Enhancing Exploration Efficiency of Fixed-Wing UAVs Through Intelligent Decision-Making and Advanced Control Integration

With the ongoing development of intelligent technologies, fixed-wing unmanned aerial vehicles (UAVs) are gaining prominence across diverse sectors. This study focuses on complex environment exploration. It employs the Dyna-Q reinforcement learning algorithm to derive optimal waypoints in obstacle-rich settings. These waypoints are pivotal for task allocation, ensuring judicious deployment of UAV clusters for targeted exploration. Concurrently, B-spline techniques refine these waypoints into feasible flight trajectories for UAVs. Subsequently, distributed model predictive control (DMPC) is utilized to maintain a designated cluster configuration. This DMPC controller ensures trajectory tracking along optimized paths, facilitating thorough area exploration. Then the approach is validated and evaluated through simulation experiments using a laboratory-developed simulation platform.