Control of Multi-Agent Collaborative Fixed-Wing UASs in Unstructured Environment

Abstract

In recent years, the study of dynamics and control of swarming robots and aircraft has been an active research topic. Many multi-agent collaborative control algorithms have been developed and have been validated in simulations, however the technological and logistic complexity involved in validation of these algorithms in actual flight tests has been a major hurdle impeding more frequent and wider applications. This work presents robust navigation algorithms for multi-agent fixed-wing aircraft based on an adaptive moving mesh partial differential equations controlled by the free energy heat flow equation. Guidance, navigation, and control algorithms for control of multi-agent unmanned aerial system (UASs) were validated through actual flight tests, and the robustness of these algorithms were also investigated using different aircraft platforms. The verification and validation flight tests were conducted using two different fixed-wing platforms: A DG808 sail-plane with a 4m wingspan T-tail configuration and a Skyhunter aircraft utilizing a 2.4m wingspan and a twin-boom configuration. The developed swarm navigation algorithm uses a virtual leader guidance scheme and has been implemented and optimized using optimal control theory. Multi-scale moving point guidance has been developed and complimented by a linear quadratic regulator controller. Several flight tests have been successfully conducted and a system of systems including software and hardware was successfully validated and verified.