This work presents a deep learning approach to classify road users as pedestrians, cyclists or cars using a lidar sensor for detection and a radar sensor for classification. A dataset was gathered on urban roads to train and test the deep learning algorithms. The results show that the system reliably classifies cars, but has trouble with pedestrians and cyclists. The results are improved after aggregating decisions with a Bayes filter. Overlapping targets remain a challenge for the system.

Increased complexity and severity of future driver assistance systems demand extensive testing and validation. For driver assistance functions the perception of the sensors is crucial. Therefore, sensors also have to be modeled. In this contribution, a statistical data-driven sensor-model, is described. The method is widely applicable and able to adapt to complex behavior. As exemplary implementation, a model of an automotive radar system, using a high precision measurement system, is presented.