GPS World

Tag: swarm autonomy

  • Baykar demos swarm UAVs without GNSS

    Baykar demos swarm UAVs without GNSS

    Turkish UAV maker Baykar demonstrated its next-generation Kamikaze UAV K2 and Sivrisinek (Mosquito) loitering munition, showcasing AI-supported swarm autonomy, GNSS-independent navigation, automatic target detection, and strike capabilities during a demonstration held at the Keşan Flight Training and Test Center.

    The K2 Kamikaze UAV and the Sivrisinek loitering munition will make their public debut at SAHA 2026, which takes place in Istanbul May 5-9.

    The April 17 demonstration opened with the sequential takeoffs of five K2 Kamikaze UAVs within five minutes. Once airborne, the platforms conducted patrol flights in “right echelon,” “line,” “V,” and “Turan” formations.

    Ten Sivrisinek loitering munitions — a new platform developed by Baykar — then joined the operation, forming a swarm beneath the K2 Kamikaze UAVs. The Bayraktar TB2, TB3, and AKINCI UCAVs accompanied the swarm flight, recording the operation from the air.


    Credit: Baykar


    AI-supported visual navigation
    Among the key technical highlights of the demonstration were the solutions developed to counter electronic warfare environments. Using AI-supported visual navigation software, the platforms demonstrated the capability to perform positioning and navigation independently of GNSS.

    Having successfully showcased autonomous navigation in a GNSS-denied environment, the K2 and Sivrisinek Kamikaze UAVs also demonstrated AI-supported automatic target detection and automatic strike capabilities.

    As part of the demonstration, a fleet of Sivrisinek loitering munitions executed a dive on the designated coordinates. A K2 Kamikaze UAV then broke off from the swarm and performed a high-speed dive on the designated coordinates, conducting a pass. In the final phase of the demonstration, a swarm group composed of 18 unmanned aerial vehicles across different classes — 5 K2s, 10 Sivrisinek, 1 Bayraktar TB2, 1 TB3, and 1 AKINCI — came together in a “V” formation to salute the delegation observing the flight.

    Developed by Baykar, the next-generation Sivrisinek loitering munition raises operational depth to a range exceeding 1,000 kilometers. Capable of uninterrupted communication within the swarm through AI support, Sivrisinek platforms can instantly share detected targets with one another.

    Performing its missions through AI-based visual positioning even in the most challenging environments — including areas where GNSS signals are unavailable or subject to intensive jamming — Sivrisinek stands out in strategic missions to be conducted on the battlefield thanks to its high autonomy capability.

  • Raytheon swarm enables one operator to guide 100+ drones

    Raytheon swarm enables one operator to guide 100+ drones

    Raytheon BBN-led team recently supported DARPA's fifth OFFSET program field exercise. (Photo: Raytheon)
    Raytheon BBN-led team recently supported DARPA’s fifth OFFSET program field exercise. (Photo: Raytheon)

    Raytheon Intelligence & Space, a Raytheon Technologies business, recently supported the fifth OFFensive Swarm-Enabled Tactics (OFFSET) DARPA program field exercise.

    Using integrated swarm technology developed by a Raytheon BBN-led team, a single operator successfully controlled a swarm — composed of 130 physical drone platforms and 30 simulated drone platforms — both indoors and outdoors in an urban setting. Raytheon BBN provides advanced technology research and development with a focus on national security priorities.

    During the exercise, the team used a combination of commercial off-the-shelf and custom-built hardware and software to deliver swarm autonomy. This enabled a single or small group of operators to direct and manage the activities of a large swarm of autonomous air and ground vehicles with minimal training.

    “Controlling a drone swarm changes the way an operator or group of operators think about the drones,” said Shane Clark, Raytheon BBN OFFSET principal investigator. “Takeaways from this exercise help inform us of the inflection points between utility and manageability.”

    Inexpensive Hardware

    A key element of the program is the use of inexpensive hardware. Without the powerful computing and sensing capabilities available in larger more expensive platforms, Raytheon BBN needed to create a broad library of simple tactic building blocks used to create plans to accomplish mission objectives.

    Raytheon BBN also designed and configured a scalable, modular and decentralized approach to manage a variety of current and future platforms and missions. Whenever possible, the drones collaborate actively to decide how to accomplish a specific mission most efficiently.

    “Our software is smart enough to assign drones with the right capabilities to the appropriate set of tasks,” Clark said. “For example, if the task is to surveil a building, multiple drones will be dispatched, with each surveilling portions of the building. The software considers each platform’s sensor capabilities, and tasks drones with downward-facing cameras to surveil the roof.”

    Once the drones are deployed, their collaboration allows them to understand what parts of a building have been explored and where the gaps are. They then autonomously select how to fill in those gaps.

    Virtual Reality Interface

    To tackle the complexities of human swarm interfaces, the Raytheon BBN team created a virtual reality interface in addition to traditional camera views. It takes feeds from all the swarm assets to create an interactive virtual view of the environment.

    “You can look behind the building to access a view of drone locations, for example, and use the virtual reality environment to test and see if your mission is viable,” Clark said. “We also developed a speech interface with the operationally deployed Tactical Assault Kit, or TAK, integration capability that enables the operator to act quickly while maintaining situational awareness over many systems simultaneously.”

    The Raytheon BBN-led team includes Smart Information Flow Technologies, or SIFT, and Oregon State University. The team is contracted by DARPA to demonstrate its swarm capabilities during Army Expeditionary Warrior Experiment 2022 taking place in February to March, and hosted by the Army Maneuver Battle Lab.