Task planning for highly automated driving

摘要: A hybrid planning approach is presented in this paper with the focus of integrating task and motion for highly automated driving. In context planning, vehicle environment states are transformed from continuous configuration space to a discrete state space. problem solved by search algorithm an optimal sequence reach goal conditions symbolic space, regarding constraints such as topology, place occupation, traffic rules. Each can be mapped specific driving maneuver dedicated method The not only bridges gap between high-level navigation low-level but also provides modular domain description that developed verified individually. Our planner evaluated several scenarios prior knowledge about road-map sensing range vehicle. Behavior otherwise complex achieve planned according rules re-planned on-line perception.