The paper discusses the development of a soft-constrained iterative linear quadratic regulator (soft-CILQR) algorithm for autonomous vehicle applications. The algorithm integrates the CILQR algorithm and a model predictive control constraint relaxation method. This system aims to calculate stable control inputs in a simpler manner. The performance of the soft-CILQR algorithm was tested through numerical simulations and vision-based maneuvers. The results showed that the soft-CILQR solver achieved smoother steering input trajectories and better tracking accuracy than the CILQR solver.
Publication date: 29 Nov 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2311.16900
