Vehicle localization by using a multi-modality top down approach

In this paper¹, a localization system for a mobile robot is proposed, using a top-down multi-sensorial approach and exploiting a map of the environment. Generally, the data sensors are associated with the map by a classical map-matching process. Because of the embedded sensors, the field of view is limited, there is a risk of false association between map and the sensor data. Popular methods try to optimize a global cost, to track multi-hypothesis or reduce the problem by using multi-sensor. These approaches are bottom-up: each sensor data is analysed even if it is not relevant (like a GPS in indoor environment). The proposed approach is based on a classical EKF combined with a Bayesian network is used to select the best feature to detect in the map with the best sensor. This selection is done by taking into account actual localization and the objectives of precision and integrity of the robot localization. Presented results show a real-time application of this method with a robot embedding several laser range-finders and a low-cost GPS. Both simulation and real data results are presented.