by an model-based systems engineering with model repository (courtesy of the INCOSE Survey).
Main Concepts. MBSE approaches are gaining increasing popularity with the widespread adoption of standard modeling languages, such as Unified Modeling Language (UML) and Systems Modeling Language (SysML).
UML is a formally defined general-purpose graphical language and is mainly used in the context of software systems development. It has been developed and is being managed by the Object Management Group and is the core standard of the Model Driven Architecture (MDA) effort, which provides a set of standards to shift from code-centric to model-driven software development. By use of an MDA-based approach, a software system is built by specifying and executing a set of automated model transformations.
SysML is defined as an extension of UML and provides a general-purpose modeling language for systems engineering applications (See FIGURE 2). SysML supports MBSE approaches in the development of complex systems that include hardware, software, information, processes, personnel, and facilities.
FIGURE 2. UML-SysML relationships. (UML 2 is the second generation version of UML introduced in 2005.)
Advantages. With respect to the conventional document-based approaches, MBSE approaches present the following advantages:
Conformance to standard specifications and availability of development tools;
Increased level of automation due to the formal specification and execution of model transformations that take as input a model at a given level of abstraction and yield as output a refined model at a lower level of abstraction;
Better understanding of the system in its operational context;
Support for simulation activities at different levels of detail and at different development stages, from concept exploration to dynamic system optimization;
Support for the coherent extension of standard modeling languages to adapt them to a specific target or domain.
These capabilities have motivated and have been sustaining an increasing trend of moving from document-centric to model-centric systems engineering.
ICML Language
UML and SysML are widely used languages for MBSE. A plethora of tools and technologies are available to compose models, transform models into documents, derive software products from models, and share and reuse models by means of repositories. However, neither of these languages offers capabilities for the representation of SIS interfaces, which are the critical interfaces for the design of Galileo receivers. For this reason, we have introduced ICML: a modeling language that can enable a full MBSE approach for the design of Galileo receivers. Moreover, ICML extends UML, and therefore it can integrate with system specifications based on compliant technologies as well as be used within standard tools.
Layout of Interface Specification. The typical layout of ICML-based interface specification is shown in FIGURE 3. The specification covers the definition of both the message structure and conversion processes. The message structure consists of five abstraction levels, and describes how the data is structured within the message. The conversion processes describe how the data values are transformed between adjacent levels of the message specification.
FIGURE 3. Layout of ICML-based interface specifications.
The message structure is defined at five levels: Data Definition, (Logical) Binary Coding, Logical Binary Structure, Physical Binary Coding, and Physical Signal, each covering specific aspects of the SIS interface specification.
For example, the Data Definition level covers the specification of the logical data structure, which includes the data items composing the message information. A data item is either of application or control type. An application data item represents a domain-specific concept that conveys the information expected by the message recipient. On the other hand, a control data item represents a domain-independent concept that can support the correctness and integrity verification of the associated application data items. A data item can also be associated with semantic and pragmatic definitions. The former specifies the meaning of the data item and the latter specifies the contextual interpretation for the semantic definition.
Analogously, the Binary Coding level covers the specification of the binary coding for each of the data items defined at the above level. For a data item, the binary coding is represented as a binary sequence and it includes at least a sequence identifier, the semantic definition, and the pragmatic definitions. Similarly to the above level, the semantic and pragmatic definitions enrich the interface specification, conveying an accurate representation of the binary coding.
The conversion processes describe the activities to be performed for deriving message values between adjacent levels of the above structural specification. As shown in Figure 3, eight processes should be defined to specify all the conversions between adjacent levels. For example, the DataDefinition2BinaryCoding process defines the activities to be performed for the derivation of the logical binary sequences representing data values. Similarly, the LogicalBinary2PhysicalBinary process defines the activities for the implementation of convolution or encryption algorithms on the logical binary sequence. However, these processes do not always need to be explicitly defined. In particular, if the implementation of a process is trivial or standard, a textual note referring to an external document may suffice for the specification purposes.
The first prototypal version of ICML has been implemented and can be used within the open source TopCased tool. The prototypal version is available under the GNU General Public License (GPL) v3.0 from the ICML project website. We applied the profile and developed the example ICML-based specification given below.
Galileo-Like Specification. An ICML-based specification of a Galileo-like OS interface, concerning only the above-defined level 3, would display as shown in FIGURE 4. This figure specifically details a part of a reduced F/NAV (the freely accessible navigation message provided by the E5a signal for the Galileo OS) structure consisting of one data frame made up of two F/NAV subframes.
FIGURE 4. Example of ICML-based specification of an F/NAV-like message structure at the Logical Binary Coding level.
Benefits. ICML can bring the above-mentioned MBSE benefits to support GNSS interoperability and to GNSS and Galileo receiver design. For example, ICML can:
provide a reference guideline for structuring the specification data and thus facilitating the communication between the Galileo SIS designers and the receiver producers;
ease visual inspection of the specification for verification purposes and for the identification of data incompatibilities of two GNSS systems;
convey the data semantics as well as the measurement units, to guarantee that the binary data from different GNSS are correctly decoded and interpreted;
support syntactical model validation using existing tools;
provide support for future advance exploitation by means of a machine-readable data format.
In particular, the availability of a machine-readable format is also the basis for advanced use cases that can exploit the capabilities of modern computer technologies.
Advanced Future Use Cases. In line with the above-mentioned MBSE model exploitations, we foresee a number of possible exploitation cases:
Automatic generation of the interface specification documents;
Collaborative development of the interface specification;
Automatic completeness and consistency checking of the interface specification;
Integration of SIS specifications with model-driven simulation engineering approaches for the simulation of single- and multi-GNSS receivers;
Integration of SIS specifications with receiver design models in SysML, for requirements traceability and reuse of existing GNSS solutions.
The automatic generation of interface specification documents can be an important capability during the lifecycle of a specification. For example, the specification may be updated several times during the interface design, and the textual documentation may need to be produced several times. Using a model-based approach, it is possible to automate the error-prone activities related to the document writing as well as other important functions such as specification versioning.
Complex system specifications are often the product of collaborating teams, which may occasionally be geographically dispersed. Using a model-based approach, the interface specification can be stored within a version control system that can be concurrently accessed by team members.
Completeness and consistency checking is also a manual activity, which demands a high degree of mental attention, and it is consequently highly error prone. Once the specification is available in a machine-readable format, the checking can be easily automated by specifying the verification rules that the interface model must satisfy.
Existing technologies support the simulation of single- and multi-GNSS receivers. As the SIS specification has a major impact on the internal structure of the receiver, the interface specification is a key input for developing GNSS simulators as well as for determining the boundary properties of the input signal into the receiver, including the admitted analog signal and the format of the digital data.
Moreover, the model-based interface specification can be integrated with a receiver design schema in SysML. This would be important to provide traceability between the interface requirements and the receiver’s functional and physical components. In the following section, we provide an outline for a preliminary integration between the interface specification and the receiver design.
Designing Galileo Receivers
Model-based interface specifications can support the design of Galileo receivers in several ways. For example, a specification can provide a link between Galileo requirements down to the Galileo receiver specifications, as shown in FIGURE 5.
FIGURE 5. Links between ICML and SysML specifications.
This capability may be useful in several scenarios. In particular, we have identified three scenarios. Scenario 1 consists of the identification of the receiver requirements that are introduced or modified by the Galileo OS SIS, with respect to existing GPS receivers. Scenario 2 concerns the linking between the ICML specification and the receiver functional schema to identify how a Galileo receiver will differ from existing GPS solutions. Scenario 3 is a development of Scenario 1 and Scenario 2, in which the physical schema definition and the physical components identification (HW and SW) may further exploit the ICML-based approach for supporting the reuse of existing GPS components.
Below, we detail Scenario 2, introducing a simplified receiver functional schema in SysML and linking the above ICML example to the schema.
Example Functional Schema. In this section, we illustrate a preliminary SysML representation for a simplified GNSS receiver. However, the figures are meant for exemplification purposes only and are not to be considered fully realistic and detailed for real GNSS receivers. Nevertheless, the SysML hierarchical modeling capabilities can be used to further refine the model, up to a potentially infinitesimal level of detail.
A GNSS receiver functional schema has been derived from A Software-Defined GPS and Galileo Receiver: A Single-Frequency Approach (see Further Reading) and its equivalent SysML internal block diagram (IBD) is shown in FIGURE 6.
In particular, the IBD illustrates the functional blocks (instances and types) and connections among these blocks that define the GNSS receiver. In particular, each of these block types is also described in other diagrams, in which the designers can specify the operations performed by the block, the attributes of the block, the referred properties, and the defined values, for example.
In this short article, we have particularly focused on the navigation data decoder. The data decoder is defined by a Block Definition Diagram (BDD) and an IBD, which are shown in FIGURES 7 and 8, respectively.
FIGURE 7. Navigation data decoder block definition diagram.FIGURE 8. Navigation data decoder internal block diagram.
In particular, the BDD indicates that the navigation data decoder is composed of four types of blocks: shift buffer, parity checker, binary adder, and data item retriever. The shift buffer receives the incoming physical sequence of bits, which is subsequently verified by the parity checker. The verified sequence is then processed to retrieve the standard binary format from the SIS-specific logical coding for the data item. This function is guided by the data item retriever, which stores the defined properties of each incoming data item, in the form of a physical sequence of bits (level 1). As a consequence, the navigation data decoder is involved with data defined at several of the above-defined ICML levels. From this description, it is also possible to sketch the preliminary IBD diagram of Figure 8.
Using a model-based approach, it becomes easier to establish links between interface elements and the functional blocks in the receiver schema.
Moreover, these links can also be decorated with a number of properties that can be used to further describe the type of the relationship between the interface element and the functional block. The link identification is important to the receiver design in several ways. For example, linking the interface elements to the receiver functional blocks, it becomes easier to identify which functional blocks are affected by each element of the SIS interface. Moreover, the tracing can be transitively extended to the physical schema, enabling the receiver designers to more immediately identify which physical components can be reused and which ones must be replaced in existing GNSS solutions.
We exemplify the tracing of interface elements on the above data decoding functional schema in FIGURE 9. This figure shows the navigation data decoder’s BDD in conjunction with ICML level 3 elements (with a white background). As in Figure 7, the relationships are drawn in red, including a richer set of relationship qualifiers. For example, the <<use>> qualifier indicates that the originating block uses the data specified in the connected ICML element. Similarly, the <<consumes>> qualifier indicates that the originating block takes in input instances of the ICML element. ICML level 4 elements are also relevant to this BDD; however, they are not shown for the sake of conciseness.
FIGURE 9. Linking level 3 elements to the navigation data decoder block definition.
Conclusions
Galileo receivers may face market barriers that are inherently raised by the costs linked with the introduction of new technologies with respect to the existing GPS ones. In this article, we have advocated that a model-based SIS interface specification can help mitigate possible extra costs in several ways. For example, the model-based interface specification can ease the communication among stakeholders, promote the reuse and adaptation of existing GPS software and chipsets, and support the implementation of receiver-side multi-GNSS interoperability. With the objective of supporting model-based interface specifications, we have designed ICML, which has been provided with a UML profile implementation in an open-source modeling tool. We have also shown an excerpt of a possible model-based specification for a simplified Galileo OS interface. Moreover, we have outlined how the model-based specification can integrate with SysML models of GNSS receivers and support the reuse and adaptation of existing solutions. A preliminary identification of potential exploitations and further benefits is also included. Further research is ongoing to generalize the existing ICML language to more complex types of SIS interfaces.
Acknowledgments
The authors would like to thank the students Serena Annarilli and Carlo Di Bartolomei (University of Rome Tor Vergata) for implementing the first prototype version of the ICML profile. The authors would also like to thank Marco Porretta, European Space Agency (ESA) / European Space Research and Technology Centre (ESTEC), for the suggestions of the GNSS example. The ICML project has been partially sponsored by the ESA Summer of Code in Space Initiative, edition 2012. No endorsement is made for the use of ICML for the official Galileo SIS interface specification.
DANIELE GIANNI is currently a requirement engineering consultant at EUMETSAT in Germany. EUMETSAT is the European operational satellite agency for monitoring weather, climate and the environment. Gianni received a Ph.D. in computer and control engineering from University of Rome Tor Vergata (Italy), in the field of modeling and simulation, in 2007. He has previously held research appointments at ESA, Imperial College, and Oxford University.
MARCO LISI is currently GNSS services engineering manager at ESA’s Directorate of Galileo Programme and Navigation- Related Activities at ESTEC in Noordwijk, The Netherlands. He was previously responsible for system engineering, operations, and security activities in the Galileo project. He is also a special advisor to the European Commission on European space policies. Lisi has over thirty years of working experience in the aerospace and telecommunication sectors, holding management positions in R&D, and being directly involved in a number of major satellite programs, including Artemis, Meteosat Operational, Meteosat Second Generation, Globalstar, Cosmo-Skymed, and more recently Galileo.
PIERLUIGI DE SIMONE is currently working on system assembly, integration, and verification for the Galileo mission in ESA. He has worked on many software developments in the fields of graphics, safe mode software, and visual programming. He has worked on many space missions including Helios, Meteosat, Metop, Cosmo-Skymed, and Galileo. His main interests are in modeling paradigms and cryptography and he holds a master’s degree in physics from University of Rome Tor Vergata.
ANDREA D’AMBROGIO is associate professor of computer science at the University of Rome Tor Vergata. He has formerly been a research associate at the Concurrent Engineering Research Center of West Virginia University in Morgantown, West Virginia. His research interests are in the areas of engineering and validation of system performance and dependability, model-driven systems and software engineering, and distributed simulation.
MICHELE LUGLIO is associate professor of telecommunication at University of Rome Tor Vergata. He works on designing satellite systems for multimedia services both mobile and fixed. He received the Ph.D. degree in telecommunications in 1994.
FURTHER READING
• Interface Communication Modeling Language (ICML)
“A Modeling Language to Support the Interoperability of Global Navigation Satellite Systems” by D. Gianni, J. Fuchs, P. De Simone, and M. Lisi in GPS Solutions, Vol. 17, No. 2, 2013, pp. 175–198, doi: 10.1007/s10291-012-0270-z.
• Use of ICML for GNSS Signal-in-Space Interface Specification
“A Model-based Signal-In-Space Interface Specification to Support the Design of Galileo Receivers” by D. Gianni, M. Lisi, P. De Simone, A. D’Ambrogio, and M. Luglio in Proceedings of the 6th ESA Workshop on Satellite Navigation Technologies and European Workshop on GNSS Signals and Signal Processing (NAVITEC), Noordwijk, The Netherlands, December 5–7, 2012, 8 pp., doi: 10.1109/NAVITEC.2012.6423066.
“A Model-Based Approach to Signal-in-Space Specifications for Designing GNSS Receivers” by D. Gianni, J. Fuchs, P. De Simone, and M. Lisi in Inside GNSS, Vol. 6, No. 1, January/February 2011, pp. 32–39.
• Related Modeling Languages
The Unified Modeling Language Reference Manual, 2nd edition, by G. Booch, J. Rumbaugh, and I. Jacobson, published by Addison-Wesley Professional, an imprint of Pearson Education, Inc., Upper Saddle River, New Jersey, 2005.
A Practical Guide to SysML: The Systems Modeling Language, 2nd edition, by S. Friedenthal, A. Moore, and R. Steiner, published by Morgan Kaufman and the Object Management Group Press, an imprint of Elsevier Inc., Waltham, Massachusetts, 2012.
• Systems Engineering
Systems Engineering: Principles and Practice, 2nd edition, by A. Kossiakoff, W.N. Sweet, S.J. Seymour, and S.M. Biemer, published by John Wiley & Sons, Inc., Hoboken, New Jersey, 2011.
A Software-Defined GPS and Galileo Receiver: A Single-Frequency Approach by K. Borre, D.M. Akos, N. Bertelsen, P. Rinder, and S.H. Jensen, published by Birkhäuser Boston, Cambridge, Massachusetts, 2007.
• Galileo Status and Plans
“Status of Galileo” (Galileo System Workshop) by H. Tork in the Proceedings of ION GNSS 2012, the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation, Nashville, Tennessee, September 17–21, 2012, pp. 2474–2502.
“Galileo Integrated Approach to Services Provision” (Galileo System Workshop) by M. Lisi in the Proceedings of ION GNSS 2012, the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation, Nashville, Tennessee, September 17–21, 2012, pp. 2572–2596.
by James T. Curran, Mark Petovello, and Gérard Lachapelle
Within a day of their initial activation over central Europe on March 12, Galileo satellites were visible over North America. The PLAN Group of the University of Calgary was successful in capturing and processing the signals from these satellites as they emerged. Galileo PRN 11, 12, and 19 were found and tracked on E1B/C. The PLAN software GSNRx was also able to track simultaneously GPS L1 and GLONASS L1 and produce combined position solutions.
Examining the Galileo navigation message transmitted on the E1B signal, it was found that the satellite health status is flagged as E1BHS=3 meaning Signal Component currently in Test, and the data validity status is flagged as E1BDVS=1 meaning Working without Guarantee. Current Galileo-ready commercial receivers may automatically discard measurements from a satellites broadcasting such messages. Parsing the received words in the I/NAV message, it was noted that more 50 percent of them were of type 0, although all words (types 0 to 10) were decoded at some point during the test.
Data was collected using a roof-mounted NovAtel 702GG antenna and an in-house two-channel digitizing front-end clocked by a high quality OCXO and also a three-channel National Instruments front-end for post-processing. The two-channel intermediate frequency data was streamed live to a laptop computer for real-time processing with GSNRx. Two RF channels were processed, the first centered at 1574.0 MHz with an IF bandwidth of 10.0 MHz, for the GPS L1 C/A and Galileo E1B/C signals and the second centered at 1602.0 MHz again with a bandwidth of 10.0 MHz, for the GLONASS L1 OF signals. The GPS and GLONASS signals were tracked using a Kalman-filter-based tracking strategy while the Galileo signals were tracked using a specialized data-pilot algorithm.
Figure 1. Scatter plot of the north and east position
Pseudorange and Doppler observations were extracted from the tracking strategies at a rate of 2 Hz. A 2D horizontal plot of the combined GPS & GLONASS and the combined Galileo, GLONASS & GPS single-frequency single-point solutions is presented in Figure 1.
Figure 2: Skyplot of the Galileo satellites.
The pseudorange residuals are plotted against time for each PRN tracked from each of the three systems in Figure 3. It is apparent that the addition of the three Galileo observations contributes to a reduction in bias and standard deviation in the horizontal directions, showing an excellent functioning of the Galileo satellites and PLAN Group equipment and software.
Figure 3. Pseudorange residuals are plotted against time for each PRN tracked from each of the three systems.Figure 4. A screenshot of the receiver processing the data.
Monitoring the Ionosphere with Integer-Leveled GPS Measurements
By Simon Banville, Wei Zhang, and Richard B. Langley
INNOVATION INSIGHTS by Richard Langley
IT’S NOT JUST FOR POSITIONING, NAVIGATION, AND TIMING. Many people do not realize that GPS is being used in a variety of ways in addition to those of its primary mandate, which is to provide accurate position, velocity, and time information.
The radio signals from the GPS satellites must traverse the Earth’s atmosphere on their way to receivers on or near the Earth’s surface. The signals interact with the atoms, molecules, and charged particles that make up the atmosphere, and the process slightly modifies the signals. It is these modified or perturbed signals that a receiver actually processes. And should a signal be reflected or diffracted by some object in the vicinity of the receiver’s antenna, the signal is further perturbed — a phenomenon we call multipath.
Now, these perturbations are a bit of a nuisance for conventional users of GPS. The atmospheric effects, if uncorrected, reduce the accuracy of the positions, velocities, and time information derived from the signals. However, GPS receivers have correction algorithms in their microprocessor firmware that attempt to correct for the effects. Multipath, on the other hand, is difficult to model although the use of sophisticated antennas and advanced receiver technologies can minimize its effect.
But there are some GPS users who welcome the multipath or atmospheric effects in the signals. By analyzing the fluctuations in signal-to-noise-ratio due to multipath, the characteristics of the reflector can be deduced. If the reflector is the ground, then the amount of moisture in the soil can be measured. And, in wintery climes, changes in snow depth can be tracked from the multipath in GPS signals.
The atmospheric effects perturbing GPS signals can be separated into those that are generated in the lower part of the atmosphere, mostly in the troposphere, and those generated in the upper, ionized part of the atmosphere — the ionosphere. Meteorologists are able to extract information on water vapor content in the troposphere and stratosphere from the measurements made by GPS receivers and regularly use the data from networks of ground-based continuously operating receivers and those operating on some Earth-orbiting satellites to improve weather forecasts.
And, thanks to its dispersive nature, the ionosphere can be studied by suitably combining the measurements made on the two legacy frequencies transmitted by all GPS satellites. Ground-based receiver networks can be used to map the electron content of the ionosphere, while Earth-orbiting receivers can profile electron density. Even small variations in the distribution of ionospheric electrons caused by earthquakes; tsunamis; and volcanic, meteorite, and nuclear explosions can be detected using GPS.
In this month’s column, I am joined by two of my graduate students, who report on an advance in the signal processing procedure for better monitoring of the ionosphere, potentially allowing scientists to get an even better handle on what’s going on above our heads.
Representation and forecast of the electron content within the ionosphere is now routinely accomplished using GPS measurements. The global distribution of permanent ground-based GPS tracking stations can effectively monitor the evolution of electron structures within the ionosphere, serving a multitude of purposes including satellite-based communication and navigation.
It has been recognized early on that GPS measurements could provide an accurate estimate of the total electron content (TEC) along a satellite-receiver path. However, because of their inherent nature, phase observations are biased by an unknown integer number of cycles and do not provide an absolute value of TEC. Code measurements (pseudoranges), although they are not ambiguous, also contain frequency-dependent biases, which again prevent a direct determination of TEC. The main advantage of code over phase is that the biases are satellite- and receiver-dependent, rather than arc-dependent. For this reason, the GPS community initially adopted, as a common practice, fitting the accurate TEC variation provided by phase measurements to the noisy code measurements, therefore removing the arc-dependent biases. Several variations of this process were developed over the years, such as phase leveling, code smoothing, and weighted carrier-phase leveling (see Further Reading for background literature).
The main challenge at this point is to separate the code inter-frequency biases (IFBs) from the line-of-sight TEC. Since both terms are linearly dependent, a mathematical representation of the TEC is usually required to obtain an estimate of each quantity. Misspecifications in the model and mapping functions were found to contribute significantly to errors in the IFB estimation, suggesting that this process would be better performed during nighttime when few ionospheric gradients are present. IFB estimation has been an ongoing research topic for the past two decades are still remains an issue for accurate TEC determination.
A particular concern with IFBs is the common assumption regarding their stability. It is often assumed that receiver IFBs are constant during the course of a day and that satellite IFBs are constant for a duration of a month or more. Studies have clearly demonstrated that intra-day variations of receiver instrumental biases exist, which could possibly be related to temperature effects. This assumption was shown to possibly introduce errors exceeding 5 TEC units (TECU) in the leveling process, where 1 TECU corresponds to 0.162 meters of code delay or carrier advance at the GPS L1 frequency (1575.42 MHz).
To overcome this limitation, one could look into using solely phase measurements in the TEC estimation process, and explicitly deal with the arc-dependent ambiguities. The main advantage of such a strategy is to avoid code-induced errors, but a larger number of parameters needs to be estimated, thereby weakening the strength of the adjustment. A comparison of the phase-only (arc-dependent) and phase-leveled (satellite-dependent) models showed that no model performs consistently better. It was found that the satellite-dependent model performs better at low-latitudes since the additional ambiguity parameters in the arc-dependent model can absorb some ionospheric features (such as gradients). On the other hand, when the mathematical representation of the ionosphere is realistic, the leveling errors may more significantly impact the accuracy of the approach.
The advent of precise point positioning (PPP) opened the door to new possibilities for slant TEC (STEC) determination. Indeed, PPP can be used to estimate undifferenced carrier-phase ambiguity parameters on L1 and L2, which can then be used to remove the ambiguous characteristics of the carrier-phase observations. To obtain undifferenced ambiguities free from ionospheric effects, researchers have either used the widelane/ionosphere-free (IF) combinations, or the Group and Phase Ionospheric Calibration (GRAPHIC) combinations. One critical problem with such approaches is that code biases propagate into the estimated ambiguity parameters. Therefore, the resulting TEC estimates are still biased by unknown quantities, and might suffer from the unstable datum provided by the IFBs.
The recent emergence of ambiguity resolution in PPP presented sophisticated means of handling instrumental biases to estimate integer ambiguity parameters. One such technique is the decoupled-clock method, which considers different clock parameters for the carrier-phase and code measurements. In this article, we present an “integer-leveling” method, based on the decoupled-clock model, which uses integer carrier-phase ambiguities obtained through PPP to level the carrier-phase observations.
Standard Leveling Procedure
This section briefly reviews the basic GPS functional model, as well as the observables usually used in ionospheric studies. A common leveling procedure is also presented, since it will serve as a basis for assessing the performance of our new method.
Ionospheric Observables. The standard GPS functional model of dual-frequency carrier-phase and code observations can be expressed as:
(1)
(2)
(3)
(4)
where Φijis the carrier-phase measurement to satellite j on the Li link and, similarly, Pijis the code measurement on Li. The term is the biased ionosphere-free range between the satellite and receiver, which can be decomposed as:
(5)
The instantaneous geometric range between the satellite and receiver antenna phase centers is ρ j. The receiver and satellite clock errors, respectively expressed as dT and dtj, are expressed here in units of meters. The term Tj stands for the tropospheric delay, while the ionospheric delay on L1 is represented by I j and is scaled by the frequency-dependent constant μ for L2, where . The biased carrier-phase ambiguities are symbolized by and are scaled by their respective wavelengths (λi). The ambiguities can be explicitly written as:
(6)
where Ni jis the integer ambiguity, bi is a receiver-dependent bias, and bi j is a satellite-dependent bias. Similarly, Bi and Bi j are instrumental biases associated with code measurements. Finally, ε contains unmodeled quantities such as noise and multipath, specific to the observable. The overbar symbol indicates biased quantities.
In ionospheric studies, the geometry-free (GF) signal combinations are formed to virtually eliminate non-dispersive terms and thus provide a better handle on the quantity of interest:
(7)
(8)
where IFBr and IFB j represent the code inter-frequency biases for the receiver and satellite, respectively. They are also commonly referred to as differential code biases (DCBs). Note that the noise terms (ε) are neglected in these equations for the sake of simplicity.
Weighted-Leveling Procedure. As pointed out in the introduction, the ionospheric observables of Equations (7) and (8) do not provide an absolute level of ionospheric delay due to instrumental biases contained in the measurements. Assuming that these biases do not vary significantly in time, the difference between the phase and code observations for a particular satellite pass should be a constant value (provided that no cycle slip occurred in the phase measurements). The leveling process consists of removing this constant from each geometry-free phase observation in a satellite-receiver arc:
(9)
where the summation is performed for all observations forming the arc. An elevation-angle-dependent weight (w) can also be applied to minimize the noise and multipath contribution for measurements made at low elevation angles. The double-bar symbol indicates leveled observations.
Integer-Leveling Procedure
The procedure of fitting a carrier-phase arc to code observations might introduce errors caused by code noise, multipath, or intra-day code-bias variations. Hence, developing a leveling approach that relies solely on carrier-phase observations is highly desirable. Such an approach is now possible with the recent developments in PPP, allowing for ambiguity resolution on undifferenced observations. This procedure has gained significant momentum in the past few years, with several organizations generating “integer clocks” or fractional offset corrections for recovering the integer nature of the undifferenced ambiguities. Among those organizations are, in alphabetical order, the Centre National d’Études Spatiale; GeoForschungsZentrum; GPS Solutions, Inc.; Jet Propulsion Laboratory; Natural Resources Canada (NRCan); and Trimble Navigation. With ongoing research to improve convergence time, it would be no surprise if PPP with ambiguity resolution would become the de facto methodology for processing data on a station-by-station basis. The results presented in this article are based on the products generated at NRCan, referred to as “decoupled clocks.”
The idea behind integer leveling is to introduce integer ambiguity parameters on L1 and L2, obtained through PPP processing, into the geometry-free linear combination of Equation (7). The resulting integer-leveled observations, in units of meters, can then be expressed as: (10)
where and are the ambiguities obtained from the PPP solution, which should be, preferably, integer values. Since those ambiguities are obtained with respect to a somewhat arbitrary ambiguity datum, they do not allow instant recovery of an unbiased slant ionospheric delay. This fact was highlighted in Equation (10), which indicates that, even though the arc-dependency was removed from the geometry-free combination, there are still receiver- and satellite-dependent biases (br and b j, respectively) remaining in the integer-leveled observations. The latter are thus very similar in nature to the standard-leveled observations, in the sense that the biases br and b jreplace the well-known IFBs. As a consequence, integer-leveled observations can be used with any existing software used for the generation of TEC maps. The motivation behind using integer-leveled observations is the mitigation of leveling errors, as explained in the next sections.
Slant TEC Evaluation
As a first step towards assessing the performance of integer-leveled observations, STEC values are derived on a station-by-station basis. The slant ionospheric delays are then compared for a pair of co-located receivers, as well as with global ionospheric maps (GIMs) produced by the International GNSS Service (IGS).
Leveling Error Analysis. Relative leveling errors between two co-located stations can be obtained by computing between-station differences of leveled observations:
(11)
where subscripts A and B identify the stations involved, and εl is the leveling error. Since the distance between stations is short (within 100 meters, say), the ionospheric delays will cancel, and so will the satellite biases (b j) which are observed at both stations. The remaining quantities will be the (presumably constant) receiver biases and any leveling errors. Since there are no satellite-dependent quantities in Equation (11), the differenced observations obtained should be identical for all satellites observed, provided that there are no leveling errors. The same principles apply to observations leveled using other techniques discussed in the introduction. Hence, Equation (11) allows comparison of the performance of various leveling approaches.
This methodology has been applied to a baseline of approximately a couple of meters in length between stations WTZJ and WTZZ, in Wettzell, Germany. The observations of both stations from March 2, 2008, were leveled using a standard leveling approach, as well as the method described in this article. Relative leveling errors computed using Equation (11) are displayed in Figure 1, where each color represents a different satellite. It is clear that code noise and multipath do not necessarily average out over the course of an arc, leading to leveling errors sometimes exceeding a couple of TECU for the standard leveling approach (see panel (a)). On the other hand, integer-leveled observations agree fairly well between stations, where leveling errors were mostly eliminated. In one instance, at the beginning of the session, ambiguity resolution failed at both stations for satellite PRN 18, leading to a relative error of 1.5 TECU, more or less. Still, the advantages associated with integer leveling should be obvious since the relative error of the standard approach is in the vicinity of -6 TECU for this satellite.
FIGURE 1. Relative leveling errors between stations WTZJ and WTZZ on March 2, 2008: (a) standard-leveled observations and (b) integer-leveled observations.
The magnitude of the leveling errors obtained for the standard approach agrees fairly well with previous studies (see Further Reading). In the event that intra-day variations of the receiver IFBs are observed, even more significant biases were found to contaminate standard-leveled observations. Since the decoupled-clock model used for ambiguity resolution explicitly accounts for possible variations of any equipment delays, the estimated ambiguities are not affected by such effects, leading to improved leveled observations.
STEC Comparisons. Once leveled observations are available, the next step consists of separating STEC from instrumental delays. This task can be accomplished on a station-by-station basis using, for example, the single-layer ionospheric model. Replacing the slant ionospheric delays (I j) in Equation (10) by a bilinear polynomial expansion of VTEC leads to:
(12)
where M(e) is the single-layer mapping function (or obliquity factor) depending on the elevation angle (e) of the satellite. The time-dependent coefficients a0, a1, and a2 determine the mathematical representation of the VTEC above the station. Gradients are modeled using Δλ, the difference between the longitude of the ionospheric pierce point and the longitude of the mean sun, and Δϕ, the difference between the geomagnetic latitude of the ionospheric pierce point and the geomagnetic latitude of the station. The estimation procedure described by Attila Komjathy (see Further Reading) is followed in all subsequent tests. An elevation angle cutoff of 10 degrees was applied and the shell height used was 450 kilometers. Since it is not possible to obtain absolute values for the satellite and receiver biases, the sum of all satellite biases was constrained to a value of zero. As a consequence, all estimated biases will contain a common (unknown) offset. STEC values, in TECU, can then be computed as:
(13)
where the hat symbol denotes estimated quantities, and is equal to zero (that is, it is not estimated) when biases are obtained on a station-by-station basis. The frequency, f1, is expressed in Hz. The numerical constant 40.3, determined from values of fundamental physical constants, is sufficiently precise for our purposes, but is a rounding of the more precise value of 40.308.
While integer-leveled observations from co-located stations show good agreement, an external TEC source is required to make sure that both stations are not affected by common errors. For this purpose, Figure 2 compares STEC values computed from GIMs produced by the IGS and STEC values derived from station WTZJ using both standard- and integer-leveled observations. The IGS claims root-mean-square errors on the order of 2-8 TECU for vertical TEC, although the ionosphere was quiet on the day selected, meaning that errors at the low-end of that range are expected. Errors associated with the mapping function will further contribute to differences in STEC values. As apparent from Figure 2, no significant bias can be identified in integer-leveled observations. On the other hand, negative STEC values (not displayed in Figure 2) were obtained during nighttimes when using standard-leveled observations, a clear indication that leveling errors contaminated the observations.
FIGURE 2. Comparison between STEC values obtained from a global ionospheric map and those from station WTZJ using standard- and integer-leveled observations.
STEC Evaluation in the Positioning Domain. Validation of slant ionospheric delays can also be performed in the positioning domain. For this purpose, a station’s coordinates from processing the observations in static mode (that is, one set of coordinates estimated per session) are estimated using (unsmoothed) single-frequency code observations with precise orbit and clock corrections from the IGS and various ionosphere-correction sources. Figure 3 illustrates the convergence of the 3D position error for station WTZZ, using STEC corrections from the three sources introduced previously, namely: 1) GIMs from the IGS, 2) STEC values from station WTZJ derived from standard leveling, and 3) STEC values from station WTZJ derived from integer leveling. The reference coordinates were obtained from static processing based on dual-frequency carrier-phase and code observations. The benefits of the integer-leveled corrections are obvious, with the solution converging to better than 10 centimeters. Even though the distance between the stations is short, using standard-leveled observations from WTZJ leads to a biased solution as a result of arc-dependent leveling errors. Using a TEC map from the IGS provides a decent solution considering that it is a global model, although the solution is again biased.
FIGURE 3. Single-frequency code-based positioning results for station WTZZ (in static mode) using different ionosphere-correction sources: GIM and STEC values from station WTZJ using standard- and integer-leveled observations.
This station-level analysis allowed us to confirm that integer-leveled observations can seemingly eliminate leveling errors, provided that carrier-phase ambiguities are fixed to proper integer values. Furthermore, it is possible to retrieve unbiased STEC values from those observations by using common techniques for isolating instrumental delays. The next step consisted of examining the impacts of reducing leveling errors on VTEC.
VTEC Evaluation
When using the single-layer ionospheric model, vertical TEC values can be derived from the STEC values of Equation (13) using:
(14)
Dividing STEC by the mapping function will also reduce any bias caused by the leveling procedure. Hence, measures of VTEC made from a satellite at a low elevation angle will be less impacted by leveling errors. When the satellite reaches the zenith, then any bias in the observation will fully propagate into the computed VTEC values. On the other hand, the uncertainty of the mapping function is larger at low-elevation angles, which should be kept in mind when analyzing the results.
Using data from a small regional network allows us to assess the compatibility of the VTEC quantities between stations. For this purpose, GPS data collected as a part of the Western Canada Deformation Array (WCDA) network, still from March 2, 2008, was used. The stations of this network, located on and near Vancouver Island in Canada, are indicated in Figure 4. Following the model of Equation (12), all stations were integrated into a single adjustment to estimate receiver and satellite biases as well as a triplet of time-varying coefficients for each station. STEC values were then computed using Equation (13), and VTEC values were finally derived from Equation (14). This procedure was again implemented for both standard- and integer-leveled observations.
FIGURE 4. Network of stations used in the VTEC evaluation procedures.
To facilitate the comparison of VTEC values spanning a whole day and to account for ionospheric gradients, differences with respect to the IGS GIM were computed. The results, plotted by elevation angle, are displayed in Figure 5 for all seven stations processed (all satellite arcs from the same station are plotted using the same color). The overall agreement between the global model and the station-derived VTECs is fairly good, with a bias of about 1 TECU. Still, the top panel demonstrates that, at high elevation angles, discrepancies between VTEC values derived from standard-leveled observations and the ones obtained from the model have a spread of nearly 6 TECU. With integer-leveled observations (see bottom panel), this spread is reduced to approximately 2 TECU. It is important to realize that the dispersion can be explained by several factors, such as remaining leveling errors, the inexact receiver and satellite bias estimates, and inaccuracies of the global model. It is nonetheless expected that leveling errors account for the most significant part of this error for standard-leveled observations.
For satellites observed at a lower elevation angle, the spread between arcs is similar for both methods (except for station UCLU in panel (a) for which the estimated station IFB parameter looks significantly biased). As stated previously, the reason is that leveling errors are reduced when divided by the mapping function. The latter also introduces further errors in the comparisons, which explains why a wider spread should typically be associated with low-elevation-angle satellites. Nevertheless, it should be clear from Figure 5 that integer-leveled observations offer a better consistency than standard-leveled observations.
FIGURE 5. VTEC differences, with respect to the IGS GIM, for all satellite arcs as a function of the elevation angle of the satellite, using (a) standard-leveled observations and (b) integer-leveled observations.
Conclusion
The technique of integer leveling consists of introducing (preferably) integer ambiguity parameters obtained from PPP into the geometry-free combination of observations. This process removes the arc dependency of the signals, and allows integer-leveled observations to be used with any existing TEC estimation software. While leveling errors of a few TECU exist with current procedures, this type of error can be eliminated through use of our procedure, provided that carrier-phase ambiguities are fixed to the proper integer values. As a consequence, STEC values derived from nearby stations are typically more consistent with each other. Unfortunately, subsequent steps involved in generating VTEC maps, such as transforming STEC to VTEC and interpolating VTEC values between stations, attenuate the benefits of using integer-leveled observations.
There are still ongoing challenges associated with the GIM-generation process, particularly in terms of latency and three-dimensional modeling. Since ambiguity resolution in PPP can be achieved in real time, we believe that integer-leveled observations could benefit near-real-time ionosphere monitoring. Since ambiguity parameters are constant for a satellite pass (provided that there are no cycle slips), integer ambiguity values (that is, the leveling information) can be carried over from one map generation process to the next. Therefore, this methodology could reduce leveling errors associated with short arcs, for instance.
Another prospective benefit of integer-leveled observations is the reduction of leveling errors contaminating data from low-Earth-orbit (LEO) satellites, which is of particular importance for three-dimensional TEC modeling. Due to their low orbits, LEO satellites typically track a GPS satellite for a short period of time. As a consequence, those short arcs do not allow code noise and multipath to average out, potentially leading to important leveling errors. On the other hand, undifferenced ambiguity fixing for LEO satellites already has been demonstrated, and could be a viable solution to this problem.
Evidently, more research needs to be conducted to fully assess the benefits of integer-leveled observations. Still, we think that the results shown herein are encouraging and offer potential solutions to current challenges associated with ionosphere monitoring.
Acknowledgments
We would like to acknowledge the help of Paul Collins from NRCan in producing Figure 4 and the financial contribution of the Natural Sciences and Engineering Research Council of Canada in supporting the second and third authors. This article is based on two conference papers: “Defining the Basis of an ‘Integer-Levelling’ Procedure for Estimating Slant Total Electron Content” presented at ION GNSS 2011 and “Ionospheric Monitoring Using ‘Integer-Levelled’ Observations” presented at ION GNSS 2012. ION GNSS 2011 and 2012 were the 24th and 25th International Technical Meetings of the Satellite Division of The Institute of Navigation, respectively. ION GNSS 2011 was held in Portland, Oregon, September 19–23, 2011, while ION GNSS 2012 was held in Nashville, Tennessee, September 17–21, 2012.
SIMON BANVILLE is a Ph.D. candidate in the Department of Geodesy and Geomatics Engineering at the University of New Brunswick (UNB) under the supervision of Dr. Richard B. Langley. His research topic is the detection and correction of cycle slips in GNSS observations. He also works for Natural Resources Canada on real-time precise point positioning and ambiguity resolution.
WEI ZHANG received his M.Sc. degree (2009) in space science from the School of Earth and Space Science of Peking University, China. He is currently an M.Sc.E. student in the Department of Geodesy and Geomatics Engineering at UNB under the supervision of Dr. Langley. His research topic is the assessment of three-dimensional regional ionosphere tomographic models using GNSS measurements.
FURTHER READING
• Authors’ Conference Papers
“Defining the Basis of an ‘Integer-Levelling’ Procedure for Estimating Slant Total Electron Content” by S. Banville and R.B. Langley in Proceedings of ION GNSS 2011, the 24th International Technical Meeting of the Satellite Division of The Institute of Navigation, Portland, Oregon, September 19–23, 2011, pp. 2542–2551.
“Ionospheric Monitoring Using ‘Integer-Levelled’ Observations” by S. Banville, W. Zhang, R. Ghoddousi-Fard, and R.B. Langley in Proceedings of ION GNSS 2012, the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation, Nashville, Tennessee, September 17–21, 2012, pp. 3753–3761.
• Errors in GPS-Derived Slant Total Electron Content
“GPS Slant Total Electron Content Accuracy Using the Single Layer Model Under Different Geomagnetic Regions and Ionospheric Conditions” by C. Brunini, and F.J. Azpilicueta in Journal of Geodesy, Vol. 84, No. 5, pp. 293–304, 2010, doi: 10.1007/s00190-010-0367-5.
“Calibration Errors on Experimental Slant Total Electron Content (TEC) Determined with GPS” by L. Ciraolo, F. Azpilicueta, C. Brunini, A. Meza, and S.M. Radicella in Journal of Geodesy, Vol. 81, No. 2, pp. 111–120, 2007, doi: 10.1007/s00190-006-0093-1.
• Global Ionospheric Maps
“The IGS VTEC Maps: A Reliable Source of Ionospheric Information Since 1998” by M. Hernández-Pajares, J.M. Juan, J. Sanz, R. Orus, A. Garcia-Rigo, J. Feltens, A. Komjathy, S.C. Schaer, and A. Krankowski in Journal of Geodesy, Vol. 83, No. 3–4, 2009, pp. 263–275, doi: 10.1007/s00190-008-0266-1.
“Undifferenced GPS Ambiguity Resolution Using the Decoupled Clock Model and Ambiguity Datum Fixing” by P. Collins, S. Bisnath, F. Lahaye, and P. Héroux in Navigation: Journal of The Institute of Navigation, Vol. 57, No. 2, Summer 2010, pp. 123–135.
Conception, Realization, Evaluation of a Seven-Element GNSS CRPA
By Frederic Leveau, Solene Boucher, Erwan Goron, and Herve Lattard
A controlled radiated pattern antenna can be an effective way to protect GPS receivers against jamming. A new CRPA, composed of seven elements, works on the E5a, E5b, E6, L2, and L1 bandwidths. This article reports on radiation pattern measurements of the array in a test facility.
Controlled radiation pattern antenna (CRPA) technique is considered to be the best GPS pre-correlation protection technique against interference. It consists of an antenna array and a processing unit that performs a phase-destructive sum of the incoming interference signals, this process being equivalent to making nulls towards interferers in the array radiation pattern.
Considering the growing Galileo system and the possible interest of the French Ministry of Defense in the Public Regulated Service (PRS) , a prospective study was undertaken to develop an array compatible with GPS M-code, Galileo PRS, and aeronautical radionavigation signals in the E5 bandwidth. The French Expertise & Procurement Defence Agency (DGA) awarded the French company SATIMO a feasibility contract to design, conceive, realize, and evaluate a circular array composed of seven elementary patch antennas (see Figure 1).
Figure 1. CRPA unit receiving satellite and jammer signals.
Product Features
SATIMO, a company specializing in R&D for antennas and in innovative antenna test ranges, has since developed this GPS-Galileo CRPA antenna, shown below.
New CRPA developed by SATIMO.
The CRPA consists of seven elementary patches covering E5a, E5b, L2, E6, L2, and L1 frequency bandwidths, using microstrip multilayer technology. Each element is housed in a 9-centimeter (diameter) by 2-centimeter (height) radome, connector excluded. In that volume, a space provision has been reserved to include a low-noise amplifier (LNA) and two filters for a sharp out-of-band rejection. As a consequence, it is possible to configure three types of arrays: passive without filters, passive with two passband filters, and finally active (including a LNA, with a gain > 26dB, NF<0.9dB) with two passband filters. The maximum gain levels in these configurations are from 3.6 dBi to 29.8 dBi. For radiation patterns, see Figure 2.
The design of the single element has been optimized to control the deviations of each patch antenna when included in a seven-element array.
To limit mutual coupling with respect to the array dimensions, the distance between the elements’ phase centers has been chosen close to 0.7 λ at L1 frequency. This value results in a 36.5-centimeter (diameter) array. The standalone antenna and the CRPA antenna have been validated through an environmental testing campaign.
Product Development
The usual iterative tuning and the optimization process for prototyping have been performed on SATIMO’s arch test range. This test facility indeed significantly reduces the time required to characterize the antenna-under-test (AUT) radiation pattern, in comparison with classical anechoic chamber test facilities.
More precisely, the arch test range instantaneously scans the field in one whole site angle cross-section plane, whereas the legacy systems mechanically scan the same cross-section plane by rotating the AUT for each incremental angle value. The spatial sampling of the near-field radiated by the AUT, thanks to a large number of probes along the arch surrounding it, enables a significant savings in time. The near-field results in the current plane can be displayed in real-time on a computer screen. Then, the rotation of AUT around its axis is automatically controlled by the measurement system, and a new acquisition is performed for each new cross-section plane. A Fourier transform computation is eventually applied to the 3D near-field to get the far-field radiation pattern.
The radiating characterization of the CRPA has been performed with a SATIMO SG24 system. With such a system, we have measured the complete 3D radiation patterns of each single element in less than 40 minutes per antenna.
Evaluation
The evaluation of the CRPA array was performed with this test bed in SATIMO’s facility (see photos below). The process begain with measuring an element alone on a ground plane, in order to extract the gain, the axial ratio, the aperture angle, the matching values, and every feature that defines a fixed-radiation pattern antenna. The evaluation secondly consisted of characterizing the array, that is, extracting the gain and the phase of each element in the array, with respect to a reference element. To implement such a reference anytime during the near-field acquisition process, the arch test range (Figure 3) is very powerful, because all the probes constantly point at the center of the array, despite AUT’s motions. On the contrary, the need for such a reference makes measurements difficult in anechoic chambers, which often require canceling out misalignments, thanks to specific motions that must be taken into account in the computations.
CRPA in measurements.CRPA in measurements.Figure 3. Arch test range working principle.
Uses
Functional tests are another important part of the CRPA unit evaluation. Usually, two kind of tests can be conducted: outdoors or in anechoic chamber.
Classical Tests. DGA plans to perform outdoor test campaigns by utilizing an array placed on the roof of an all-terrain vehicle (see photo). The array will be connected to a CRPA GPS processing unit and to a receiver in the vehicle. Some interferers will be located along the trajectory of the vehicle, according to various scenarios defining their waveforms and their power levels. The CRPA capability to reject those interferers can then be assessed. These kinds of outdoor tests naturally suit CRPA’s processing unit and array characterization, as they involve radiated GPS and interfering signals. However, these kinds of tests are not reproducible and are quite complicated to set up.
Outdoor jamming test campaign.
Some tests in anechoic chambers could be an alternative in order to obtain reproducible test results, but in that case, transmitting GPS constellation signals indoor becomes a challenge. An option could be the use of a GPS signal simulator, but this means a unique direction of arrival of GPS signals. Moreover, no dynamic trajectory could be done.
New Test Bed. DGA recently acquired a test bed, developed by INEO Defense, that enables evaluating CRPA units in conducted mode, for example. There is no longer a need to radiate either GPS signals or interfering signals. The purpose of this test bed, called BAnc de Caractérisation des Antennes Réseaux Antibrouillage (BACARA), or test bed to characterize anti-jamming antenna arrays (Figure 4 and Figure 5), is to replace the array and simulate its GPS and jamming environment. This means that it is able to create elementary antenna phase delays and gains resulting from the array geometry, by using finite impulse response (FIR) filters (Figure 6). This is the reason why this test bed must be fed with the array phase and gain measurement results obtained with the arch test range.
Figure 4. BACARA test bed.Figure 5. BACARA working principle.Figure 6. BACARA working principle.
Alternatively, these results can be obtained with traditional anechoic chamber measurements. 10 channels of a multi-channel GPS simulator, each one matched with a satellite, are used by the test bed. Thus, BACARA coherently sums GPS constellation simulator output channels and interfering signals, so as to accurately simulate the array’s behavior in the laboratory. As a result, for any CRPA processing unit, it is possible to compare the array’s impact on a processing unit with an ideal array being composed of perfect elementary antennas.
Unfortunately, BACARA currently operates on L1 or L2, but not on the E6 and E5 bandwidths. On the other hand, this test bed is able to simulate dynamic trajectories, with the mobile positions and attitudes. Up to 10 internal jammers with various waveforms can be set up, and their power levels over time are computed by software like Warfare or Matlab. A numerical calibration allows some transparency of the test bed for CRPA units under test.
Figure 7. BACARA graphical user interface.Figure 8. Examples of available simulated array geometry.
Conclusion
SATIMO, a company specializing in electromagnetic field measurements in the microwave frequency range and part of the Microwave Vision Group, has developed an array for the reception of M-code, PRS, and aeronautical radionavigation signals. This antenna array has been fully evaluated and qualified through electrical and environmental tests. The measurement methods have enabled the company to demonstrate the feasibility of the performances expected. Functional evaluations restricted to GPS are still under way. To do so, DGA will utilize its complementary outdoor and indoor test means, especially its laboratory test bed BACARA, as a tool to precisely evaluate GPS CRPA units.
Frederic Leveau works at the French MoD (DGA Information Superiority) as a radionavigation expert. His main interests are Galileo PRS prospective studies and developments and the integration of CRPA systems within French platforms.
Solene Boucher works at the French MoD (DGA Information Superiority) as a radionavigation expert. Her main interests are Galileo PRS prospective studies and developments. She is also responsible for the test bed BACARA.
Erwan Goron is an engineer at SATIMO Industries (Microwave Vision Group). His main activity is antenna conception.
Herve Lattard is an engineer at SATIMO Industries (Microwave Vision Group). His main activity is antenna conception.
A Case History Using the New Galileo E6-B/C Signal
By Sergei Yudanov, JAVAD GNSS
A method of decoding an unknown pseudorandom noise code uses a conventional GNSS antenna and receiver with modified firmware. The method was verified using the signals from the Galileo In-Orbit Validation satellites.
Decoding an unknown GNSS pseudorandom noise (PRN) code can be rather easily done using a high-gain steerable dish antenna as was used, for example, in determine the BeiDou-M1 broadcast codes before they were publicly announced. The signal-to-noise ratio within one chip of the code is sufficient to determine its sign. This article describes a method of getting this information using a conventional GNSS antenna and receiver with modified firmware. The method was verified using the signals from the Galileo In-Orbit Validation (IOV) satellites. In spite of the fact that only pilot signal decoding seems to be possible at first glance, it is shown that in practice data signals can also be decoded.
Concept
The idea is to do coherent accumulation of each chip of an unknown signal during a rather long time interval. The interval may be as long as a full satellite pass; for medium Earth orbits, this could be up to six hours. One of the receiver’s channels is configured in the same way as for signal tracking. The I and Q signal components are accumulated during one chip length in the digital signal processor, and these values are added to an array cell, referenced by chip number, by the processor. Only a limited amount of information need be known about the signal: its RF frequency; the expected chip rate; the expected total code length; and the modulation method.
The decoding of binary-phase-shift-keying (BPSK) signals (as most often used) is the subject of this article. It appears that the decoding of more complicated signals is possible too, but this should be proved. A limitation of this method (in common with that of the dish method) is the maximum total code length that can be handled: for lengths greater than one second and bitrates higher than 10,000 kilobits per second, the receiver’s resources may not be sufficient to deal with the signal.
Reconstructing the Signal’s Phase
This method requires coherency. During the full accumulation period, the phase difference between the real signal phase and the phase of the signal generated by the receiver’s channel should be much less than one cycle of the carrier frequency. Depending on the GNSS’s available signals, different approaches may be used. The simplest case is reconstruction of a third signal while two other signals on different frequencies are of known structure and can be tracked.
The main (and possibly the only significant) disturbing factor is the ionosphere. The ionospheric delay (or, more correctly, the variation of ionospheric delay) is calculated using the two known tracked signals, then the phase of the third signal, as affected by the ionosphere, is predicted.
The final formula (the calculations are trivial and are widely available in the literature) is:
where:
φ1 , f1 are the phase and frequency of the first signal in cycles and Hz, respectively
φ2 , f2 are the phase and frequency of the second signal in cycles and Hz, respectively
φ3 , f3 are the phase and frequency of the third signal in cycles and Hz, respectively.
It was confirmed that for all pass periods (elevation angles less than 10 degrees were not tested), the difference between the calculated phase and real phase was always less than one-tenth of a cycle. GPS Block IIF satellites PRN 1 and PRN 25 were used to prove this: the L1 C/A-code and L5 signals were used as the first and second signals, with the L2C signal as the third unknown.
If two known signals are not available, and the ionospheric delay cannot be precisely calculated, it is theoretically possible to obtain an estimate of the delay from one or more neighboring satellites with two signals available. Calculations and estimations should be carried out to investigate the expected precision.
The Experiment
The Galileo E6-B/C signal as currently transmitted by the IOV satellites was selected for the experiment, as its structure has not been published. The E6 signal has three components: E6-A, E6-B and E6-C. The E6-A component is part of the Galileo Public Regulated Service, while the two other components will serve the Galileo Commercial Service. The E6-B component carries a data signal, while the E6-C component is a pilot signal.
From open sources, it is known that the carrier frequency of the E6 signal is 1278.75 MHz and that the E6-B and E6-C components use BPSK modulation at 5,115 chips per millisecond with a primary code length of one millisecond. E6-B’s data rate is 1,000 bits per second and the total length of the pilot code is 100 milliseconds (a secondary code of 100 bits over 100 milliseconds is also present in the E6-C signal, which aids in signal acquisition).
A slightly modified commercial high-precision multi-GNSS receiver, with the E6 band and without the GLONASS L2 band, was used for this experiment. The receiver was connected to a conventional GNSS antenna, placed on a roof and was configured as described above. The E1 signal was used as the first signal and E5a as the second signal. The E6 code tracking (using 5,115 chip values of zero) was 100 percent guided from the E1 code tracking (the changing of the code delay in the ionosphere was ignored). The E6 phase was guided from E1 and E5a using the above equation. Two arrays for 511,500 I and Q samples were organized in firmware. The integration period was set to one chip (200 nanoseconds).
Galileo IOV satellite PRN 11 (also variously known as E11, ProtoFlight Model and GSAT0101) was used initially, and the experiment started when the satellite’s elevation angle was about 60 degrees and lasted for only about 30 minutes. Then the I and Q vectors were downloaded to a PC and analyzed.
Decoding of Pilot Signal (E6-C)
Decoding of the pilot signal is made under the assumption that any possible influence of the data signal is small because the number of ones and zeros of E6-B in each of 511,500 chips of the 100-millisecond integration interval is about the same. First, the secondary code was obtained. Figure 1 shows the correlation of the first 5,115 chips with 5,115 chips shifted by 0 to 511,500 chips. Because the initial phase of the E6 signal is unknown, two hypotheses for computing the amplitude or signal level were checked: [A] = [I] + [Q] and [A] = [I] – [Q], and the combination with the higher correlation value was selected for all further analysis.
Figure 1. Un-normalized autocorrelation of E6-C signal chips.
In Figure 1, the secondary code is highly visible: we see a sequence of 100 positive and negative correlation peaks (11100000001111 …; interpreting the negative peaks as zeros).This code is the exact complement (all bits reversed) of the published E5a pilot secondary code for this satellite. More will be said about the derived codes and their complements later. It appears that, for all of the IOV satellites, the E6-C secondary codes are the same as the E5a secondary codes.
After obtaining the secondary code, it is possible to coherently add all 100 milliseconds of the integration interval with the secondary code sign to increase the energy in each chip by 100 times. Proceeding, we now have 5,115 chips of the pilot signal — the E6-C primary code.
To understand the correctness of the procedure and to check its results, we need to confirm that there is enough signal energy in each chip. To this end, a histogram of the pilot signal chip amplitudes can be plotted (see Figure 2). We see that there is nothing in the middle of the plot. This means that all 5,115 chips are correct, and there is no chance that even one bit is wrong.
Figure 2. Histogram of pilot signal chip amplitude in arbitrary units.
But there is one effect that seems strange at first glance: instead of two peaks we have four (two near each other). We will shortly see that this phenomenon results from the influence of the E6-B data signal and it may be decoded also.
Decoding the Data Signal
The presence of four peaks in the histogram of Figure 2 was not understood initially, so a plot of all 511,500 signal code chips was made (see Figure 3).
Interestingly, each millisecond of the signal has its own distribution, and milliseconds can be found where the distribution is close to that when two signals with the same chip rate are present. In this case, there should be three peaks in the energy (signal strength) spectrum: –2E, 0, and +2E, where E is the energy of one signal (assuming the B and C signals have the same strength).
Figure 3. Plot of 511,500 signal code chip amplitudes in arbitrary units.
One such time interval (starting at millisecond 92 and ending at millisecond 97) is shown in Figure 4. The middle of the plot (milliseconds 93 to 96) shows the described behavior. Figure 5 is a histogram of signal code chip amplitude for the signal from milliseconds 93 to 96.
Figure 4. Plot of signal code chip amplitude in arbitrary units from milliseconds 93 to 96.
Then we collect all such samples (milliseconds) with the same data sign together to increase the signal level. Finally, 5,115 values are obtained. Their distribution is shown in Figure 6.
The central peak is divided into two peaks (because of the presence of the pilot signal), but a gap between the central and side peaks (unlike the case of Figure 5) is achieved. This allows us to get the correct sign of all data signal chips. Subtracting the already known pilot signal chips, we get the 5,115 chips of the data signal — the E6-B primary code. This method works when there are at least some samples (milliseconds) where the number of chips with the same data bit in the data signal is significantly more than half.
Figure 5. Histogram of signal code chip amplitude.Figure 6. Histogram of the signed sum of milliseconds chip amplitude with a noticeable presence of the data signal.
Proving the Codes
The experimentally determined E6-B and E6-C primary codes and the E6-C secondary codes for all four IOVsatellites (PRNs 11, 12, 19, and 20) were put in the receiver firmware. The receiver was then able to autonomously track the E6-B and E6-C signals of the satellites.
Initial decoding of E6-B navigation data has been performed. It appears that the data has the same preamble (the 16-bit synchronization word) as that given for the E6-B signal in the GIOVE Interface Control Document (ICD). Convolutional encoding for forward error correction is applied as described in the Galileo Open Service ICD, and 24-bit cyclic redundancy check error detection (CRC-24) is used. At the time of the analysis, all four IOV satellites transmitted the same constant navigation data message.
Plots of PRN 11 E6 signal tracking are shown in Figure 7 and in Figure 8. The determined codes may be found at env-gpsworld-integration.kinsta.cloud/galileo-E6-codes. Some of these codes may be the exact complement of the official codes since the code-determination technique has a one-half cycle carrier-phase ambiguity resulting in an initial chip value ambiguity. But from the point of view of receiver tracking, this is immaterial.
Figure 7. Signal-to-noise-density ratio of E1 (red), E5a (magenta), E5b (blue), and E6 (green) code tracking of Galileo IOV satellite PRN 11 on December 21–22, 2012.Figure 8. Pseudorange minus carrier phase (in units of meters) of E1 (red), E5a (magenta), E5b (blue), and E6 (green) code tracking of Galileo IOV satellite PRN 11 on December 21–22, 2012.
Acknowledgments
Special thanks to JAVAD GNSS’s DSP system developers. The system is flexible so it allows us to do tricks like setting the integration period to one chip, and powerful enough to be able to do required jobs within a 200-nanosecond cycle. This article was prepared for publication by Richard Langley.
Manufacturers
A JAVAD GNSS TRE-G3T-E OEM receiver, a modification of the TRE-G3T receiver, was used in the experiment, connected to a conventional JAVAD GNSS antenna. Plots of E6 code tracking of all four IOV satellites may be found on the company’s website.
Sergei Yudanov is a senior firmware developer at JAVAD GNSS, Moscow.
NOAA’s National Geodetic Survey (NGS) is conducting a 12-year project, called Gravity for the Redefinition of the American Vertical Datum (GRAV-D), to redefine the vertical datum of the United States by flying airborne gravity missions. The accuracy of the resulting vertical datum depends directly on the quality of the aircraft’s GNSS position solutions.
In August 2010, NGS issued a Kinematic GPS Challenge to seek community input on the best practices for processing this large positioning data volume. Ten international groups answered the call, submitting 16 different position solutions calculated with a variety of software and techniques. However, the majority of solutions were corrupted by a characteristic “sawtooth” pattern which was tracked back to the aircraft receiver used in the initial challenge; for this challenge reissue, a second onboard GNSS receiver is used. Also in this new call for participation, inertial measurement unit (IMU) data are made available for joint GPS+IMU processing.
“To further facilitate our software and method development, we invite interested researchers and practitioners to compute and submit solutions from samples of actual GRAV-D data,” said Gerry Mader and Theresa Diehl, NGS, in an invitation email. “In this new call, NGS requests that all participants submit a GPS-only solution utilizing the new aircraft GPS data. For those able to process with IMU data, we request additional submission of a second IMU+GPS solution. NGS would like to receive all solutions by April 1, 2013.
“This is a strictly voluntary exercise for those interested in such a comparison and we will share our results with the participants. We are also interested in possibly co-authoring a publication with the participants on the topic if results are significant.”
Detailed information on the challenge is available here:
Those interested in participating should read through the PDF (link above), then email Gerry Mader (gerald.l.mader at noaa.gov) and Theresa Diehl (theresa.diehl at noaa.gov) with any questions.
