Backstepping design of adaptive robust learning controller for fast trajectory tracking of ball-beam dynamic systems

This paper presents a robust learning controller for ball-beam dynamic systems. The error dynamics equations are derived for ball and beam dynamics respectively and compensated by a robust adaptive learning control input. In the learning controller, all the error signals, parameter error, tracking error and learning control input error are shown to be bounded and the tracking error converges to zero as the learning proceeds. Computer simulation results are given to demonstrate the feasibility and effectiveness of the proposed learning scheme.