A NODAL LINK PERCEPTRON NETWORK WITH APPLICATIONS TO CONTROL OF A NONHOLONOMIC SYSTEM

A new perceptron neural network (PNN) for functional approximation and control of a general class of nonlinear systems is introduced. The ba sic structure of the network along with the conditions for its exponen tial convergence under a suitable training law are derived, A novel di screte-time control strategy is formulated that employs the PNN for di rect online estimation of the feedforward control input. The developed controller can be applied to both discrete- and continuous-time plant s. Unlike most of the existing direct adaptive or learning schemes, th e nonlinear plant is not assumed to be feedback linearizable. The deve loped controller is then applied for tracking control of a nonholonomi c (free-flying) robot. The simulation results of this application demo nstrate a perfect tracking performance after the network is fully trai ned.