Robust Implementation of Automated Collision Avoidance Using an Updating Particle Reference

Övrigt konferensbidrag, 2018

In recent work, an autonomous collision avoidance system was described, targeting speed and path changes based on optimal control theory applied to a friction-limited particle. This avoids artificial separation of the control problem into path-planning and path following, fully integrates speed control and stability criteria with the lateral motion control, and provides a modular framework for motion control at the limits of friction. The present paper extends the previous work by providing an updating reference aimed at improving the robustness and real-world applicability of the controller. The controller is tested and evaluated under conditions of uncertain geometry and surface friction, using a high-fidelity simulation model as well as physical testing on the track.