Friction-Robust Autonomous Racing Using Trajectory Optimization Over Multiple Models

Engineering notes

Key topics: autonomous driving, autonomous vehicle, motion planning, planning, control. See the paper for implementation details and experimental results.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

Autonomous vehicle control in low-friction environments should be capable of using all of the available traction at the road to accomplish maneuvering objectives. In these environments, however, the limit of traction is difficult to estimate, which challenges standard motion planning techniques. In this paper, we introduce a trajectory optimization framework that natively incorporates the complex and nonlinear effects of friction uncertainty into the planning process to improve both the performance and robustness of maneuvering at high accelerations. The core approach of the method is to explicitly consider a range of possible dynamics models, inclusive of their closed-loop behavior, simultaneously in the optimization. We illustrate this method through a racing example, where the minimum-time objective facilitates intuitive performance and robustness metrics (lap time and tracking error limits), all while necessitating vehicle maneuvering through nonlinear and friction-sensitive regions of the state space. Experiments on an autonomous VW Golf GTI on a challenging winter ice track demonstrate the efficacy of this approach.

Links and sources

• Official / arXiv page

Comments