欠驱动机械臂无源-自适应迭代学习控制

针对一类欠驱动串联机械臂,基于其动力学方程的变换,提出了一种分步控制策略。该策略首先利用全驱动关节与欠驱动关节之间的动力学耦合,驱使欠驱动关节达到理想位置进而对其进行锁定。由于欠驱动关节理想夹角的随机性,使得锁定欠驱动关节后的机械臂成为一结构不确定系统。针对此类系统,结合迭代学习和耗散控制理论,提出了一种无源-自适应迭代学习控制方法,使其能够进行轨迹跟踪。通过对三连杆机械臂的仿真验证了其有效性及实用性。

Study of passivity-adaptive iterative learning control of underactuated robots

Aiming at a kind of underactuated serial robots,a multistep control strategy based on its transformation of dynamics equation is proposed.This strategy firstly derives the underactuated joints to desired positions by the coupling of actuated and underactuated joints,and brakes the underactuated joints later.Because of the randomicity of the desired angles of underactuated joints,the robot underactuated joints braked becomes structure uncertain.Combining with the theory of iterative learning and dissipation,a passivity-adaptive iterative learning controller is provided to make this uncertain system capable of handling trajectory tracking.And its validity and practicality are validated by the simulation to a three-link robot.