When completed, the Karlatornet will be Sweden’s tallest building and redefine the skyline of the city of Gothenburg, rising to 74 stories and 246 meters (the Burj Khalifa in Dubai, currently the tallest human-made structure, is 828 meters high). Ensuring that the building remains stable even when deformed by very strong winds, sun exposure, seismic activity, settling or shrinkage will require very high precision construction methods. To ensure its vertical alignment, the engineers are using the core wall control survey (CWCS) method, which relies on active GNSS control points, and SinoGNSS T300S high accuracy GNSS receivers.
A SinoGNSS multi-constellation T300 GNSS receiver and a 360° prism mounted at the top of a building under construction. (Photo: ComNav Technology)
The CWCS method was first used during the construction of the Burj Khalifa and was subsequently applied in the construction of many other high-rise buildings around the world. Swedish surveying company Teodoliten decided to apply this method for the Karlatornet project. Core walls, which are key structural components of high-rises, require a layered construction approach, with multiple concrete pours for each core wall element. CWCS makes it possible to precisely align these core walls with the vertical axis of the building during construction, using GNSS receivers, total stations, inclinometers, and other tools.
When constructing a tall building, it is essential to continuously measure the positions of various elements at its summit to ensure their vertical alignment. This is typically done by placing at the top of the building four GNSS receivers — referred to in this context as active GNSS control points — each with a 360° prism at the bottom. By sighting the prisms and using the 3D coordinates from the GNSS receivers, a surveyor then sets up a total station. This obviates the need for an extensive array of ground control points, which are often not visible from the top, and for holes in the slabs to accommodate vertical laser plummets. Additionally, observations from a nearby reference station are used to post-process the data from the GNSS receivers in post-processed kinematic (PPK) mode to achieve an accuracy of a few millimeters. The Karlatornet project uses four SinoGNSS multi-constellation T300 GNSS receivers for the active control points.
It is also necessary to correct for the movement of the survey platform as the building’s main axis is flexed by the loads applied to it during construction. This is achieved by placing a series of high-precision dual-axis inclinometers along the core walls and then applying a least squares adjustment that takes into consideration the floor height of the measuring devices.
The SinoGNSS T300 receivers operating in GNSS-RTK mode also will be used to monitor and document post-construction building deformations.
What positioning technology is right for your UAV?
One of the things to evaluate is accuracy. Accuracy is important for two reasons: you want your UAV to be where it’s supposed to be, and you want to be able to accurately georeference the data you’re gathering with your payload. But (as you will no doubt have already discovered), accuracy isn’t as simple as looking for a number. You’ll have spotted various abbreviations accompanying those numbers — GPS, PPK, RTK, and GCP most commonly, but you may also see GNSS thrown into the mix. In this article, OxTS will explain what these abbreviations mean, and what they mean for your UAV project.
What are GCPs?
GCPs stands for ground control points, and they are the most inexpensive method of ensuring your data is accurately georeferenced. They are physical targets that you place on the ground, and for which you know the coordinates. Once your UAV has finished its survey, those points can be used to reference the position of your UAV in the global frame.
The biggest drawback with GCPs is that they don’t help your UAV know its own position. GCPs only help provide your UAV with a general position reference. So, GCPs aren’t any use if you want your UAV to fly pre-programmed flight plans. For that, you’ll need a solution such as an inertial navigation system (INS) paired with a GNSS receiver. GCPs can also be time consuming to use and cause additional difficulties at the post-processing stage.
What is GNSS?
GNSS stands for global navigation satellite system, which are systems that use satellite-based radio navigation to provide positioning, navigation and timing anywhere on Earth. The U.S. GPS is one of four GNSS constellations; the other three are the Russian GLONASS, the Chinese BeiDou, and the European Galileo. There are also two regional satellite navigation systems — the Indian NavIC and the Japanese QZSS.
Many UAVs will have a GNSS receiver built in — it’s what enables them to know where they are on the planet, after all. Using GNSS only, most UAVs can get accuracy of 3 m to 5 m. This level of accuracy isn’t too bad for some applications, but not accurate enough if you’re trying to use the position data for mapping activities.
What is PPK?
Most UAVs advertise their ability to perform PPK — which stands for post-processed kinematics. It’s a method of squeezing extra accuracy out of your GNSS signal. OxTS has a blog post here that describes how it works.
The main thing to note about PPK is that you can’t use it in real time. UAVs with PPK capabilities can provide data that’s centimeter-level accurate in optimum conditions, but that accuracy can’t be used for navigating the drone itself. It also means that for activities that require centimetre-level accuracy in real time, PPK doesn’t deliver.
What is RTK?
RTK is the best you can get when it comes to position accuracy. RTK stands for real-time kinematic, and just like PPK it can use it to obtain centimeter-level accuracy — but, in real time, rather than in post-processing.
For most mobile mapping activities RTK accuracy is the goal, particularly if you’re using a lidar sensor to create georeferenced pointclouds.
Without RTK accuracy for the duration of your lidar survey, your point cloud may be unusable. The additional accuracy RTK offers could be used to tackle more challenging environments — providing that you have the tools to remain with RTK accuracy for as long as possible in the absence of GNSS.
Most off-the shelf UAVs won’t have RTK capabilities built in; however, to get this level of accuracy, it’s likely that you’ll either need to purchase a top-of-the-range UAV or invest in a custom UAV (either built by you, or by a professional company).
What’s right for me?
If you’re involved in mobile mapping activities, then at the very least you will need PPK capabilities. Without those, you won’t be able to georeference your data with enough accuracy to be of use to anyone.
When considering the difference between PPK and RTK, you need to consider:
In what environment is your UAV operating? Do you need more accuracy than just a GNSS signal (remembering that PPK can only be applied after the survey takes place)?
Is range of no particular importance – or is the payload on your drone sufficiently large that you need to calculate range very carefully? If so, RTK will give your UAV additional accuracy and, therefore, fuel efficiency.
The final word in accuracy: gx/ix PPK and RTK from OxTS
If you read the blog post mentioned above, you’ll know that RTK (and PPK) rely on having an optimal number of satellites visible. If those satellites are lost, then so is RTK lock. That is, unless you use an OxTS INS with gx/ix tight coupling technology. Gx/ix allows our INS devices to maintain RTK and PPK level accuracy even if the number of visible satellites starts to drop. Essentially, it protects the accuracy of your scan for longer — and it is available on the OxTS xNAV650, our UAV-mountable INS.
The OxTS xNAV650 INS combines a best-in-class inertial measurement unit, with a survey-grade GNSS receiver to output highly accurate navigation data (position, heading, pitch and roll). The xNAV650 is used across the world for applications where reliability and accuracy are critical.
A roundup of recent products in the GNSS and inertial positioning industry from the October 2022 issue of GPS World magazine.
OEM
Software
Aids GNSS/INS installation
Photo: Septentrio
The RxLeverArm software tool aids integration of GNSS receivers that include inertial navigation systems (GNSS/INS). RxLeverArm is part of Septentrio’s RxTools software package included with every Septentrio GNSS/INS receiver. The new tool visualizes, validates and automatically calibrates the exact distance between the INS sensor and the antenna, removing the need for accurate distance measurements with complex instruments. For lever-arm compensation, users only need to measure the rough distance between the INS sensor and the main GNSS antenna reference points on the vehicle. Data is then logged under open-sky conditions, which allows the RxLeverArm tool to optimize the initial rough distance measurement and prevent common errors such as sign inversion.
Enables proof of concept for IoT products and applications
Photo: u-blox
The u-blox XPLR-IOT-1 IoT explorer kit is an all-in-one package to test, evaluate and validate applications for the internet of things (IoT). The board hosts an ultra-low-power MAX-M10S positioning module capable of concurrently tracking four GNSS constellations, delivering highly reliable location data. Integrating relevant u-blox technologies and services into a capable prototyping platform with a vast selection of sensors and interfaces as well as cloud connectivity, XPLR-IOT-1 makes it easier to explore the potential of IoT applications.
The LC29H is a dual-band multi-constellation GNSS module built using the Airoha AG3335 platform. It is available in multiple variants and optionally integrates real-time kinematic (RTK) and dead reckoning. The LC29H series offers high performance with power efficiency to meet the market needs of high-precision positioning at the centimeter and decimeter levels. The LC29H concurrently receives and processes signals from GPS, GLONASS, BeiDou, Galileo and QZSS. The modules are suited to an expanding market for autonomous lawn mowers, drones, precision agriculture, micro-mobility scooters and delivery robots.
The Cicerone LoRa/GNSS board is a high-performance, low-power, Arduino MKR-compatible development board based on the u-blox MAX-M10S GNSS module and the MAMWLE LoRa module. It delivers high-performance GNSS, long-range wireless connection, and high-performance processing in a low-power solution for optimal battery life. The board allows users to build tracking applications worldwide with meter-level accuracy and to communicate long-range, low-power data via LoRaWAN. The integrated Li-Po charging circuit enables the Cicerone board to manage battery charging through the USB port. It has a compact 63 mm x 25 mm form factor and is compatible with all Arduino MKR shield boards. These boards all share a common pinout to enable developers to easily add expansions with minimal software changes.
The Snapdragon W5 Gen 1 and W5+ Gen 1 platforms are designed to advance ultra-low power and breakthrough performance for next-generation connected wearables with a focus on extended battery life and premium user experiences. They incorporate innovations including low power islands for GNSS, Wi-Fi and audio; ultra-low power Bluetooth 5.3 architecture; and low power states such as Deep Sleep and Hibernate. New enhancements to the flagship Snapdragon W5+ platform offer 50% lower power, 2x higher performance, 2x richer features, and 30% smaller size, compared to the previous generation. The purpose-built platform is comprised of a 4 nm-based system-on-chip and 22 nm-based highly integrated always-on co-processor. By using these platforms, manufacturers can scale, differentiate and develop products faster in the continuously growing and segmenting wearables industry, Qualcomm said. Qualcomm also announced two reference designs from Compal and Pegatron, which showcase the capabilities of the platform and the company’s collaboration with ecosystem partners, helping customers develop products faster.
The pocket-sized vRTK GNSS real-time-kinematic (RTK) receiver is equipped with dual cameras to enable non-contact image surveying. It also has a nine-axis IMU module with auto installation for tilt surveying. Visual positioning technology combines imagery with high-precision positioning equipment, allowing users to obtain the location of the target from a distance. The Live View Stakeout function improves stakeout speed, while non-contact measurement greatly improves the usable range of GNSS. The vRTK receives 1,408 channels (GPS, GLONASS, BeiDou, Galileo, QZSS, IRNSS and SBAS). A new generation of GNSS engine supports the new frequency points B1C, B2a and B2b RTK decoding of BeiDou-3 satellites.
The SXblue SMART features an engine capable of tracking all-in-view GNSS signals, with interference mitigation and optimization for handling a wide frequency band. Weighing 850 g including battery, the SXblue SMART is compact and rugged. Its radio link is based on the Farlink protocol that allows a range of up to 8 km while reserving a wide bandwidth for transmission of real-time kinematic (RTK) data. In addition to a tilt sensor for measurements in hard-to-reach places, the SXblue SMART features a high-performance attitude measurement module that can detect and measure movement of the device. Also integrated are an inertial measurement unit and a thermometer for monitoring and controlling its internal temperature.
Emlid Studio is a new post-processed kinematic (PPK) application designed specifically for post-processing GNSS data. It allows users to convert raw GNSS logs into RINEX, post-process static and kinematic data, geotag images from drones (including DJI brand), and extract points from survey projects completed with Emlid’s ReachView 3 app. With Emlid Studio, users can post-process data recorded with Emlid Reach receivers and other GNSS receivers or NTRIP services. Post-processing requires RINEX observation and navigation files. Raw data in UBX and RTCM3 format also can be used through conversion.
The P1 GNSS receiver has a high-precision module that tracks GPS, GLONASS, BDS, Galileo, QZSS and SBAS to deliver centimeter-level real-time kinematic (RTK) accuracy even in harsh environments. It is also equipped with an anti-jamming and anti-spoofing algorithm. The P1 GNSS receiver has integrated the GNSS module and GNSS antenna while keeping the device as small as a smartphone, which makes it portable enough to be worn around the neck or placed in a pocket. With 4G/Bluetooth communication, the P1 supports real-time positioning data transmission, providing users with a stable correction data steam and positioning data uploads. The P1 also can be mounted on a pole.
Nuwa surveying smartphone app version 2.3.3.2 has vector map import and digital surface stakeout. The Nuwa app runs on Android and is reliable and easy to operate. It has rich and powerful functions that can help surveyors complete measurements more efficiently and accurately. The app is designed to work with the David and Oscar GNSS receivers from Tersus GNSS, plus other receivers that support NMEA-0183. Features include the ability to configure base, rover and static surveys; graphical interface with background map (online/import); CAD stakeout, road stakeout and earthwork; data management (import/export multiple formats); and Bluetooth and USB connection support.
Version 3.2 of the survey application 1Edit allows the use of Web Maps (WMS) to be used as background layers, making it easier for surveyors to identify assets and changes in context. It provides easier configuration of background maps and supports hybrid working practices for surveyors. Where offline background maps are required, 1Edit supports multiple raster files and handles large image files, providing visual context for geospatial data when there is no data signal. Enhanced support for complex geometries increases efficiency as features with multiple parts share common attributes and IDs.
The Smarty U.S. Geocoding QGIS Plugin provides an easy way for users of the software platform to validate, standardize, and convert addresses to their latitude and longitude coordinates (geocodes). The plugin allows manual address entry as well as batch geocoding via CSV. It features a 95% match rate with the actual rooftop and parcel, as well as providing sub-address geocoding that can match secondary addresses such as apartment units and office-suite rooftops in building. The free plugin also includes supplemental metadata useful for many geographic information system (GIS) purposes.
Datasets for the United States, UK, Canada, Australia and Europe
Photo: Maptitude
Maptitude 2022 is a major release of the geographic information system (GIS) and mapping software. It includes up-to-date, accurate data encompassing expenditure, geodemographic segments, gross domestic product, medical and banking locations, branded business locations, traffic counts, building footprints, address points and financial assets, as well as the tools to leverage this information to improve the location intelligence of organizations in markets such as healthcare, franchising, communications, logistics, retail, real estate and banking.
The Mesa Pro rugged tablet features 11th-generation Intel Core processors, a Windows 11 operating system, device customization options, a large sunlight-readable display and the “Juniper Rugged” company design. Standard Mesa Pro units come with an 11th Gen Intel Core i5 processor and 16 GB of LPDDR4x RAM. Core i7 and Celeron versions are also available. Each Mesa Pro configuration offers powerful performance and allows users to select the computing performance that fits their needs and budgets.
Data fusion across multiple data sources, including ADS-B
Photo: Vigilant Aerospace
FlightHorizon COMMANDER is a situational awareness and safety system for UAV airspace management. The system provides airspace managers with either a 2D or 3D view of all aircraft in the selected airspace using a combination of sensors and data sources to create an airspace safety picture for pilots, airspace managers and command centers. The system is based on an exclusively licensed NASA patent and prototype that has been used in extensive flight testing. FlightHorizon COMMANDER functions as a visualization tool for airspace management, an active situational awareness tool, and a detect-and-avoid system that enables unmanned aircraft to avoid other aircraft and keeps drone pilots and airspace managers aware of the location and air traffic around their UAS and in their airspace.
The Draganfly Heavy Lift Drone is a versatile, multi-rotor unmanned aerial vehicle designed to enhance deliveries and flight times. Compatible with a variety of interchangeable payloads, the heavy-duty drone can carry more and fly longer than the typical professional drone. It has a payload/cargo-lift capacity of 30 kg (67 lbs) and up to 55 minutes of flight time. The industrial UAV handles heavy winds and high elevations with ease. Its lifting capacity permits flexibility in carrying large high-end sensors such as hyperspectral and bathymetric lidar to conduct large-area surveys.
Allows rapid MWIR integration for commercial, industrial and defense applications
Photo: Teledyne FLIR
Part of the Neutrino IS series, the Neutrino LC CZ 15-300 is a new mid-wavelength infrared (MWIR) camera module with integrated continuous zoom lenses. Designed for integrated solutions requiring crisp, long-range MWIR imaging, the camera offers size, weight, power and cost (SWaP+C) benefits to original equipment manufacturers (OEMs) and system integrators for airborne, unmanned, C-UAS, security and targeting applications. The LC CZ 15-300 offers high performance, 640 x 512 high-definition MWIR imagery and 15 mm to 300 mm zoom capability for ruggedized products requiring long life, low power consumption and quiet, low-vibration operation. The camera module and lens are designed for each other, providing optimal performance.
A miniature drone with flapping wings was demonstrated at the Teknofest Black Sea aviation and defense industry event, which took place Aug. 30 to Sept. 4 at the Samsun Çarşamba Airport. With its low detectability, the nano drone is being developed to perform reconnaissance and surveillance missions. It is still in research and development.
The Eyeonic Vision Sensor can perceive, identify and avoid objects at a range of more than 1 kilometer. The sensor is a frequency modulated continuous wave (FMCW) lidar transceiver that uses a silicon photonic chip. Long-range visibility is a requirement for autonomous vehicles, which require sufficient awareness to evade obstacles at highway speeds. This capability requires vision sensors to provide millimeter-level accuracy and depth at instantaneous velocity. The highly detailed and ultra-long-range information from the Eyeonic Vision Sensor enables robots to classify and predict their environments. The sensor is designed to be integrated into autonomous vehicles, security solutions and industrial robots.
The nROK 1030 is a compact, rugged entry-level vehicle computer with an advanced GNSS receiver. The u-blox NEO-M9N module supports GPS, GLONASS, Galileo, BeiDou and QZSS signals. An Intel Atom x6211E dual-core processor 1.3 GHz/3 GHz (burst) is designed for harsh in-train environments. Its fanless, compact design is suitable for vehicles with limited space. The nROK 1030 has onboard CAN 2.0B for vehicle diagnostics and driver behavior management. WLAN Wi-Fi 6/6E/Wi-Fi 5 and WWAN 5G NR/LTE wireless data connectivity is optional. The nROK 1030 is flexible to meet the demands of various rolling-stock applications, such as wireless gateway, infotainment and digital radio data/voice transmission systems.
The Google Smartphone Decimeter Challenge (SDC) competition, co-sponsored by the Institute of Navigation (ION), took place this summer. For the competition, teams developed high-precision GNSS positioning using a pool of smartphone GNSS + inertial measurement unit (IMU) datasets accompanied by high-accuracy ground truth. Teams competed to achieve the best location accuracy with the datasets provided. Winners received cash prizes and sponsored attendance at the ION GNSS+ 2022 conference in Denver, Sept. 19-23, to present their results.
Origins
The SDC has its origins in the Android Operating System, which is an open-source platform. In 2016, Google made GNSS raw measurements available as a public application programming interface (API) on all Android phones. Since then, the available measurements have become more sophisticated and more accurate. For example, dual-frequency carrier-phase data is now available on many Android phones. This enables new areas of research.
Goals
The competition had two goals:
• Stimulate the research and development of high-accuracy algorithms that can produce submeter position accuracy on phones.
• Establish a publicly accessible repository of labeled data so that all future research on location algorithms can be judged in a consistent way against a standard set of data.
The first goal was met beyond our expectations. A total of 1,381 teams participated in the two competitions of 2021 and 2022. Discussion among competitors on the competition platform (kaggle.com) was wide-ranging, incredibly collegial, and beneficial to the entire community.
Competitors have written and shared detailed descriptions, and these have been reviewed and commented on by other competitors. Moreover, winners have written formally peer-reviewed papers and made presentations at the ION GNSS+ conferences, which are available from ion.org.
The second goal is a work-in-progress and is intended to be the legacy of the events.
Legacy
Disciplines such as machine learning have established benchmarks that make it possible to compare new approaches to previous ones in a proper quantitative way. In the GNSS community, this convention has been missing — a glance across papers at conferences will show that different algorithms tend to be presented with different test data and different metrics. Usually, the authors collect this data, and it is often fairly sparse (one or two drive tests, for example). Also, the reader never knows whether the data was cherry-picked (were bad results not mentioned?).
The SDC data provides:
• 206 different drive tests
•86 total hours of dual-frequency (L1, L5) data with code and carrier-phase measurements
•All labeled with ground-truth positions and velocities collected using NovAtel SPAN ISA-100C, with precise lever-arm compensation and validated with Google’s analysis tools.
The Kaggle site allowed users to submit their results, then automatically scored them against the ground-truth data. We advocate that all GNSS researchers use this resource to measure their location algorithm improvements in a standard way. This creates trust in published results, accelerating the recognition and adoption of truly great improvements for the benefit of the entire industry and GNSS users worldwide.
The top three winners of this year’s Smartphone Decimeter Challenge described their projects to Matteo Luccio, GPS World editor-in-chief.
Suzuki
Taro Suzuki, Chiba Institute of Technology
1st Place Winner: Two-Step Optimization of Velocity and Position using Smartphone’s Carrier Phase Observations
What is your research focus and how does it relate to the contest?
My current research focuses on the accurate positioning of vehicles and mobile robots in urban environments where GNSS multipath occurs. I usually use commercial GNSS receivers for my research. This competition is very relevant to my current research, except that the smartphone is replacing a receiver.
How long have you been developing the technology or approach you used to win the contest?
The competition was held for three months, but I concentrated my efforts on the past three weeks. However, I used technologies and resources developed in my previous research (for example, source code developed in last year’s competition).
Have you participated in previous editions of this contest?
Yes, I participated in the last competition and won. The approach used in this year’s competition is based on the method used to win last year’s competition, with additional innovations and improvements.
Where, in what GNSS signal conditions, and at what speeds were the test data collected?
The competition provides a training dataset, which contains raw GNSS observations from a smartphone installed on a vehicle as it travels on real roads. In addition to GNSS observations, the training dataset contains the ground truth of the smartphone’s position. The training dataset includes a wide range of GNSS signal conditions, such as driving on highways around San Francisco and Los Angeles, driving on tree-lined urban streets, and driving in tunnels and under overpasses. I have developed an algorithm that uses a training dataset containing ground truth to accurately estimate the location of smartphones in a variety of GNSS signal reception environments.
What accuracies were you able to obtain?
The competition metric was “average of 50th and 95th percentile horizontal errors.” The metrics are computed for each of the 36 runs in the test dataset, which are divided into public and private groups, then the metrics are averaged in each group to compute the final score. My final score was 1.382 m for public and 1.229 m for private. The best score given after the competition was 1.372 m for public and 1.197 m for private. The final result achieved sub-meter accuracy in the median (50th percentile).
What are the key features of your approach?
The key point of my method is global optimization using graph optimization, unlike a conventional Kalman filter or least-squares-based positioning methods. In addition, highly accurate relative position estimation using the time difference of carrier wave phases of smartphones contributed to the accuracy. Because the competition dataset included environments such as tunnels and elevated structures in which GNSS cannot be received at all, I devised an algorithm with two optimization steps (first velocity optimization, then position optimization) and applied it to the competition. This method enables highly accurate position estimation for vehicle driving data in various GNSS signal reception environments using only smartphone GNSS observation data.
What end-user applications are you expecting your approach to enable?
Decimeter-accurate location estimation could lead to lane-level navigation for vehicles, pedestrian navigation, and advanced location-based smartphone games.
Dai
Shubin Dai, Kaggle Community
2nd Place Winner: Improving Smartphone GNSS positioning using Gradient Descent Method
What is your research focus and how does it relate to the contest?
I am a data scientist and one of the top competition grandmasters on Kaggle. My research interests include computer vision, natural language processing, autonomous driving, and reinforcement learning. I placed in the top three in 14 related competitions (13 of which were solo). So, despite my lack of background knowledge in the GNSS field, these methods, skills and experiences helped me find a solution.
How long have you been developing the technology or approach you used to win the contest?
I spent about 50 days on this competition, including learning principles of GNSS and understanding all kinds of algorithms by reading books, papers and source codes. The Kaggle platform is very helpful when we want to get started in a new field.
Have you participated in previous editions of this contest?
I did not participate in the competition held last year, but I learned a lot from solutions of recent years, particularly the third-place solution.
Where, in what GNSS signal conditions, and at what speeds were the test data collected?
The benchmark datasets include raw GNSS measurement and raw readings from inertial sensors, using smartphones (Xiaomi Mi 8, Google Pixel 4, etc.) enabled with dual-frequency and ADR (accumulated delta range) in driving scenarios, collected in the San Francisco Bay area.
In the GSDC2021 dataset, there are 29 drives with 73 phone GNSS logs in the training set and 19 drives with 48 phone logs in the test set. Compared to 2021’s competition, in the GSDC2022 dataset we can see more data overall and a wider variety of routes: 62 drives with 170 phone logs are provided in the training set and 36 drives with only one phone per drive are provided in the test set.
The drives in the training set took 15 to 60 minutes at an average speed of 18 m/s.
What accuracies were you able to obtain?
According to the metric of this competition, the score is calculated as the mean of the 50th and 95th percentile distance errors. The score on my local validation set is 1.929 m, the score on the public test set is 1.608 m, and the score on private test set is 1.499 m. When we calculate the mean error, the score is 1.401 m on a validated set, the mean error of 40% of the trips are under 1 m. I think the competition metric is more reliable as the 95th percentile distance error is also important.
By the way, my local validation set is more difficult to optimize than the test set, so the mean error on the test set is expected to be lower than 1.401 m.
What are the key features of your approach?
The competition data is noisy due to multipath effects, non-line-of-sight receptions, receiver noise and missing data, therefore it’s quite challenging. I found that the optimal estimation for each point locally is not stable and can be affected by noise at that point on the track. If we can find a solution to a whole track globally, the noise can be reduced as the model must follow all kinds of constraints, such as geometry constraints, speed constrains, and global acceleration constraints.
Although we could extend the WLS and Kalman filter solution to take more points on a track into consideration, it’s not so easy to model all kinds of constrains. On the other hand, if we use a global optimization method, such as factor graph optimization and neural networks, we can add the constrains easily, which makes it more efficient to conduct experiments.
Following the solution of the third-place winner in last year’s competition, I used the global optimization method by taking into account gradient descent, pseudorange, pseudorange rate, accumulated carrier phase (ADR), phone speed and acceleration constraints of every time epoch on a track. When optimizing the track using gradient descent, the losses are designed to filter out abnormal data and reduce the noise by a series of physical and geometrical rules. I spent much time searching for the constraints, proving them and turning them into losses that can be used to update the coordinates iteratively during the competition.
What end-user applications are you expecting your approach to enable?
According to the setting of this competition, we can post-process data collected using Android phones, which is easily obtained. The track obtained can then be optimized using the solutions from this competition. The solutions from the first and the second place can both be considered as a framework that can be extended by adding more constrains to it to improve accuracy.
Everett
Tim Everett, RTK Consultants LLC
3rd Place Winner: An RTKLIB Open-Source-Based Solution
What is your research focus and how does it relate to the contest?
I develop and maintain the demo5 fork of the popular RTKLIB open-source GPS/GNSS software tool. I have optimized this software for low-cost precision GNSS solutions, so it is very closely related to the goals of this competition. My background is in control system theory and I worked in product and technology development for servo systems in the disk drive industry for 25 years before switching to the GNSS field. The mathematics turns out to be quite similar between the two as both are problems in precision positioning, just different in scale. In disk drives, it is nanometers over centimeters and in precision GNSS, it is centimeters over kilometers.
How long have you been developing the technology or approach you used to win the contest?
I have been developing and maintaining low-cost precision GNSS solutions in the RTKLIB software for about six years but have only worked with smartphone solutions in the last year or two.
Have you participated in previous editions of this contest?
I did not participate in last year’s competition but I did work with the data after the contest was over and shared a solution using RTKLIB that would have placed fifth in the competition.
What accuracies were you able to obtain?
I achieved a score of 1.648 m on the private leaderboard. This represents the average of the 50th percentile and the 95th percentile of the errors as scored by Kaggle. Kaggle does not provide any further breakdown of this number but, based on the training data for which ground truths were provided, this corresponded to a 50th percentile error of roughly 0.9 m and a 95th percentile error of roughly 2.3 m. With a small tweak to my solution after the competition was over, I was able to improve my private leaderboard score to 1.593 m, which would have been within 1 cm of the third-place solution.
What are the key features of your approach?
My approach was to use the existing post-processing kinematic (PPK) solution algorithm in RTKLIB but to reoptimize it for the unique characteristics of the smartphone observation data. A PPK solution is the post-processing equivalent of a real-time kinematic (RTK) solution and is a differential solution that relies on differencing the receiver observations with observations from a nearby base station to cancel out most of the largest error sources — including atmospheric, orbital and clock errors — since these errors are common between the two sets of proximate observations.
Because smartphones have very poor GNSS antennas and they were mounted inside vehicles, the signal quality is much lower and the multipath much greater than those for which the RTKLIB algorithm was optimized. In addition, the smartphones were using the L5 frequency band, whereas RTKLIB was optimized for the more commonly used L2 frequency band. One of the main goals of my optimization process was to include many low-quality observations in the solution that would normally be discarded, but to de-weight them appropriately.
What end-user applications are you expecting your approach to enable?
RTKLIB software is currently used to provide precision solutions for many end-user applications such as surveying, drone photogrammetry, sports tracking, precision agriculture, utility location, marine navigation and ground subsistence monitoring. Although smartphones won’t replace dedicated low-cost GNSS receivers, the challenging nature of the smartphone data severely stresses the RTKLIB algorithms and exposes numerous opportunities for improvement that are much less obvious with more typical, higher quality data. I have pulled these improvements into the main branch of the demo5 version of RTKLIB, and hence this work should immediately improve the quality of all these applications and extend their use into more challenging environments.
Photo: Google
Acknowledgements: Thanks to the Institute of Navigation (ION) for co-sponsoring the 2022 Smartphone Decimeter Challenge. Thanks to Luke Walcher and Tolu Ojelade for their contributions to the photos used in this article.
Emlid has released Emlid Studio, a new post-processed kinematic (PPK) application designed specifically for post-processing GNSS data. The app is free and available for Windows and Mac users.
Emlid Studio features a simple interface to make post-processing easy. The app allows users to convert raw GNSS logs into RINEX, post-process static and kinematic data, geotag images from drones (including DJI brand), and extract points from survey projects completed with Emlid’s ReachView 3 app.
With Emlid Studio, users can post-process data recorded with Emlid Reach receivers and other GNSS receivers or NTRIP services. Post-processing requires RINEX observation and navigation files. Raw data in UBX and RTCM3 format also can be used — Emlid Studio will automatically convert them into RINEX.
The post-processing workflow is straightforward. Users can receive precise positioning of a single point or track depending on the positioning mode. Users can simply add several RINEX files and enter the antenna height, click the Process button, and Emlid Studio will do the rest. Once the resulting position file is ready, the plot will show the result.
Another tool is available for the users of Reach receivers and the ReachView 3 app. The Stop & Go feature allows users to improve the coordinates of points collected in single or float modes.
Geotagging for drone mapping. Adding geotags to images’ EXIF data requires aerial photos and the POS file with the events. Emlid Studio also provides a chance to update data from the RTK drone in case of a float or single solution during a survey. A set of RINEX logs from a base and drone, an MRK file and images from the drone are dragged and dropped into specific file slots, providing result in seconds.
Septentrio now offers Qinertia post-processing software from SBG Systems on AsteRx-i3 D Pro+, AsteRx-i3 S Pro+ and AsteRx SBi3 Pro+ receivers
Septentrio will now offer post-processing solutions for its GNSS/INS (inertial navigation system) receivers, using SBG Systems’ Qinertia software.
The AsteRx-i3 Pro+ receivers are fully compatible with Qinertia post-processing software, so no data manipulation is required before the post-processing step.
Land or aerial mapping applications, which do not have access to real-time GNSS corrections, benefit from post-processing software for higher positioning and orientation (heading, pitch and roll) accuracy. With the addition of post-processing, Septentrio GNSS/INS products cover the full mapping workflow.
“As a result of our cooperation with SBG Systems, Septentrio’s mapping customers who use GNSS/INS are benefiting from a quicker and more reliable workflow,” said Danilo Sabbatini, product manager at Septentrio. “The intuitive user interface of Qinertia software makes it easy for users to further improve their positioning and orientation accuracy in the post-processing step.”
In case of GNSS outage or correction link failure, post-processing recovers accuracy for recorded positioning and inertial data.
After the mission, Qinertia gives access to real-time kinematic (RTK) corrections from more than 8,000 base stations to deliver centimeter level accuracy. Trajectory and orientation are greatly improved by post processing GNSS and IMU data forward and backward. The Qinertia GNSS/INS post-processed kinematic (PPK) solution provides accuracy, reliability, advanced quality-control indicators, and a modern application programming interface (API).
Qinertia recently added an image geotagging feature, and specific outputs dedicated to photogrammetry.
Qinertia post-processing software will be used on Septentrio receivers. (Photo: SBG Systems)
Emlid is offering two positioning modules for mapping with unmanned aerial vehicles (UAVs). Both the Reach M+ and Reach M2 provide centimeter-level accuracy in real-time kinematic (RTK) and post-processed kinematic (PPK) modes, enabling precise UAV mapping with fewer ground control points.
The Reach M+ single-band receiver has a baseline up to 20 kilometers in PPK. The Reach M2 is a multi-band receiver with a baseline up to 100 kilometers in PPK.
Usually autopilot triggers the camera and records the coordinate it has at that moment. When the drone is flying at 20 m/s and GPS works at 5 Hz, the UAV autopilot will have position readings only every four minutes, which is not suitable for precise georeferencing. In addition, there is always a delay between the trigger and the actual moment the photo is taken.
Reach solves this problem by connecting directly to the camera’s hot-shoe port, which is synced with the shutter. The time and coordinates of each photo are logged with a resolution of less than a microsecond. Reach captures flash sync pulses with sub-microsecond resolution and stores them in a raw data RINEX log in the internal memory. This method allows ground control points to be used only to check accuracy.
The Reach RS2. (Photo: Emlid)
The Reach M2 PPK UAV mapping kit consist of the Reach M2 multi-band GNSS receiver onboard the aircraft that records the location of each photo at a frequency of 20 Hz. It is combined with the Reach RS2 GNSS multi-band receiver on the ground, drastically reducing the number of ground control points and simplifying the setup process on site, while maximizing the accuracy to centimeter levels even in remote areas.
The M2 tracks GPS/QZSS (L1, L2), GLONASS (L1, L2), BeiDou (B1, B2), Galileo (E1, E5) and SBAS (L1C/A), and receives a fixed solution almost instantly.
Rokubun has launched JASON, a satellite navigation service for accurate geolocation. With JASON, users can achieve highly accurate geolocation without a base station, Rokubun said.
JASON works under a “freemium” pricing model, making it possible to use it for free or to subscribe to monthly professional or premium plans.
JASON is a post-processed kinematic (PPK) satellite positioning service in the cloud that allows users to achieve highly accurate geolocation by uploading GPS/GNSS campaign data. JASON will try to obtain the best possible positioning solution on a best-effort basis. It will use PPK with the nearest GNSS continuously operating reference station (CORS) in the service’s database or with he user’s own provided station if no close CORS are available, or precise point positioning. The data can be processed interactively by using the Rokubun website or automating the workflow via the Rokubun API.
JASON also features a free data-conversion service for GNSS raw measurements.The service is compatible with Argonaut, u-blox, Septentrio, Android GnssLogger and any receiver able to export industry standard RINEX v2 or v3 file formats.
According to Rokubun’s CEO Xavier Banqué-Casanovas, JASON cloud service allows users to get the best possible performance out of their GNSS equipment using an internet browser, without the need for special installation or updates requirements.
Qinertia, SBG Systems’ PPK software, now supports third-party IMUs and offers a GNSS post-processing license covering all major GNSS receivers
Screenshot: SBG Systems
SBG Systems’ INS/GNSS post-processing kinematic (PPK) software Qinertia now covers all surveyors’ projects by offering a license dedicated to GNSS post-processing. Open to the world, Qinertia supports all major GNSS receivers and is now open to third-party inertial measurement units (IMUs).
Qinertia has been designed to offer a comprehensive suite of post-processing software to geospatial professionals. It accepts all major GNSS manufacturers, and supports proprietary protocols from NovAtel, Septentrio, Trimble and u-blox for a straight-forward workflow.
The full-featured post-processing software offers native support for u-blox F9 real-time kinematic (RTK) receivers, reducing the workflow to a simple “drag and drop” to guarantee data integrity and accuracy.
Qinertia has been designed to help surveyors get the most of their surveys easily with a simple workflow, powerful quality control tools and tightly coupled algorithms. All of this is available to any surveyor with the new support of third-party IMUs or GNSS receivers. Several IMUs and inertial navigation systems (INS) have already been successfully integrated with Qinertia including Northrop Grumman’s LN-200 and LCI-100 and the Inertial Sense µIMU.
The new Qinertia GNSS license allows surveyors to post-process both static and kinematic GNSS data. In just a few clicks, surveyors can improve their trajectories, access RTK corrections worldwide, or even control a base-station’s precise location using precise point positioning (PPP) static computations.
GIS and Photogrammetry. Whether they fly a UAV or drive a car, professionals can improve their image location accuracy. Qinertia has been designed to help surveyors get their GIS or photogrammetry projects way more precise, by exporting a centimetric position for each picture at the exact shutter event.
Septentrio’s post-processing kinematic (PPK) software has been upgraded with multi-GNSS and BaseFinder functionality. BaseFinder improves project efficiency by automatically finding the most suitable reference station data needed for centimeter-level accuracy.
Both GeoTagZ and PP-SDK now feature BaseFinder, which speeds up survey workflow by automatically finding reference data needed for augmenting GNSS logs with sub-centimeter accuracy. BaseFinder accesses an online database of reference networks and extracts the most suitable corrections available. BaseFinder is available via an app or via an API and can be incorporated into any existing software.
PPK is often used for ground surveys with aerial drones, allowing high precision georeferencing without the need for a real-time base station link or ground control points (GCPs).
“Surveying without a base station will allow users to reduce costs and set-up time. With this PPK upgrade we are improving the end-user experience as well as developer experience,” said Danilo Sabbatini, product manager at Septentrio.
The new release of this GNSS post-processing software also includes two additional GNSS constellations: European Galileo and Chinese BeiDou. Having access to all the signals from all GNSS constellations improves reference network compatibility. It also improves positioning availability in difficult environments. This is particularly important when working in areas of low satellite visibility such as near tall structures or under foliage.
When doing photogrammetry with a drone, GNSS data is often recorded and then post-processed together with base station data to achieve sub-centimeter positioning accuracy. This base station data can be obtained either with proprietary base stations or by using base station data from a public reference network (see diagram below). Septentrio receivers are designed to bring accurate and reliable positioning to photogrammetry, aerial inspection, marine survey as well as mobile mapping.
GPS Post-Processing SDK architecture, bringing high-accuracy positioning without the need for a real-time correction stream. (Diagram: Septentrio)
Wingtra has officially launched the WingtraOne PPK high-precision mapping drone. Wingtra said its drone, which features vertical take-off and landing, is designed to set a new benchmark for large-scale surveying and mapping applications.
WingtraOne PPK offers large area coverage, ultra-high accuracy and brilliant image resolution. It features an advanced PPK module and high-quality cameras like the 42-megapixel full-frame camera Sony RX1RII, it is now possible to reach down to 1-centimeter absolute accuracy in aerial mapping.
To prove this accuracy claim, the Wingtra team performed test flights in a gravel quarry. The process was documented and is now explained in a white paper on the company website.
Conventional drone mapping on centimeter accuracy requires ground control points (GCPs) to correct the final map. Besides requiring additional surveying equipment and being extremely time consuming, setting up GCPs might be downright risky or just not possible in the area of interest.
More advanced solutions achieve similar levels of accuracy by using GPS correction technology for the georeferencing of the aerial imagery: namely RTK (real-time kinematics) or PPK (post processed kinematics).
RTK requires real-time base station connectivity and corrects GPS signals during the flight, while PPK corrects them after the flight and therefore offers greater robustness and consistency.
Moreover, PPK is independent from base stations or base station networks. It is highly reliable, accurate and time saving to use, Wingtra said. Neither special flight preparations nor intensive post-processing steps are required to achieve down to 1-cm accurate aerial maps.
Belgian GNSS receiver manufacturer Septentrio was selected by Swiss drone manufacturer Wingtra to supply GNSS OEM receiver boards and PPK processing software for the recently-launched WingtraOne PPK drone.
The combination of vertical take-off and landing (VTOL) technology and a high-spec post-process kinematics (PPK) brings wide-area coverage at ultra-high precision.
Following a flight, the GNSS data of the WingtraOne is processed offline using Septentrio’s PPK software. This combines the drone data with correction data from a nearby reference receiver to get accurate cm-level geolocations for every photograph.
The on-board high-resolution Sony RX1RII camera, AsteRx-m2 UAS receiver board combined with Septentrio’s PPK library, and Pix4D photogrammetry processing software are together able to produce ground precisions of 1.3 centimeter (cm) horizontal and 2.3 cm vertical.
Image: Wingtra
“With the WingtraOne PPK, we can offer a world first in drone photogrammetry — wide coverage at ultra-high precision,” said Armin Ambühl, CTO of Wingtra. “In a single one-hour flight, the WingtraOne can cover 130 ha (320 acres) delivering mapping with GSDs [ground sample distance] below 1 cm/pixel with absolute accuracy down to 1.27 cm.”
He continued, “WingtraOne’s advantage is twofold: it combines VTOL with the latest PPK technology from Septentrio. With VTOL we can offer the best of both worlds: multirotors and fixed-wings. Vertical take-off and landing means hands-free operation and a smoother ride for the on-board camera payload. Secondly, efficient flying in fixed-wing mode means far greater coverage than any comparable multirotor.”
“We are proud and excited to be part of this innovative project with Wingtra pushing the boundaries of aerial photogrammetry,” said Gustavo Lopez, product manager at Septentrio. “The WingtraOne incorporates our AsteRx-m2 UAS OEM board and, photogrammetry applications requiring high-precision, low-latency positioning are what it does best. The board is specifically designed for quick and easy integration and, with Septentrio’s world-first, multi-frequency PPK, cm-level precision can now reach the parts dual-constellation solutions feared to tread.”
