Optimal discrete-time sliding-mode control based on recurrent neural network: a singular value approach

In this paper, a strategy involving the combination of optimal discrete-time sliding-mode control and recurrent neural networks is proposed for class uncertain linear systems. First, performance index based on reaching law signal defined. Then, constrained quadratic programming problem formulated considering limitations as static constraint. The dynamic algebraic model network derived optimization conditions their relationship with projection theory. method prevents chattering by selecting proper parameters twisting law. convergence analysed using Lyapunov stability A singular value-based analysis employed robustness method. closed-loop system are studying eigenvalues matrix value approach. algorithm assessed in simulated example terms elimination, solution feasibility, encountering uncertainties compared recently DSMC methods literature.