Automated vehicles will need to be able to identify their position in the traffic environment accurately – particularly in relation to the challenges of driving scenarios in urban surroundings. A priori information out of digital maps can be used as an additional source of information (virtual sensor) and fused with the detection of the environment by in-vehicle sensors. Digital maps, in combination with vehicle sensors, provide the key to reliable and safe autonomous vehicles.
SP 2Digital map and localisation
Automated driving vehicles shall have demands on digital maps that cannot yet be met by today’s maps. Many key attributes of these maps will differ from the maps currently in use, which have been primarily designed for correct navigation. Automated driving requires maps that, for example, can specify the position of the kerb with an accuracy of just a few centimetres. ‘High-definition’ maps like this can be considered as an additional sensor, which can provide data with a practically unlimited coverage.
Use of the map as an additional sensor and a plausibility check using against data from the in-vehiclesensors are the key to realize automated driving in an urban environment. The use of high definition digital map data as an additional source of information to look beyond sensors and cameras and ultimately see around corners is the optimal way forward. New, innovative approaches for sensor fusion, localisation and plausibility checks are being developed and validated to guarantee a safe performance in the diverse urban scenarios, particularly in relation to the presence of vulnerable road users.
Moreover, the map can be considered as a ‘natural’ format to store all incoming data that is provided by the vehicle’s sensor-based detection of the environment. This format allows all environmental data to be exploited by the algorithms used by the self-driving vehicle – and could thereby form the interface between the real world and the autonomous vehicle.
A major challenge will be to quantify the validity and quality of the data with sufficient accuracy prior to their use.
Sensing the environment and situational understanding
Digital map and localisation
Concepts and pilot applications
Automated driving through urban junctions
Automated driving on urban streets
Interaction with vulnerable road users