Complete Coverage Path Planning of an Unmanned Surface Vehicle Based on a Complete Coverage Neural Network Algorithm

In practical applications, an unmanned surface vehicle (USV) generally employs a task of complete coverage path planning for exploration in a target area of interest. The biological inspired neural network (BINN) algorithm has been extensively employed in path planning of mobile robots, recently. In this paper, a complete coverage neural network (CCNN) algorithm for the path planning of a USV is proposed for the first time. By simplifying the calculation process of the neural activity, the CCNN algorithm can significantly reduce calculation time. To improve coverage efficiency and make the path more regular, the optimal next position decision formula combined with the covering direction term is established. The CCNN algorithm has increased moving directions of the path in grid maps, which in turn has further reduced turning-angles and makes the path smoother. Besides, an improved A* algorithm that can effectively decrease path turns is presented to escape the deadlock. Simulations are carried out in different environments in this work. The results show that the coverage path generated by the CCNN algorithm has less turning-angle accumulation, deadlocks, and calculation time. In addition, the CCNN algorithm is capable to maintain the covering direction and adapt to complex environments, while effectively escapes deadlocks. It is applicable for USVs to perform multiple engineering missions.

Automated summary

The CCNN algorithm simplifies the neural activity calculation process, reducing calculation time and increasing path movement directions in grid maps to further reduce turning angles and make the path smoother. Additionally, an improved A* algorithm effectively decreases path turns to escape deadlocks.