Multi UAV Cooperative Reconnaissance Based on Dynamic Programming IDQN Algorithm

In this paper, multiple UAV cooperative reconnaissance tasks in complex environments are realized based on deep reinforcement learning. With the goal of quickly completing reconnaissance task allocation, safely avoiding obstacles and finally completing flight tasks, task allocation and trajectory planning are tightly coupled, and a multiple UAV cooperative reconnaissance algorithm based on dynamic programming multi-agent independent deep Q networks (IDQN) is proposed, Implemented collaborative reconnaissance task allocation and trajectory planning flight decision-making for multiple unmanned aerial vehicles in complex environments, and completed collaborative reconnaissance tasks for multiple unmanned aerial vehicles.