Motion planning based on decentralized sensor data (MODES)

The development of future intelligent mobility in urban traffic presents complex challenges, especially in critical scenarios such as intersections. In this sense, the field of view of the sensors mounted aboard the autonomous vehicle may not be enough to estimate many critical situations (see Fig. 1). To overcome this limitation, new methods of traffic observation and communication between the vehicle and external sensors, V2X-communication, are emerging. This is intended to open up new opportunities to increase the traffic safety of automated and connected driving.

Figure 1. Critical situations at intersections in urban areas

The main objective of this part of the SAVe project is the development of motion planning algorithms for autonomous vehicles that allow to find safe and comfortable trajectories in complex crossings and if necessary to re-plan the trajectories when critical situations occur. The algorithms are based both on the data captured by the sensors of the autonomous vehicle and on external information sent through a wireless communication infrastructure. In this way, the autonomous vehicle can perceive information about objects that are not within its direct field of vision. The validation of the algorithms is subject to selected tests that are carried out in a simulation environment as well as on the CARISSMA test track.
As shown in Fig. 2, the information captured by the traffic observation sensor, by other cars and a HD map is sent to the autonomous vehicle, via V2X-communication, and combined with the information provided by the sensors mounted on the vehicle. This allow the algorithms in the autonomous vehicle to better predict the movement of traffic objects, both static and dynamic. As a result, the traffic scenario can be interpreted better by algorithms and the motion planning leads to safe and comfortable trajectories.

Figure 2. Scheme for the study of algorithms in automated and connected driving

Funding