Highly Automated Driving on Motorways

Object-Level Fusion for Surround Environment Perception in Automated Driving Applications

“…A lane change maneuver, however, is not possible and must be made by the driver. The system is able to detect frontal traffic and the lane using a combination of radar, camera, laser scanner and ultrasonic sensors [29]. The sensors used are integrated into the vehicle body and retain the production look and feel of the vehicle, as opposed to the more complex sensor configurations used in some fully automated driving projects, described in Section 1.2.5.…”

“…29) where p(c k | Z k−1 ) is the predicted classification vector at time k given measurement data up to time k − 1, p(z k | c k ) is the new classification information at the current time k, p(c k | Z k ) is the updated classification vector at time k with new measurement data from time k and η is a normalization factor. The predicted classification vector p(c k | Z k−1 ) can be realized by defining a classification transition probabilities p(c k | C k−1 i…”