Multi-UAV Autonomous SDR Coordination Experiment on AERPAW

Kevin Dee

Research Scientist, University of California (Irvine)

As a research experimenter developing distributed autonomous UAV coordination algorithms for wireless sensing applications, Kevin Dee wanted to deploy and validate a real-time multi-drone experiment on the AERPAW platform that integrated autonomous trajectory generation, SDR-based communication measurements, and coordinated vehicle control across multiple UAVs.

The experiment involved two UAV-mounted portable nodes equipped with SDRs, where one drone acted as a transmitter operating at 3.32 GHz and the second acted as a receiver logging metrics such as SNR, signal power, and communication quality. A fixed node coordinated the experiment by issuing waypoint and “goto” commands over the cellular IP network to both drones. The UAVs autonomously synchronized their movements while maintaining collision avoidance and geofencing constraints.

To support deployment, Kevin first developed and validated the guidance and coordination algorithms in MATLAB simulations. Using MATLAB Coder, he converted the algorithms into deployable code for the AERPAW environment and iteratively refined the experiment through emulation and field testing, incorporating safety and operational feedback from the AERPAW team. This included:

  • Adjusting UAV separation distances during takeoff and return phases
  • Increasing the collision exclusion radius from 6m to 8m
  • Scaling the experiment geometry to improve safety margins
  • Verifying anti-collision services in the OEO console
  • Monitoring timing between autonomous waypoint commands
  • Comparing simulated trajectories against logged UAV GPS trajectories for validation

The experiment successfully demonstrated coordinated multi-drone autonomous behavior in the AERPAW testbed. During field execution, the UAVs dynamically synchronized their flight behavior, including coordinated takeoff, transit, mission execution, and landing, despite differences in vehicle cruise speeds. Logged field trajectories closely matched the MATLAB simulation outputs, validating both the guidance algorithms and the AERPAW multi-vehicle coordination framework.

This work represented one of the first externally developed multi-drone experiments executed successfully on the AERPAW platform without direct implementation assistance from the AERPAW team, serving as a strong validation of the platform’s documentation, coordination services, and autonomous experimentation capabilities.

Collaborators and advisors included Mohammadreza Barzegaran, Hamid Jafarkhani, and Lee Swindlehurst under NSF CPS Grant CNS-2209695.