摘要
采用模糊神经网络并结合模糊逻辑控制的方法,研究在动态环境下移动机器人的在线路径跟踪与实时避障问题。针对移动机器人的运动学模型,依据点与直线间的距离关系设计了移动机器人的路径跟踪算法,提出了一种基于模糊神经网络的移动机器人在线路径跟踪方案,采用改进的BP算法对网络进行学习与训练,利用梯度下降法调整网络的权值与阈值,使其实际输出与期望输出的误差总均方差最小。同时,运用模糊逻辑控制,实现了移动机器人的实时避障。仿真实例证实了控制方案的有效性,表明了所提出的跟踪算法与控制方案具有良好的动态路径跟踪与实时避障能力。
The problems of mobile robot on-line path tracking and real-time avoiding obstacles in a dynamic environment are studied.A method of fuzzy neural networks（FNN）combined with fuzzy logic control（FLC）is presented.A scheme of mobile robot on-line path tracking is designed based on FNN according as kinematics models of mobile robot and the interval relation between point and line.A reformative BP algorithm is used to complete the network learning and training.The gradient descending method is utilized to adjust the weight and threshold value of this network,which makes the error mean square deviation between reality output and expected output be minimum.The FLC is utilized to fulfill the real-time avoiding obstacles.Simulation results show that the method is efficient and possesses better abilities of dynamic path tracking and real-time avoiding obstacles with regard to an unknown dynamic environment.
出处
《控制工程》
CSCD
北大核心
2010年第3期397-400,共4页
Control Engineering of China
基金
哈尔滨工业大学机器人技术与系统国家重点实验室开放基金资助项目(SKLRS200708)
天津工程师范学院科研发展基金资助项目(KJ2008028)
关键词
移动机器人
路径跟踪
模糊神经网络
避障
mobile robot
path tracking
fuzzy neural networks
avoid obstacles
