Joint Task Allocation and Path Planning for Multi-UAV SAR Based on QDE

In practice, classical path planning algorithms need further improvement to perform better when applied to an unknown environment. In order to improve the accuracy to avoid a moving object in dynamic environment for unmanned aerial vehicle (UAV), this paper proposes a Q-learning combine Differential Evolution algorithm (QDE) method for task allocation and path planning in a multi-UAV Synthetic Aperture Radar (SAR) system. The paper first formulates the task allocation problem as a constrained single-objective optimization problem and solves it using a Differential Evolution (DE) algorithm. Subsequently, a reinforcement learning-based path planning method is proposed, which generates optimal flight paths for UAVs by constructing reward and cost matrices. Simulation results show that the method effectively avoids mountain obstacles and threats from other UAVs, successfully completing imaging tasks.