基于网络结构的并行路径规划算法

算法继承了人工势场法的基本思想，通过寻找路径点的能量函数的极小值点而使路径避开障碍物。势场由排斥场和吸引场叠加而成，在算法中对于排斥场和吸引场的强度引入了一个平衡系数，并引入了模拟退火的思想和一些启发性知识，以避免某些局部极值的情况。算法具有很大的并行性，收敛速度较快，易于从二维空间扩展到三维空间，对人工势场法给予了较大的改进，取得了较好的仿真效果。

Algorithm based on massively parallel connectionist network for path planning

The algorithm given in this paper inherits the basic idea of potential energy field method. By searching the minimum value of the energy function,the collision free path can be found. Because a balance coefficient between the reject potential energy field and the attract potential energy field is introduced in the algorithm, and simulated annealing and some heuristic method are used in the algorithm, some local minimum value can be avoided. Compared with the method only using potential energy field, the parallel degree of this algorithm is higher,the convergence velocity of the algorithm is quicker,and it is easier to develop the algorithm from two dimensions to three demensions. The simulation shows good results.