ILC For Robotic Manipulators Based On Kinematic Model

Robotic manipulators are suitable for dealing with the repetitive tasks during the process of light industry or handicraft.In fact,robotic manipulators are already widely used in car production and electronic manufacturing industry.However,these robotic manipulators are introduced to handle the high-precision tasks.It takes a lot to purchase and maintenance these facilities,so that the medium and small enterprises are hard to afford the cost.It is necessary to choose a kind of low-cost manipulator system and to put forward the corresponding control method.Three kinds of control methods are put forward in this paper to deal with the control task of robotic manipulator system with kinematic controller.The main research results can be summa-rized as follows:(1)A kind of PD-type closed-loop iterative learning control(ILC)method is put forward to deal with the control task when the model of the manipulator is unknown.First of all,ILC con-troller is chosen because of its effectiveness for the tracking tasks of nonlinear system.Secondly,closed-loop ILC is chosen to deal with the environment disturbance.Finally,in consideration of the integral effect of ILC,a PD-type closed-loop ILC method is put forward to solve the control problem.Simulation results verify the effectiveness of this method.(2)A kind of ILC method based on kinematic model is put forward to deal with the control task when the model of the manipulator is available.By taking use of the kinematic model,the controller can get the reference joint angles to improve the learning process.As the simulation results show,tracking error decreases faster when the ILC method based on kinematic model is used.(3)A kind of ILC method based on adaptive kinematic model is put forward to deal with the control task with model mismatch.By estimating the parameters based on Kalman filter,the reference kinematic model is updated during the iterative process.The simulation shows that the ILC method based on adaptive kinematic model can solve the control problem with model mismatch.(4)The robotic manipulator physical platform is built,and experiments of the control method in(1)and(2)are studied based on the physical platform.As the results show,the output fluctuates because of the observation error,but the tracking error decreases during the iterative process.Error analysis of the results is presented.