移动机器人编队的递归模糊神经网络滑模控制  

Sliding Mode Control of Mobile Robot Formations Based on Recurrent Fuzzy-Neural Network

作　　者：李艳东[1] 朱玲[1] 孙明[1] 

机构地区：[1]齐齐哈尔大学计算机与控制工程学院,黑龙江齐齐哈尔161006

出　　处：《计算机仿真》2013年第7期349-353,382,共6页Computer Simulation

基　　金：国家青年科学基金(61100103);齐齐哈尔大学青年教师科研启动支持计划项目(2011k-M02);黑龙江省教育厅科学技术研究项目(12531761)

摘　　要：研究非完整移动机器人编队控制优化问题,由于动态模型存在诸多不稳定性,针对领航者-跟随者I-ψ控制结构,提出了一种Back stepping运动学控制器与自适应神经滑模控制器相结合的新型控制策略。采用动态递归模糊神经网络(dynam-ic recurrent fuzzy neural network,DRFNN)对跟随者及领航者动力学非线性不确定部分进行在线估计,并通过自适应鲁棒控制器对神经网络建模误差进行补偿。所提方法不但解决了移动机器人编队控制的参数与非参数不确定性问题,同时也确保了机器人编队在期望队形下对指定轨迹的跟踪;根据Lyapunov方法的设计过程,保证了控制系统的稳定;仿真结果表明了改进方法对机器人编队优化控制的有效性。This paper presented a formation control scheme for multiple nonholonomic mobile robots.A new control structure based on leader-follower made possible the integration of a backstepping kinematics controller and a sliding mode controller with adaptive neural network.The DRFNN was developed to achieve on-line estimation for the part of dynamics nonlinear uncertain of follower and leader robot,and the adaptive robust controller compensated modeling errors of uncertainties by neural network.This method can not only solve the problem of parameters and non -parameter uncertainties of mobile robots,but also ensure that the desired trajectory tracking of robot formation in the case of maintaining formation.The stability of the control system was proved by using the Lyapunov theory.The simulation results show the effectiveness of the proposed method.

关 键 词：编队控制 滑模控制 动态递归模糊神经网络 非完整移动机器人 不确定性 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]