空间柔性机械臂控制策略研究

【作者】 张晓东；

【导师】 孙汉旭； 贾庆轩；

【作者基本信息】 北京邮电大学 ， 机械电子工程， 2009， 博士

【摘要】 随着21世纪航天事业的迅速发展和对太空探索的不断深入,未来空间任务越来越多,如空间站的建造维护、空间设备的维修、科学实验等,目前尚不能全靠宇航员来完成,由于空间机械臂具有适应太空作业环境的能力,利用空间机械臂完成一些太空作业任务具有重要的意义,已经成为空间技术研究领域内的一个重要的研究方向。本论文的研究内容来源于教育部科学技术研究重大项目“空间柔性机器人动力学/控制耦合系统的研究”,本论文以此为依托,重点对大型空间柔性机械臂动力学建模理论和控制策略进行了深入的分析和研究,本文的主要工作如下:首先,设计了一种大型的空间六自由度机械臂,提出了需要完成的指标性的作业任务,在对空间机械臂进行精度分析的基础上,针对具体的作业任务对机械臂进行精度分配;建立由柔性臂杆和柔性关节构成的空间多自由度柔性机械臂动力学模型,并验证动力学模型的有效性;建立空间机械臂的地面气浮模拟试验平台,针对地面气浮模拟平台上空间机械臂的构型特点,建立了平面两连杆柔性机械臂动力学模型,分析机械臂的弹性-阻尼约束边界条件,验证柔性机械臂在此约束条件下的动力学特性。其次,通过对机械臂关节的非线性特性进行分析,建立了考虑关节非线性特性的柔性关节动力学模型;为克服柔性关节非线性因素的影响,提出双位置闭环伺服控制,输出反馈抗饱和的高精度PID控制和基于LuGre摩擦模型的自适应补偿控制策略;通过机械臂单关节测试平台上的试验表明:所提出的控制策略不仅有效的提高了非线性影响下的位置控制精度和稳定性,并且可实现饱和非线性影响下空间机械臂系统的全局渐进稳定。再次,建立不确定性影响下的柔性关节机械臂级联动力学方程,基于反演设计的思想,设计虚拟控制量实现了级联方程的递阶控制,在反演镇定函数中综合积分项,进一步消除轨迹跟踪稳态误差,并且设计自适应滑模控制器消除不确定因素影响,基于Lyapunov理论证明了整个系统的稳定性和对误差的收敛性;为避免由于虚拟控制反复求导而导致的计算复杂度问题,将动态滑模面控制与自适应神经网络相结合,提出了空间柔性关节机械臂的自适应神经网络动态滑模面控制策略,并进行控制策略的稳定性分析、仿真研究和试验验证研究。再其次,以具有柔性关节和柔性臂杆的空间柔性机械臂为研究对象,针对空间柔性机械臂的定位控制任务,通过输入输出线性化将非线性柔性机械臂动力学方程简化为近似线性系统,采用具有输出反馈的自学习输入成形控制策略进行定位控制时的振动抑制;针对轨迹控制任务,基于奇异摄动方法将空间柔性机械臂简化为快慢两个独立的子系统,采用复合控制策略实现精确轨迹跟踪的同时实现有效的振动抑制;在空间机械臂地面气浮模拟试验平台上的试验表明:提出的控制策略不仅可保证高精度的定位控制精度,而且可有效的抑制弹性振动,满足空间柔性机械臂的指标要求。最后,针对空间机械臂的点位控制和轨迹跟踪作业任务,在六自由度空间柔性机械臂实验平台上进行试验研究,对提出的作业任务指标进行验证。更多还原

【Abstract】 In the 21st century, with the rapid development of aerospace industry and the deepening of space exploration, more and more future space missions, such as the construction of the space station, space equipment maintenance, and other scientific experiments need to be finished, but it is still hard to be completed all by astronaut. Space manipulator has the capability of adapting space environment, so it is significant to complete some tasks using space manipulator, and it has become an important research direction in space technology field. This paper come from the Ministry of Research Science and Technology major projects "space flexible robot dynamics/control coupling system research", Based on this project, this paper mainly focus on the large flexible space manipulator dynamics modeling theory and control strategy research, the paper’s main work are as follows:Firstly, a six freedom, large space manipulator is designed, and raises the needed tasks target, distributes the accuracy of flexible manipulator aimed at the specific task target on the basis of the space flexible manipulator accuracy analysis; The dynamics modeling is build on the flexible links and flexible joints manipulator, and verify the effect of dynamic model; An air space simulation test table on the ground is build and aimed at the configuration features of space manipulator on test table, a simplified flat two-link manipulator is set up; The flexibility-damping boundary conditions of flexible manipulator is analyzed, and the dynamic characteristics are researched under this restrictive conditions.Secondly, the manipulator joint nonlinearities are analyzed, and the nonlinear dynamics model of flexible joints is set up; To overcome the influences of such nonlinear factors, a kind of two-position closed-loop servo control system, an anti-saturation output feedback high-precision PID control and the adaptive friction compensation strategy based on LuGre friction model are proposed; Experiment on the designed manipulator single joint test table verify that the control strategy not only improved the nonlinear position control precision and stability, but also the global stability of space manipulator system can be realized under the influence of saturation, and control algorithms structure is simple, control parameter setting rules are clear.Thirdly, flexible-joint manipulator cascade dynamic equation is established under the influence of uncertainty; Based on the backstepping methods, a virtual control variable is designed to achieve cascade equation’s control, an integration function is synthesized to elimination of tracking steady-state errors furthermore, and the adaptive sliding controller is designed to eliminate the influence of uncertainties, and improve the whole system’s stability and the convergence to errors by Lyapunov theory; In order to avoid the complexity for the repeated derivation of virtual control, combine the dynamic sliding-control and Neural networks, a space flexible-joint manipulator adaptive dynamic sliding-control strategy is proposed. Then, the stability analysis, simulation and test verification are finished.Fourthly, aimed at the flexible-joints and flexible-link space manipulator, for the positioning control of flexible manipulator, simplify the nonlinear dynamics of flexible manipulator through the input and output linearization, a recursive learning the output feedback control strategy is proposed for the vibration suppression in position control; In the trajectory control tasks, flexible space manipulator will be simplified to two separate subsystems based on the singular perturbation method, and then using composite control strategy to achieve precision tracking and achieve effective vibration suppression at the same time, and simulation research are studied; Finally, verify the control strategy in the space manipulator ground simulation test table.Finally, aimed at the set-point control and tracking control tasks of space manipulator, the indicators task targets are verified on the six freedom flexible space manipulator experiment platform.更多还原