Model predictive control with fuzzy logic switching for path tracking of autonomous vehicles

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence in Automotive Engineering
• Nonlinear System Control

Background:

• Autonomous vehicles require robust path tracking for safe navigation.
• Vehicle dynamics are highly nonlinear, posing challenges for traditional control methods.
• Model Predictive Control (MPC) offers a framework for handling constraints and optimizing control actions.

Purpose of the Study:

• To develop an integrated path tracking control strategy for autonomous vehicles.
• To address the challenges of vehicle nonlinearity and operational variability.
• To improve the precision and reliability of steering angle and angular velocity control.

Main Methods:

• Implementation of a multi-input multi-output linear model predictive control (LMPC) utilizing Laguerre networks.
• Integration of a fuzzy logic switching system to manage multiple linearized vehicle models.
• Application of gap metric analysis for optimal selection of linearized models.
• Simulation of various vehicle maneuvers using generated paths from path planning algorithms.

Main Results:

• The proposed LMPC with fuzzy logic switching demonstrated superior path tracking performance compared to a Linear Quadratic Gaussian (LQG) controller.
• The controller effectively managed physical constraints on control signals and measurement noise.
• Optimized control signals for steering angle and angular velocity were achieved across different operating points.

Conclusions:

• The integrated LMPC and fuzzy logic control strategy provides an effective solution for autonomous vehicle path tracking.
• This approach offers enhanced performance and robustness in handling nonlinear vehicle dynamics.
• The developed controller represents a significant advancement in autonomous driving control systems.