A labeled PHD filter for extended target tracking in lidar data using geometric invariance properties: Vehicular application

In the field of road safety, a key problem concerns the advanced driver assistance systems (ADAS). One of the objectives is to define systems that monitor the vehicle's environment and inform the driver about the surrounding vehicles. The final idea is to warn him about potentially hazardous situations. In this context, the scanning laser rangefinder is a very popular sensor. It has the advantage of providing distributed information about the relative distance with a high measurement rate. This paper is dedicated to the definition of a method for joint detection and tracking of extended targets (vehicles) based on the use of the geometric invariants of these objects. The proposed solution fits into the framework of sequential Monte Carlo methods and the multi-target aspect is addressed with the PHD theory. The proposed solution is then validated on synthetic and real sequences of road scenarios.