生物激励神经网络路径规划仿真研究与改进

生物激励神经网络移动机器人路径规划方法是一种新颖的方法,可用于在动态不确定环境下生成实时的避障轨迹.本文的仿真结果表明当该方法被应用于点对点路径规划时,生成路径可能不满足路径长度要尽可能短的约束条件;当该方法被应用于全覆盖路径规划时,生成路径可能不满足覆盖过程应有规律和重复覆盖应尽可能少的约束条件.本文对上述出现的不合理现象进行了理论分析并分别提出了在点对点路径规划中引进目标制导和在全覆盖路径规划中引进规则制导的改进方法.仿真结果表明改进方法是有效的.

Simulation Research and Improvement on Biologically Inspired Neural Network Path Planning

Biologically inspired neural network approach of mobile robot path planning is an original approach, which can be applied to generate real-time collision-free trajectory under dynamic uncertain environment. The simulation in this paper shows that the generated path may not accord with the restriction that the length of the path should be short to the greatest extent when the approach is applied in point to point path planning; the generated path may not accord with the restriction that the coverage process should be of regularity and the repetitious should be little to the greatest extent when the approach is applied in the complete coverage path planning. The analysis of the above unideal phenom- ena is made in the paper. And the improving methods of introducing goal navigation in point to point path planning and introducing rule navigation are individually proposed in the paper. The simulation results show that the new approaches are valid.