柔性关节机器人输出反馈自适应轨迹跟踪控制研究

发布时间：2018-05-03 14:19

本文选题：柔性关节机器人 + 奇异摄动　；参考：《东华大学》2017年硕士论文

【摘要】：与传统的关节和连杆均为刚性的机器人相比,柔性关节机器人具有结构紧凑、质量轻和能耗低等特点,被广泛应用于业界。本文基于奇异摄动理论和积分流概念,深入研究了柔性关节机器人的输出反馈自适应轨迹跟踪控制方法,研究成果将为柔性关节机器人的高性能实际应用提供一定的理论依据。首先,根据Spong提出的柔性关节简化模型,利用拉格朗日方程推导出柔性关节机器人的动力学模型,再通过引入奇异摄动理论和积分流概念将其解耦成两个子系统,即慢子系统和快子系统。慢子系统表示柔性关机机器人系统模型中的刚性部分,快子系统是用来描述关节处的柔性特性。从而,柔性关节机器人的控制器设计转变为针对两个子系统的控制器设计。其次,针对柔性关节机器人系统模型中存在参数不确定性和上界未知的外部扰动,设计了一种基于改进参数线性化表达式的鲁棒自适应控制器作为慢子控制律。同时,通过对快子系统方程进行变形,将其转变成典型的目标跟踪类问题,从而使快子控制律的设计更加灵活。此外,为实现仅需位置反馈的输出反馈控制,在快、慢子系统中均引入一种近似微分滤波器来得到伪速度误差信号以消除对速度的测量。然后,针对不同(强或弱)柔性关节系统,本文提出了一种有界自适应控制器来解决柔性关节机器人中存在参数不确定性和输出力矩有界的问题。控制律的设计中,一类饱和函数被用来确保关节执行器输出力矩有界,一种投影自适应控制律用来克服慢子控制律中前馈项的参数不确定性。同时,为使整个闭环控制系统实现仅需位置反馈的输出反馈控制,可分别在慢子系统和快子系统中引入线性滤波器和高增益观测器来得到无法直接测量状态量的估计值。更为重要的是,通过奇异摄动理论和积分流概念推导出校正控制律,实现了对强柔性系统的柔性补偿,使得奇异摄动方法也能适用于强柔性机器人系统。最后,对于柔性关节机器人的稳定性分析,考虑到快、慢子系统的稳定性无法直接保证整个复合控制系统的稳定性。因此本文针对整个奇异摄动系统给出两种不同的稳定性分析思路。此外,通过仿真实验对提出的两类控制算法分别进行了验证。结果表明,本文提出鲁棒自适应输出反馈控制器和有界自适应输出反馈控制器具有更优的轨迹跟踪性能。
[Abstract]:Compared with the traditional robot with rigid joints and connecting rod, flexible joint robot is widely used in the industry because of its compact structure, light weight and low energy consumption. Based on the singular perturbation theory and the concept of integral flow, the output feedback adaptive trajectory tracking control method of flexible joint robot is studied in this paper. The research results will provide some theoretical basis for the practical application of flexible joint robot with high performance. Firstly, according to the simplified flexible joint model proposed by Spong, the dynamic model of flexible joint robot is derived by using Lagrangian equation, and then it is decoupled into two subsystems by introducing singular perturbation theory and the concept of integral flow. Slow subsystem and fast subsystem. The slow subsystem represents the rigid part of the flexible shutdown robot system model, and the fast subsystem is used to describe the flexible characteristics of the joint. Thus, the controller design of flexible joint robot is transformed into the controller design for two subsystems. Secondly, a robust adaptive controller based on the improved parametric linearization expression is designed to solve the external disturbances with uncertain parameters and unknown upper bounds in the flexible joint robot model. At the same time, the fast subsystem equation is transformed into a typical target tracking problem by deforming the fast subsystem equation, which makes the design of the fast subsystem control law more flexible. In addition, in order to realize the output feedback control with only position feedback, an approximate differential filter is introduced in both fast and slow subsystems to obtain pseudo-velocity error signals to eliminate the measurement of velocity. Then, for different (strong or weak) flexible joint systems, a bounded adaptive controller is proposed to solve the problem of parameter uncertainty and output moment boundedness in flexible joint robot. In the design of control law, a kind of saturation function is used to ensure that the output torque of joint actuator is bounded, and a projection adaptive control law is used to overcome the parameter uncertainty of feedforward term in the control law. At the same time, in order to realize the output feedback control of the whole closed-loop control system, the linear filter and the high gain observer can be introduced into the slow subsystem and the fast subsystem to obtain the estimated value of the state quantity which can not be measured directly. More importantly, through the singular perturbation theory and the concept of integral flow, the correction control law is derived, and the flexible compensation for the strong flexible system is realized, so that the singular perturbation method can also be applied to the strong flexible robot system. Finally, for the stability analysis of flexible joint robot, considering the speed, the stability of the slow subsystem can not directly guarantee the stability of the composite control system. So this paper gives two different methods of stability analysis for the whole singularly perturbed system. In addition, two kinds of control algorithms are verified by simulation experiments. The results show that the robust adaptive output feedback controller and the bounded adaptive output feedback controller have better trajectory tracking performance.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】