Research On Path Planning Method Of Multi-UAV Cooperative Detection

Task assignment and trajectory planning,as the key technologies to guide multi-UAV cooperation,have always been a focus on Unmanned Aerial Vehicle(UAV).However,the current research on coverage path planning(CPP)of multi-UVA cooperative detection is mostly confined to static planning and lacks the dynamic adaptability.Moreover,how to delimit molecular region according to UAV’s performance is rarely considered.To address above issues,this paper focus on studying the region division and CPP methods for MultiUAV cooperative detection.Aiming at dividing areas according to the performance of UAV,a region division method is proposed,which is improved by the method of dividing areas based on robots’ initial position(DARP).By modeling the performance of UAV in the proposed method,the region partition is effectively restricted.The simulation results show that the proposed region division method can reasonably allocate the detection region in a relatively short time.To solve the problem of CPP in dynamic task scenarios,the method based on spanning tree coverage(STC)is introduced for static planning and the method based on Q-learning is used for dynamic planning.Moreover,an improved STC based path planning method is proposed to deal with the diversity problem of the number of path turns.By combining the ant colony algorithm,the number of path turns in the proposed method is optimized,resulting in covering the region quickly.Aiming at increasing the coverage rate of Qlearning,an improved Q-learning method is proposed for path planning.By adjusting the training criteria and selection strategy of Q-learning algorithm,the complete coverage of areas in the proposed method is successfully achieved.The simulations show that the proposed path planning method can improve the UAV’s adaptability to dynamic tasks,and optimize the number of turns for static tasks.Furthermore,a multi-UAV cooperative detection trajectory planning test verification system based on the proposed path planning method is implemented.This system extends the multi-UAV cooperative detection path planning from theoretical research to practical application,which is meaningful.