Indirect adaptive control via parallel dynamic neural networks

Stability conditions for a parallel dynamic neural network by means of Lyap unov-like analysis are determined. The new learning law ensures that the id entification error converges to zero (model matching) or to a bounded zone (with unmodelled dynamics). Based on the neural identifier we present a loc al optimal controller and analyse the tracking error. Our principal contrib utions are that we provide a bound for the identification error of the para llel neuro identifier and that we then establish a bound for the tracking e rror of the neurocontrol.