IJCAS

This study introduces a Takagi-Sugeno (T-S) fuzzy modeling framework for kinematic modeling of mecanum wheeled mobile robot (MWMR). T-S fuzzy systems are particularly effective in capturing complex nonlinear dynamics and unmodeled subsystems inherent to MWMR architectures. Optimization of parameters within the T-S structure is achieved through a genetic algorithm (GA), enabling precise alignment between the T-S derived model and physical system behavior. Notably, the proposed methodology achieves convergence to optimal T-S model within 200 generations of the GA, without necessitating an explicit analytical formulation of the complete MWMR dynamics. Validation experiments reveal the optimized T-S model achieves 0.015 m/s a mean squared error (MSE) difference relative to empirical velocity profiles from the MWMR platform. Rigorous numerical assessment demonstrates the formulated T-S model achieves exceptional dynamic congruence with the physical MWMR platform, manifesting peak velocity discrepancies of 57×10−4 m/s accompanied by standard deviations of 0.027 m/s across experimental trials. Comparative evaluation against conventional probabilistic modeling techniques highlights superior predictive accuracy and dynamic fidelity of the proposed T-S framework. Observed results substantiate the model’s capacity to replicate nonlinear kinematic interactions and transient velocity characteristics under experimentally validated boundary conditions, corroborating theoretical expectations through empirical system identification.

Keywords: Genetic algorithm, mecanum-wheeled mobile robot, system approximation, Takagi-Sugeno modelling.