关键词: 生物激励神经网络, 变电站巡检, 路径规划, 任务分配

Abstract: In order to improve the efficiency of patrol inspection of variable power station, a patrol inspection method based on improved biologically inspired neural network was proposed considering the existence of obstacles. By improving the excitation input and path decision of the biologically inspired neural network, the method avoided the path misjudgment problem and the problem of many turning times caused by the algorithm's dependence on the distribution of neuron active values when making path decisions. When it was applied to the routing path planning of multiple substations, multi-point cooperative routing and area coverage routing of variable power stations could be realized.The simulation results show that, compared with the biologically inspired neural network(BINN) algorithm, the improved BINN algorithm has obvious advantages in the path planning of the variable power station, the path length was reduced by 50.22 m, the rotation angle was reduced by 11.65°, and the number of turns was reduced by 5 times, 10.11 m, 40.23°, and 15 times respectively. Compared with the conventional path planning algorithm PRM algorithm and Dijkstra algorithm, the improved BINN algorithm still has obvious advantages in various indicators such as path length, rotation angle and number of turns, which can improve the work efficiency of patrol inspection of variable power plants.

Key words: biologically inspired neural network(BINN), substation patrol inspection, path planning, task allocation