AI-based sideslip angle estimation for robust path-tracking control using line integral Lyapunov function and <i>H</i> _∞ for autonomous vehicles

Abstract

Abstract This paper proposes a robust control approach for autonomous vehicle path-tracking based on sideslip angle estimation. First, artificial intelligence techniques, namely artificial neural networks and adaptive neuro-fuzzy inference systems (ANFIS), are employed for sideslip angle estimation. The optimal model structure is determined based on statistical error metrics, including mean square error (MSE), mean absolute error, and root MSE. Next, the controller is designed using a Takagi–Sugeno fuzzy multi-model combined with a line integral Lyapunov function and H ∞ performance, leading to a convex optimization problem formulated as linear matrix inequalities, from which optimal gains are obtained. The approach is validated through MATLAB/Simulink and CarSim co-simulations under various scenarios and compared with sliding mode control, Lyapunov quadratic function-based control, and PID methods. Results show that the ANFIS-based method achieves better accuracy and robustness across different conditions.