基于实际意义误差函数的机器人避碰模糊神经网络算法

Fuzzy Neural Network Algorithm for Robot to Evade Obstacles Based on Error Function with Actual Significance

介绍一种机器人在不确定环境下,通过探障传感器,在探测过程中采用模糊神经网络躲避障碍物的方法。该模糊控制规则采用神经网络,来实现以避免控制规则的死点从而完成避碰路径规划,计算机仿真实验结果验证了该法的有效和实用性。

The paper introduces a kind of approach for robot to evade obstacles with the fuzzyneural network by probe sensor under uncertain surroundings. The neural network isadopted to avoid the blocking zone of the fuzzy control rules so as to make out pathplanning for evading obstacles. Computer simulation test result proves the availabilityand practicability of the method.