Adaptive Model Predictive Control for 4WD-4WS Mobile Robot: A Multivariate Gaussian Mixture Model-Ant Colony...

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence and Optimization

Background:

• Accurate trajectory tracking is essential for autonomous mobile robots in dynamic environments.
• Model Predictive Controller (MPC) performance relies heavily on optimal parameter tuning.
• Existing optimization algorithms struggle with interdependencies between variables and balancing exploration/exploitation.

Purpose of the Study:

• To develop and validate an advanced optimization algorithm, MGMM-ACOR, for tuning MPC parameters.
• To improve the robustness and accuracy of trajectory tracking for a 4WD-4WS mobile robot.
• To address limitations in conventional optimization methods for complex robotic control tasks.

Main Methods:

• Implementation of Multivariate Gaussian Mixture Model Continuous Ant Colony Optimization (MGMM-ACOR).
• Integration of MGMM-ACOR with a nonlinear Model Predictive Controller (MPC) for a 4WD-4WS mobile robot.
• Two-phase validation: benchmark function testing and real-world trajectory tracking simulations (circular, eight, and obstacle avoidance).

Main Results:

• MGMM-ACOR demonstrated superior convergence speed and solution accuracy compared to ACO, ACOR, and PSO variants on benchmark functions.
• The integrated MGMM-ACOR-MPC system achieved stable, collision-free trajectory tracking.
• The proposed method outperformed conventional ACOR approaches in trajectory error, control effort, and computational latency.

Conclusions:

• MGMM-ACOR effectively optimizes MPC parameters by modeling variable correlations, leading to enhanced robotic trajectory tracking.
• The algorithm provides a robust solution for autonomous navigation, ensuring safety and efficiency in complex scenarios.
• This work advances the state-of-the-art in robot control optimization and autonomous system performance.