Vehicle localization based on Interval Constraint Propagation algorithms

Vehicle localization can be stated as a Constraint Satisfaction Problem (CSP). Interval Constraint Propagation (ICP) algorithms allow us to solve CSPs. This paper discusses the performance of some ICP algorithms applied to vehicle localization. Each of them is fed with the same data obtained with our experimental vehicle equipped with odometers, a gyro and a GPS receiver. Results present the efficiency of these algorithms in terms of heading and positioning accuracy. It is shown that local consistency algorithms do not correct the heading imprecision.