基于时延估计的机器人有限时间跟踪控制  

Finite-time Tracking Control for Robots Based on Time-delayed Estimation

作　　者：刘海涛[1,2] 张铁[2] 

机构地区：[1]广东海洋大学工程学院,广东湛江524088 [2]华南理工大学机械与汽车工程学院,广东广州510640

出　　处：《机床与液压》2012年第23期12-16,共5页Machine Tool & Hydraulics

基　　金：国家863重点项目(2009AA043901-3);粤港关键领域招标(20090101-1);广东省科技计划项目(2010B080703004);广东省省部产学研项目(2010B090400259)

摘　　要：针对具有各种不确定性的工业机器人复杂系统,提出一种基于时延估计的连续有限时间轨迹跟踪控制方法。该方法不需要机器人系统复杂的动力学模型知识,从而避免了机器人逆动力学的在线实时计算;系统的未知不确定性和外界干扰利用时延估计在线获得,并对控制过程进行有效的补偿,因而算法简单,易于实际应用。引入非线性的分数幂误差项,实现了系统的有限时间跟踪控制,因而收敛速度快,跟踪精度高,鲁棒性强;利用有限时间稳定性相关理论证明该闭环系统是有限时间稳定的,通过2自度工业机器人的数值仿真说明了该算法的有效性。A continuous finite-time tracking control method was presented based on time-delayed estimation for complicated in- dustrial robot with uncertainties. Without complicate dynamic knowledge, the inertia matrix was replaced by a positive constant sym- metrical matrix to avoid all on-line computational problems associated with the robot inverse dynamic, and unknown uncertainties and external disturbances were online obtained by the time-delay estimation to compensate for the control. So the proposed method was quite simple and convenient to application in high-speed robots. Nonlinear exponential errors were used to realize the finite-time tracking con- trol, so fast convergence, high-accuracy and robustness were gotten. The finite-time stability of robot close-loop system was proved by LaSalle＇s invariant principle and finite-time stability. Simulations on a 2-DOF industrial robot demonstrate the effectiveness of this algo- rithm.

关 键 词：工业机器人 时延估计 有限时间稳定 跟踪控制 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]