基于改进神经网络的多AUV全覆盖路径规划

针对三维环境下的多自主水下机器人(Autonomous Underwater Vehicle,AUV)全覆盖路径规划问题,提出一种基于改进神经网络—Glasius生物启发神经网络（Glasius Bio-inspired Neural Network,GBNN）的全覆盖路径规划算法。对AUV的水下工作环境构建离散的三维栅格地图;根据栅格地图,建立相对应的三维GBNN模型;根据GBNN活性值的动态变化,AUV规划各自的搜索路径,对水下任务区域进行全覆盖搜索。仿真结果表示,多AUV可以协同完成覆盖搜索任务,能够自动避开各类静态和动态的障碍物,自动逃离路径的死锁区。

Multi-AUV Complete Coverage Path Planning Based on Improved Neural Network

Aiming at the working space search task of multiple AUVs (Autonomous Underwater Vehicle) in 3-dimensional underwater environments, a complete coverage path planning algorithm based on an improved neural network-Glasius Bio-inspired Neural Network (GBNN) is presented in this paper. A discrete 3-D grid map of the underwater environment is constructed. A 3-D GBNN model is established topologically according to the map. Based on the dynamic activities of GBNN model, each AUV plans its own coverage path independently, and covers the whole working space collaboratively. The simulation results show that the multiple AUVs can collaboratively cover the working space completely, automatically avoid the obstacle and escape from the deadlock in the path.