Task allocation and path planning problems of multi-AUV system based on auction-dynamic neural network

How to carry out reasonable task allocation and path planning in the real environment of the seabed are the two major problems of the multi-AUV system. In this paper, the task preprocessing of the 26 set targets is obtained by using the K-means method to obtain three cluster center points. The task allocation process of the system is solved using a distributed auction algorithm, and according to the results of the allocation, the dynamic neural network (DNN) algorithm is used for path planning. The proposed algorithm can be applied to the real environment where ocean currents and obstacles exist. Finally, in the simulation part, the results under the particle swarm (PSO) algorithm are compared, to verify the reliability of the algorithm.