Computing the Lateral Compliance of the Racing Line Using Trajectory Optimization

While race car drivers speak often on racing lines, data reveals they do not follow consistent paths lap over lap. Furthermore, autonomous racing controllers designed to track optimal paths fail to keep pace with the best professional drivers. In this paper, we assess the importance of the racing line from an optimization perspective by evaluating its dynamic sensitivity to lateral perturbations. After applying the method to the Laguna Seca Raceway, we find that there is in fact a family of trajectories that are significantly different in path but similar in lap time. This finding is consistent with prior experimental work and indicates that extracting the peak dynamic performance from a vehicle may require reasoning through this whole family of solutions—instead of tracking a single line—to operate the vehicle at its true limits.