Cooperative Path Planning for Single Leader Using Q-learning Method

Cooperative navigation is a popular method in Autonomous underwater vehicles (AUVs) navigation in recent years. In this paper, the Q-learning method is applied to the cooperative navigation system to solve the path planning problem of the leader AUV, which can make the observation error of the system to be minimum. On the other hand, the algorithm proposed in this paper focuses on the time consumption of the algorithm, the purpose is to complete the path planning calculation in a shorter time. First, the path planning problem is proposed based on the error model of AUV cooperative navigation. Then the path planning problem is modeled by Markov decision framework (MDP). Finally, the Q-learning method is used to plan the path of the leader AUV. The simulation results show that the path generated by the Q-learning method can minimize the navigation error of the AUV during the entire cooperative navigation process. Compareing with other path planning methods, the Q-learning-based path planning algorithm has the less time complexity.