A successive learning neuro control system shooting irregular moving object

A successive learning control system based on a neural network technique with a genetic algorithm has been developed to simulate a human real-time learning process. As an application experiment, a 3 degree-of-freedom arm robot shoots a ball equipped with a CCD camera at an irregular moving basket. Where the hitting rate is improved by the successive learning control and the final value was 23% in average, 40% in maximum.<
>