基于视觉伺服的平面非完整移动机器人鲁棒镇定和跟踪方法研究

【作者】 杨芳；

【导师】 王朝立；

【作者基本信息】 上海理工大学 ， 系统分析与集成， 2011， 博士

【摘要】 现有的非完整运动学系统的镇定和跟踪控制器设计时大多假定运动学模型是精确已知的。但当模型参数不精确或测量不准确时，实际系统的运动学模型存在复杂的不确定性。轮式移动机器人是一类典型的非完整系统,近年来,由于机器人动力学内在非线性特性以及在实际中的多方面的应用,具有非完整约束的移动机器人的控制问题引起了学术界的广泛的关注。由于未校准参数、外部扰动的存在以及测量和未建模动态等的不确定性，实际中很难获取到机器人系统精确的运动学模型。而且这种不确定常常使得模型中含有状态和控制耦合的不确定非线性成分，输入前面带有不确定性，这给控制研究带来了极大的困难。本论文就是借助于视觉伺服反馈,利用自适应控制、变结构控制、动态反馈控制、两步法等多种控制方法结合以及Backstepping技巧和Lyapunov稳定性理论研究视觉参数未知情形下不确定的非完整移动机器人的鲁棒镇定、轨迹跟踪以及同时镇定和跟踪控制问题。具体展开了以下几个方面的工作:1、探讨平面非完整移动机器人在单目摄像机下的不确定运动学和动力学模型。2、研究了视觉参数未知情形下,(2,0)型不确定非完整移动机器人在质心与几何中心重合和不重合两种情形下的运动学和动力学系统的鲁棒镇定问题,给出了运动学镇定控制器以及基于动力学的更加实际的力矩镇定控制器设计,并给出严格的证明。3、研究了视觉参数未知情形下,(2,0)型不确定非完整移动机器人在质心与几何中心重合和不重合两种情形下的运动学和动力学系统的轨迹跟踪控制问题,设计了运动学跟踪控制器以及基于动力学的力矩跟踪控制器,并给出严格的证明。4、研究了视觉参数未知情形下,(1,1)型不确定非完整移动机器人基于链式标准形式的轨迹跟踪控制问题,设计了基于动力学的力矩跟踪控制器,并严格证明了控制系统的稳定性。5、研究了视觉参数未知情形下,(2,0)型不确定非完整移动机器人的同时镇定和跟踪控制器设计问题,设计了统一的跟踪和镇定控制器,并给出严格的证明。6、在上面几个内容研究成果的基础上，搭建由平面非完整移动机器人和单目视觉传感器构成的实验平台,进行仿真实验检验。更多还原

【Abstract】 The designing of stabilizing and tracking controller of nonholonomic kinematicsystems is mostly investigated based on the assumption that the kinematics model isexactly known. But when the model parameters are not accurate or inaccuratemeasurement exists, the kinematics model of the actual system is uncertain. Wheeledmobile robots are a class of typical nonholonomic system. The control of mobile robotswith nonholonomic constraints has attracted much attention due to the inherentnonlinearity in dynamics of the robots and the usefulness in many applications. In thereal world, the precise kinematic model of mobile robot systems is difficult to obtainedbecause of the uncalibration parameters, external disturbances and unmodeled dynamics,measurement uncertainties. The uncertain nonlinear part with state and control couplingincluded in the model can bring the uncertain parameters in front of the inputs, it willbring great difficulties to the study. In this paper, based on visual servo feedback, theadaptive control, variable structure control, feedback control, two step method,Backstepping techniques and Lyapunov stability theory are used to study the robuststabilization, trajectory tracking and simultaneous stabilization and tracking control ofnonholonomic mobile robots when visual parameters are unknown. The main studiesare as follows:1、Discuss the uncertain kinematics and dynamics models of planar nonholonomicmobile robots with the monocular camera.2、The robust stabilization problem of the type (2,0) nonholonomic mobile robot atthe two cases that the mass center and geometry center is coincide and not coincide arediscussed, the kinematic stabilization controller is firstly designed, and then, the morerealization torque stabilization controller is designed, the strict proof is given.3、The trajectory tracking control problem of the type (2,0) nonholonomic mobilerobot at the two cases that the mass center and the geometry center is coincide and notcoincide are discussed, the kinematic tracking controller is firstly designed, and then,the more realization torque tracking controller is designed, the strict proof is given. 4、The trajectory tracking control problem of the type (1,1) nonholonomic mobilerobot is discussed, based on the standard chained system model, the torque trackingcontroller is designed, the strict proof is given.5、The simultaneous stabilization and tracking control problem of the type (2,0)nonholonomic mobile robot is discussed, the torque controller is designed which canguarantee the uniform stabilization and tracking control when the visual parameters areunknown, the stability of the proposed control system is rigorously proved.6、Based on the research results mentioned above, the experimental platform ofplanar nonholonomic mobile robot with monocular camera sensor is set up, simulationsor experiments are done to illustrate the effectiveness of the proposed controllers.更多还原