Laser range finder based obstacle tracking by means of a two-dimensional Kalman filter

This article presents a implementation of a two- dimensional tracking filter in an automotive application. A headway alert system is a considered application for this tracking system, whose key component is a Kalman filter estimating and filtering two-dimensional trajectories of vehicles and various other obstacles that may occur in a traffic scenario. A reliable tracking system is essential for developing a robust headway alert application to identify obstacles and to avoid false warnings. The data, on which the Kalman filter algorithm is based on, is provided by a prototype laser range finder mounted in front of the host car and delivering range and angular data from obstacles in front. These data are filtered and combined with ego-motion data from the host car to decide whether to warn the driver about a potential threat.