Segmentation Object - Tracking Object - Classification Ego - motion - estimation

Pedestrian protection systems will become an essential safety function in future vehicles as the government will force automotive manufacturer to improve pedestrian safety within the next few years. For such systems the ability to recognise pedestrians is mandatory. This includes detection, tracking and classification of these vulnerable road users. This paper presents a prototype system for pedestrian recognition that has been tested in a series passenger car. For environment sensing a high resolution Multilayer-Laserscanner with a horizontal field of view of about 180 degree is used. It offers a range image of the environment in front of the ego vehicle. Based on these range images, new pedestrian recognition algorithms were developed. Using these algorithms, moving pedestrians can be classified and tracked from scan to scan, even at large distances. Moreover, their movement can be predicted for the future allowing an assessment whether the situation is dangerous or not.