Model Reference Adaptive Control of Spherical Robot Based on Generalized Dynamic Fuzzy Neural Network Inverse System

To address the challenges posed by uncertain factors, such as variations in model parameters and external disturbances that affect the performance of spherical robot control system, a Model Reference Adaptive Control (MRAC) system based on a Generalized Dynamic Fuzzy Neural Network (GD-FNN) inverse system is proposed. Initially, the dynamic model of the spherical robot is derived using Kane’s equation, which is based on the robot’s mechanical structure, and its reversibility is analyzed. Subsequently, utilizing the nonlinear approximation capability of the GD-FNN to handle uncertainty, an inverse system for the spherical robot is constructed. The GD-FNN inverse system is then connected in series with the original system to form a pseudo-linear system. Finally, a closed-loop controller is designed for the pseudo-linear system using the MRAC method, and the system’s stability is proven. The proposed design enables effective attitude balance control for the spherical robot. Experimental results validate the feasibility and effectiveness of the proposed control system.