基于改进的生物激励神经网络机器人路径规划算法

Robot path planning algorithm based on improved biological-inspired neural network

摘要:

针对传统生物激励神经网络（BINN）在点对点全局路径规划中存在的路径偏离和路径非最优问题,提出了基于路径修正和无障碍理想路径制导的生物激励神经网络算法。路径规划的初始阶段，通过判断起始单元外部激励输入和起始单元活性值大小决定是否触发路径生成策略，从而实现初始路径修正；在生成下一位置单元的算法中结合无障碍理想路径的导向，引入实际路径单元与无障碍理想路径单元间的理想路径接近率使路径神经元活性值增大，从而实现路径优化。在静态复杂环境下，分别以三种算法进行了对比实验。实验结果表明，改进后的路径规划算法相比传统生物激励神经网络算法和基于目标制导的生物激励算法，不仅解决了路径规划初始阶段的路径偏离问题，而且使路径长度和路径转折次数更低，效率更高。

Abstract:

In order to solve the problem of path deviation and non-optimal path in the traditional biological-inspired neural network (BINN) in point-to-point global path planning, BINN based on path modification and ideal barrier-free path guidance was proposed. In the initial stage of path planning, determine whether to trigger the path generation strategy by judging the external stimulus input of the starting unit and the activation value of the starting unit, so as to achieve the initial path modification; combine the guidance of the ideal path without obstacles in the algorithm for generating the next position unit The introduction of the ideal path approach rate between the actual path unit and the barrier-free ideal path unit increases the path neuron activity value, thereby achieving the purpose of optimizing the path. In a static and complex environment, three experiments were carried out for comparison experiments. The experimental results show that the improved path planning algorithm not only solves the problem of path deviation in the initial stage of path planning, but also makes the path length and number of path turns lower than traditional biological inspired neural network algorithms and target-guided biological inspired algorithms. higher efficiency.