PhD Cooperative autonomous driving vehicles: where are we?

For a vehicle to function autonomously, it should be fully aware of its surroundings, own location and state. For platooning, vehicles in addition need the location and state of (at least) their predecessor. When considering both longitudinal and lateral control in platooning, this becomes even more relevant in order to avoid corner cutting.

For determining position information, sensors signals (e.g. RADAR, camera, IMU) on the ego vehicle have to be fused. One therefore needs to create an observer that fuses all relevant signals incorporating the (sometimes poor) reliability of sensor outputs. Subsequently, this needs to be combined with communicated trajectories and positions of upstream vehicles to plan a coherent path to be followed. In addition, as the path to be followed is dependent on the road layout, it should not only be coherent but also feasible. Finally, the generated path should also be tracked within the physical boundaries imposed by the road layout.

This research should answer the following two questions:

Based on practical limitations, e.g., road layout, bounds should be determined indicating the performance of the closed loop control system. This research will make it possible to achieve real-world longitudinal and lateral cooperative driving and provides both a theoretical and practical foundation. From this, various research opportunities arise, e.g., the extension to heterogeneous platoons. By eliminating corner cutting it sets the foundations for achieving cooperative driving in urban environments, which is still an open research area.

For practical implementation, several cooperative automated vehicles are available in the Automotive Lab , which have successfully been used in cooperative driving experiments.