Smooth path and speed planning for an automated public transport vehicle

This paper presents a path and speed planner for automated public transport vehicles in unstructured environments. Since efficiency and comfort are two of the key issues in promoting this kind of transportation system, they are dealt with explicitly in the proposed planning algorithm. To that end, a global path planner has been designed with bounded continuous curvature and bounded curvature derivative to ensure smooth driving. This will allow the public transport system to know a priori which is the shortest path within a selected area that guarantees lateral accelerations and steering wheel speeds below given pre-set thresholds. A closed-form speed profiler uses semantic information provided by the path planner to set a continuous velocity reference that takes into account not only bounds on lateral and longitudinal accelerations consistent with comfort, but also a bound on longitudinal jerk. The suitability of the above two features was compared to manual driving in a real instrumented public transport vehicle on a test track.