本文介紹了基于神經網絡和微粒群優化算法的移動機器人動態避障路徑規劃算法。通過神經網絡改進的微粒群算法，充分利用了神經網絡的融合性和并行性來提高微粒群算法中適應度函數的準確性。通過神經網絡描述機器人工作空間的動態環境約束并找到最優的適應度函數，在微粒群算法中使用該函數，求得微粒群算法最優無碰路徑。
關鍵詞：神經網絡；微粒群優化算法；動態避障；路徑規劃
Abstract: This paper presents a mobile robot obstacle avoidance dynamic path planning
algorithm based on neural network and Particle Swarm Optimization (PSO). Using neural network to improve the PSO algorithm, and make integration and parallel of neural network to enhance the accuracy of PSO algorithm’s fitness function. Optimization function for the extreme, points to determine the equations of motion, and finally to the point iterative path tends to set the optimal path planning, use this function in PSO algorithm, finally get PSO obtained optimal collision path.
Key words: neural network; particle swarm optimization; obstacle avoidance dynamic; path
planning