Fuzzy Rules Data-Driven Equivalent Model with Multi-gradient Learning for Discrete-Time Nearly Optimal Control

Abstract A behavior of switchable control directions is investigated for the non-holonomic robotic system considered as a class of unknown nonlinear discrete-time systems. The data-driven equivalent model is established by a multi-input fuzzy rule emulated network and the multi-gradient learning law is developed to tune all adjustable parameters. Thereafter, the nearly optimal controller is derived using the dynamics of the equivalent model and the closed-loop performance is analyzed through rigorous mathematical analysis. The experimental system is constructed to validate the effectiveness of the proposed scheme and the advantage of the multi-gradient approach.