A New Landmark and Sensor Selection Method for Vehicle Localization and Guidance

Markov localization is one of the effective techniques for determining the physical locations of an autonomous vehicle whose the perceptions of the environment are limited. To improve the localization, a multi-sensor approach is used. A landmark selection process is usually employed. The aim of this selection strategy is to select the landmark that answers at best to a criterion. In general, the selected landmark is the one that improve the most the vehicle's location. In this paper, we extend the landmark selection problem into a resource selection (i.e. sensor and feature detection algorithm) problem. This selection is also based on a criterion. However, this criterion is defined in function of the application's objectives. Here, the application concerns vehicle's guidance. This last one requires an accurate and reliable estimation. Thus, we propose a novel selection strategy of the landmark, the sensor, and the feature detection algorithm to offer an accurate and reliable localization. We demonstrate the practicality of this approach by guiding an experimental vehicle in real outdoor environment.