Learning from neural control

One of the amazing successes of biological systems is their ability to "learn by doing" and so adapt to their environment. In this paper, we firstly present an adaptive neural controller which is capable of learning the system dynamics during tracking control to periodic reference orbits. A partial persistent excitation (PE) condition is shown to be satisfied, and accurate NN approximation for the unknown dynamics is obtained in a local region along the tracking orbit. Secondly, a neural learning control scheme is proposed which can effectively recall and reuse the learned knowledge to achieve local stability and better control performance. The significance of this paper is that it presents a dynamical deterministic learning theory, which can implement learning and control abilities similarly to biological systems.