Algorithm for intelligent collaborative target search and trajectory planning of MAV/UAV

Abstract

Based on the manned aerial vehicle (MAV) / unmanned aerial vehicle (UAV) intelligent cooperation platform, the search of multiple interfered signal sources with unknown locations and trajectory planning were studied.Considering the real-time and dynamic nature of the search process, a MAV/UAV intelligent collaborative target search and trajectory planning (MUICTSTP) algorithm based on multi-agent deep reinforcement learning (MADRL) was proposed.Each UAV made online decision on trajectory planning by sensing the received interference signal strength (RISS) values, and then transmitted the sensing information and decision-making actions to the MAV to obtain the global evaluation.The simulation results show that the proposed algorithm exhibits better performance in long-term RISS, collision, and other aspects compared to other algorithms, and the learning strategy is better.

Keywords

• MAV/UAV;intelligent collaborative;MADRL;trajectory planning;RISS