Discrete-time adaptive control of feedback linearizable nonlinear systems using neural networks

A two-layer neural network-based controller in discrete-time which feedback linearizes a MIMO nonlinear system is presented. The neural network (NN) controller exhibits learning-while-functioning-feature and its structure is derived using filtered error notions. A uniform ultimate boundedness of the closed-loop system is given in the sense of Lyapunov and without using certainty equivalence. In addition, new online tuning algorithms are derived, which are similar to ε-modification for the case of continuous-time systems. These weight tuning algorithms guarantee tracking as well as bounded NN weights in nonideal situations