一种改进的模糊CMAC神经网络车辆自主避障算法  

A New Autonomous Obstacle-avoiding Algorithm Based on the Improved Fuzzy CMAC Neural Network

作　　者：王冠[1] 齐龙妹 

机构地区：[1]西北工业大学自动化学院,陕西西安710129

出　　处：《机械与电子》2014年第3期52-54,共3页Machinery & Electronics

摘　　要：考虑到模糊控制算法在自主避障上的缺陷,设计了一种改进的模糊CMAC神经网络车辆自主避障算法。采用模糊CMAC神经网络(FCMAC)的各层节点来实现模糊控制器变量的输入、模糊化、模糊逻辑的前提条件匹配运算、模糊量的归一化、控制量的输出和控制规则的调整,借助神经网络的自学习能力来完成模糊控制。通过仿真实验,验证了改进的模糊CMAC神经网络算法的可行性和有效性。Autonomous obstacle avoiding based on the improved fuzzy CMAC put forward the neural network for vehicles is derived.With the increase of environmental complexity,fuzzy al-gorithm becomes complicated,and only through the fuzzy algorithm is difficult to determine and ad-j ust the system behavior described in the member-ship functions and fuzzy rules.Because of neural network to control the issue with high availability, and CMAC neural network has a unique “address-ing the way”,The traditional fuzzy control algo-rithm is modified through CMAC neural network to reduce the computation of fuzzy algorithm to make the robot more efficient obstacle avoidance.＆amp;nbsp;Simulation results show that the improved fuzzy CMAC algorithm is feasible.

关 键 词：自主避障 模糊控制 模糊CMAC神经网络 

分 类 号：V249.4[航空宇航科学与技术—飞行器设计]