Path Planning for Swarm AUV Visiting Communication Node

This paper proposes a method for path planning of an underwater robot swarm. The method is based on biological inspired neural network to plan path between robots and communication nodes. The robot swarm is used to search wild sea area. To solve the long distance communication problem, we deploy some communication nodes ahead, forming a communication network under the water. The robots visit the nodes to communicate. With this method, robots can also avoid obstacles in real time. Firstly, put the landscape into grid map. Then build biologically inspired neural network based on the grid map. The node attracts the robots and the obstacles reject the robots through neural activity. At last, robots plan their path by the activity with a steepest gradient descent rule. Simulation result shows the method may lose in local optimum, so we improve the method to avoid repetitive path. The results show that the improvement effective for path planning.