No abstract

No abstract

This paper presents a method for joint detection and tracking of vehicles with a scanning laser rangefinder. The lidar measurements of an object have the particularity to be spatially distributed, which generally leads to a detection step before any tracking. Differently, the proposed method relies on the raw measurement processing without any detection step, which improves the overall performance in multiobject tracking while providing good estimation accuracies. The solution uses the sequential Monte Carlo methods by incorporating the geometric invariant of the objects of interest (vehicles). This approach also offers an efficient solution to the problem of multitarget tracking by integrating naturally the track management in the filtering process.Index Terms-Extended target tracking, sequential Monte Carlo methods, track-before-detect, scanning laser range finder, advanced driver assistance systems.

The aim of this paper is to derail a fusion niethodfor land vehicle navigation using (parrial) GPS measurement and odomerric information. We focus here on a global esriniariori merhod which uses rhe available GPS signal with partial GPS ourage. The method deals with odometer and pseudo-range measures tliar contain the location parainefers of the vehicle. This estimation problem is solved by a Particle Filter, which has proved ifs benefits compared ro an Extended Kaluian Filter; since it deals wirh non-linear models arid sraristics.The Global Positioning System (GPS) is a key sensor for many navigation systems, and more particularly in the land vehicle navigation applications. In urban environment, a central problem lies in the periods of shading of the GPS receiver's antenna [6]. In order to provide a continuous navigation, the GPS is usually hybrided with additional sensors [9, 13, 11, 141. One of the most popular is the odometer, which is a wheel speed sensor available as standard component in Antilock Braking Systems (ABS). Odometers are electronic devices that generate digital pulses for each revolution of the wheel and allows an estimation of the distance covered by the vehicle.Many works on the subject use the Extended Kalman Filter (EKF) to achieve the GPSlodometer hybridation [I, 121.The EKF performs well in many practical situations, hut the unavoidable linearization stage of the state equations may lead to convergence or instability problems. Recent filtering methods, often merged as Sequential Monte-Carlo Methods [41, avoid these drawbacks due to their ability to deal with non-linear models and statistics. Some works are developed in this case for land vehicle navigation, and more specifically for GPSLlNS hybridation [?I, GPS/ABS hybridation and map matching 151, for instance.In these works, the odometer sensors are only used whenever the GPS measures are not available. The aim of our work is to propose a global modelling of the GPS/odometer fusion problem, with partial or total GPS outages (i.e. with less than four available satellites). The basic idea is to use the pseudo-range measures even if they are not enough to obtain a position estimation. To this end, one develops a particle-based solution that allows a direct integration of the non-linear odometric and pseudo-range measurements. The proposed method is compared with an usual Em-based navigation method with GPS satellites shading. 3-D navigation modelproblem.One deals here with a state modelling of this estimation be the state vector composed by the characteristics parameters of the vehicle. (y", vu3 U ) are respectively, the acceleration, the velocity and the position components along U

No abstract