Motion Control of a Mobile Robot with Differential Drive Using Model-Predictive Control and Artificial Neural Networks

Mobile robots are vital to automation, logistics, healthcare, and defense, and their dependable functioning depends on accurate and effective control. The nonholonomic restrictions of Differential Drive Mobile Robots (DDMR) in particular present difficulties, necessitating sophisticated control systems to attain precise trajectory tracking and stability. In this paper, a novel control method for DDMR is proposed, which integrates Model Predictive Control (MPC) in the inner loop with Artificial Neural Networks (ANN) in the outer loop. The technology is intended to reduce steady-state errors, improve position regulation, and improve trajectory tracking. The suggested ANN-MPC system's performance is assessed by contrasting it with two other controllers: a traditional PI-PI system and a hybrid ANN-FLC system. According to simulation data, the ANN-MPC method performs better than both options in terms of steady-state error, settling time, and tracking accuracy. In dynamic situations, the suggested system exhibits enhanced robustness, reduced overshoot, and faster convergence. These results demonstrate how well predictive control can be combined with AI-based methods, opening the door for more study on multi-robot coordination and real-time implementation.