Paper Title
Real Time Localization and Mapping of a Mobile Robot using Visual Features

Abstract
This paper proposes a new system for real-time localization and mapping of a mobile robot in unknown environments that will provide robot navigation based on visual features. The Iterative Closest Point Algorithm (ICP) method is employed to estimate the pose of the mobile robot using a stereo camera and a three-step approach. The vision system first identifies characteristic features from two sets of images and matches them by applying the Random Sample Consensus (RANSIC) method. The second step involves building a local 3D map of the robot’s environment using visual Speeded-Up Robust Feature (SURF) to match points and image disparities, thereby allowing robot to localize its position based on motion estimation. Finally, the current visual information and the local maps collected by the robot sensors are used to build a real time global map. The experimental results used to evaluate accuracy of our algorithms are presented and discussed. Keywords - Localization and Mapping, Vision System, Mobile Robot, Unknown Environments.