UAV Implementation of Distributed Robust Target Location in Unknown Environments

This paper presents the implementation of a target seeking application on a multi-agent unmanned aerial vehicle (UAV) testbed, in which the target to be located emits a signal that attenuates with increasing distance from its source. A heterogeneous swarm of multirotors collects local measurements of the signal emanating from the target, and each UAV shares its information with neighboring UAVs in a fully distributed manner to form an estimate of the signal's gradient, informing the direction of travel to move toward the target. The UAVs navigate around obstacles in an unknown environment, implementing a localized robust hybrid controller for obstacle avoidance when an obstacle is within the UAV's detection radius. The hybrid controller enables the UAVs to robustly avoid obstacles in the presence of potentially adversarial exogenous inputs through the use of overlapping control modes which preclude the topological obstructions that can arise in traditional obstacle avoidance schemes. The algorithms are implemented on the VT SpaceDrones platform, using mulitrotors equipped with Raspberry Pi microcontrollers. Each UAV performs its own target seeking and obstacle avoidance calculations in a distributed architecture, receiving position data from an OptiTrack motion capture system. The distributed implementation illustrates the adaptation of the control law to real world challenges, including unsynchronized clocks among different UAVs, limited computational power, and communication latency. Experimental and theoretical results are compared.