Multi vehicle trajectory coordination for automated parking

Automated parking will probably be among the first tasks performed by fully autonomous vehicles. To smoothly integrate into the existing infrastructure and traffic scenarios an autonomous vehicle has to make assumptions on the intentions of human drivers and maneuver safely at low speeds in a highly structured environment without Car2X communication. This work presents a trajectory planning and coordination strategy for such situations. Based on the observed situation and an estimated intention of other participants in the scene their driving behavior is quantified and an optimized trajectory plan is calculated using mixed-integer linear programming. It balances own and other intentions and avoids collisions. The approach resolves conflicts occurring in parking situations as demonstrated in two simulated scenarios.