Simultaneous Localization and Map building paradigm based on omnidirectional stereoscopic vision

This paper deals with the localization and map building paradigm in an unknown indoor environment. We propose an exploration method based on the use of the sensorial data provided by an omnidirectional stereoscopic vision system. The first part of our study is linked with the problem of sensorial model construction with two omnidirectional images obtained by a rigid translation along a rail of our SYCLOP sensor. We propose an approach based on the fusion of several criteria which is realized according to Dempster-Shafer rules. The second part is devoted to the matching problem of the stereo sensorial model with an environment map integrating all the previous primitive observations. We propose two matching approaches based on different selection criterion: the Hausdorff distance and the cumulated cartesian distance. The third part presents our incremental map building paradigm based on the hypothesis of a non a priori knowledge. We deal with the problem wish consists in allowing a robot to localize itself and to construct concurrently a representation of its environment. In this part we discuss the problem of interaction between the localization stage and the mapping stage.