Stochastic neural adaptive control for nonlinear time varying systems based on Newton and gradient optimizations

The authors present a stochastic neural adaptive control algorithm for nonlinear time-varying systems. The implicit neural identification is derived based on the Newton optimization approach. Using the one-step-prediction quadratic performance index, the authors design a control law which in combination with the identification algorithm constitutes an effective neural adaptive control algorithm. The identification and control are robust and computationally efficient for real-time control systems design