Motion Planning for Automated Vehicles

As autonomous vehicles drive in more challenging environments, it is important that they plan and follow paths that are safe and feasible considering the vehicle's dynamics. We are incorporating knowledge of vehicle handling, road conditions and sensor limitations into the motion planning and control problem. For example, if the vehicle is traveling on a snowy road, we want to be able to plan trajectories that accurately model the limited amount of force between the tires and the road. Or, if a vehicle is traveling behind a large truck which blocks the vehicle's sensors from detecting the presence or absence of another vehicle on part of the road, we want to model this uncertainty and potentially plan information-seeking trajectories to reduce our uncertainty. We are also interested in how to interface with higher level planners, and how to appropriately merge the tasks of motion planning and control.