Investigations on Collective Perception for Left Turning Manoeuvres

Abstract

At the DLR Institute of Transportation Systems, methods for cooperative perception in the Vehicle-to-X (V2X) context are being developed. The core of these investigations is the derivation of abstract information such as position, speed or size of other road users from sensor data, the distribution of this information and the fusion of data from different sensor sources in real time. That way an automated vehicle obtains the sensor information of other road users as well as further sensor information provided by the traffic infrastructure. This information can compensate for weak spots in the autonomous vehicle's on-board sensors through complementary sensors or different visual perspectives. In addition, the extended Field of View (FOV) also includes more distant areas of the vehicle, for example at the next but not yet immediately visible intersection. This enables automated and networked vehicles to drive more predictively and thus more safely.