A direct adaptive iterative learning control for robotic systems using only position measurement

In this paper, we propose a position based direct adaptive iterative learning control for robotic systems with repetitive tasks. As we assume that the joint velocities are not measurable, we introduce a sliding window of measurements to design the iterative learning controller without state observer. Based on a derived error model, a fuzzy neural network is applied to design a fuzzy neural learning component for compensation of the unknown certainty equivalent controller. Then, an averaging filter approach is applied to overcome the relative degree problem. Finally, the uncertainties due to approximation error and state estimation errors can be solved by introducing a robust learning component. Based on a Lyapunov like analysis, we show that the finiteness of control parameters and control input can be guaranteed for all the time interval during each iteration without using parameter projection. The norm of output tracking error will asymptotically converge to a tunable residual set as iteration goes to infinity.