Position estimation of robotic platform using optical flow

Abstract

This paper deals with a base research of an alternative type of navigation for our project ARES (Autonomous Research Exploration System). This system is focused on exploration of unknown areas. The navigation of this platform is based on fusion of several navigation methods. The first method is based on the visual odometry using SURF (Speeded Up Robust Feature). The second navigation method is based on Hector Slam and Lidar sensor. The third method is based on optical flow. The implementation of third method based on the Lucas-Kanade method is descripted in this paper.

Keywords: Robot navigation, Visual Odometry, Lucas-Kanade, Optical Flow