Priority-Aware Tracking of Infeasible Trajectories via Smooth Input Saturation and Asymmetric Control Structure

This paper presents a novel priority-aware control strategy for tracking infeasible trajectories. While the upper-level planner generates trajectories known to be feasible, a sudden change in traffic or road friction can cause the planned trajectory to be impossible to follow accurately. In this situation, it is considered optimal to prioritize tracking the lateral path over longitudinal acceleration. The difficulties in priority-aware control originate from the nonlinear constraints on tire forces. Existing NMPC approaches with explicit nonlinear constraints may suffer from numerical sensitivity and chattering. On the other hand, rule-based control strategies are commonly used but do not explicitly perform optimization. To overcome these issues, the proposed controller utilizes a smooth input saturation method to design a well-defined control problem. The input saturation is intended to improve numerical behavior by avoiding explicit active-set switching. To induce lateral priority, the integrated cost is designed to be asymmetrical. This integrated control structure allows the controller to set the level of priority by adjusting its weighting factors. Simulation results show different levels of lateral prioritized control in low-friction roads. The compromise in lateral control improves longitudinal control, reducing collision risk by about 10% even in saturated control situations.