Building UAV-Based Testbeds for Autonomous Mobility and Beamforming Experimentation

Abstract

Unmanned aerial vehicles have been deployed in many applications such as search and rescue, reconnaissance, and disaster recovery. However, UAV mobility can threaten the ability to maintain robust transmissions in practical deployments. On one hand, advanced software methodologies and extensive experiments are required to ensure safe and autonomous flights. On the other hand, to unlock additional capacity in drone communications, additional techniques must be leveraged such as directionality via MIMO-based beamforming, requiring accurate channel information to be fed back in- flight. Software defined radio (SDR) platforms play a major role in filling these gaps in multiple frequency bands, customizable design, and performance characterization. In this work, we present the hardware setup as well as software architecture of our proposed testbed leveraged for two different applications: autonomous mobility and beamforming. In the autonomous mobility case, we build a robust UAV-control framework on a customizable drone platform using MAVLink. Our experiments have demonstrated the feasibility of intelligent automated flight patterns. In the beamforming case, we implement a beamforming scheme on a drone-based SDR platform and evaluate its performance in various contexts. Our evaluations reveal that the drone-based beamforming can improve throughput significantly over conventional schemes.