Backlash Compensation in Nonlinear Systems by Dynamic Inversion Using Neural Networks: Continuous and Discrete Time Approaches

Two different dynamic inversion compensation schemes for control of nonlinear system with input backlash are presented; one in continuous time and one in discrete time. Both schemes use the backstepping technique with neural networks (NN) for inverting the backlash nonlinearity in the feedforward path. The technique provides a general procedure for using NN to determine the dynamics preinverse of an invertible dynamical system. Tuning algorithms are given for the NN weights so that the backlash compensation schemes guarantee bounded tracking and backlash errors, and also bounded parameter estimates. This paper presents a mathematical approach based on rigorous proofs that guarantees both stability and performance. Simulation examples are given to verify closed-loop performance.