Integrated Trajectory Planning and Tracking Control for Autonomous Vehicles under Uncertain Environments

Abstract

 Traditional autonomous vehicle architectures exhibit a distinct functional disconnect between trajectory planning and tracking control. While this hierarchical design offers advantages in terms of computational efficiency, it often leads to a severe compromise in vehicle dynamic feasibility when confronted with complex environments characterized by stochasticity and high-frequency disturbances. This paper explores a coupled control framework designed to integrate these two typically isolated layers into a unified optimization manifold, specifically addressing systemic uncertainties such as fluctuations in road adhesion coefficients and unmodeled vehicle sideslip dynamics. The study investigates the potential of this unified control approach to maintain vehicle stability while simultaneously satisfying rigorous safety boundaries. The results indicate that, although this coupled method enhances tracking accuracy to some extent in moderately uncertain environments, its overall efficacy remains fundamentally contingent upon the precision of the underlying disturbance observer.