移动机器人编队的神经网络滑模控制  

Sliding mode control of mobile robot formations based on neural networks

作　　者：朱玲[1] 李艳东[1] 孙明[1] 王宗义[2] 

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 [2]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《电机与控制学报》2014年第3期113-118,共6页Electric Machines and Control

基　　金：国家青年科学基金(61100103);黑龙江省教育厅科学技术项目(12531761)

摘　　要：针对非完整移动机器人编队控制问题,基于领航者-跟随者l-ψ控制结构,提出了一种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。采用径向基神经网络(radial basis function neural network,RBFNN)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。实验结果表明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,还确保了机器人编队在期望队形下对指定轨迹的跟踪;基于Lyapunov方法的设计过程,保证了控制系统的稳定。A combined kinematic/adaptive neural sliding mode control law is proposed for leader-follower based formation control of multiple nonholnomic mobile robots. The radial basis function neural network （RBFNN） was introduced to approximate the nonlinear uncertain part of dynamics of the follower and its leader, and the adaptive robust controller compensated modeling errors of uncertainties by neural net- work. The results show that this approach not only solves parameter and non-parameter uncertainties of mobile robots, but also ensures 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.

关 键 词：编队控制 滑模控制 非完整移动机器人 不确定性 领航者-跟随者 动力学 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]