Multi-Robot One-Target 3D Path Planning Based on Improved Bioinspired Neural Network

Abstract

3D Path planning for multi-robot one-target pursuit is an interesting topic. Bioinspired neural network is frequently implemented for path planning of multi-robot, and the bioinspired neural network neural activity value computational cost and time cost will increase sharply with the increase of the number of neurons. This paper explores an improved 3D path planning method based on multi-scale map method to reduce the time cost. Combining the multi-scale map method with the Dijkstra algorithm, the optimal paths of 3D coarse-scale map of multi-robot one-target can be generated. The weights created from the coarse-scale map are employed to yield the 3D path planning of the fine-scale map for the same terrain. Therefore, this improved bioinspired neural network algorithm has proven the ability to calculate the multi-robot 3D optimal paths. By introducing this multi-scale map method into the multi-robot bioinspired neural network algorithm, the time cost and mathematical complexity of the path planning algorithm can be greatly reduced. MATLAB simulation results further reveal the effectiveness and superiority of this method.