A fine contact motion of manipulators based on learning control

The transitional motion when a robotic manipulator and a workpiece first make contact is fundamental and occurs frequently. The authors propose a learning controller to improve the contact motion. The learning controller has two features. It observes input and output signals at different periods to improve the transient response of the system; and it can modify an initial state of the system, which includes an approaching velocity of the manipulator. By modifying the approach velocity, a fine contact motion can be achieved, because the magnitude of the impact force depends on the approach velocity. Simulation results verify the effectiveness of the proposed method.<<ETX>>