Skip to Main Content
In many mobile robot application, navigation is a critical issues. This paper deals with the localization and building problem in an unknown indoor environment. We propose an exploration based on the use of the sensorial data provided by one CCD camera and odometer system. In this paper, the first part of our study is linked to the problem of Vehicle Model and the algorithm model. The second part is devoted to the extracting the features of nature landmark form the CCD image and the matching problem. We can calculate estimation of landmark localization and heading angle with environment maps integrating the previous primitive observations. The third part presents our incremental map building in indoor environment of a non a priori knowledge. We deal with the problem that consists in allowing a robot to localize itself and to construct concurrently a representation of its environment. In the end, this SLAM Algorithm has been implemented with a monocular vision and odometer for indoor environment. The experimental results indicate that the proposed method is feasible and the robot localized coordinate is the higher precision. SLAM algorithm which relies on robot odometer information and nature landmark can retard the positional error growth. The proposed technique is validated in experimental setups.
Image and Signal Processing, 2008. CISP '08. Congress on (Volume:3 )
Date of Conference: 27-30 May 2008