Simple visual self-localization for indoor mobile robots using single video camera

Abstract

In this paper, we present a simple method for visual localization of indoor mobile robots based on a natural landmark model. Only two natural lines and an intersection point are used for localization to take the advantage of fast detection. To track the landmark model, Lucas-Kanade optical flow algorithm is applied by using gradient descent. Then, by adopting reasonable assumption for indoor environments, a quick localization method for mobile robots from correspondent landmark is proposed in a linear technique. Experimental results that demonstrate the robustness of the method with respect to image illumination and noises are also presented. The performance in indoor environments shows the feasibility of the proposed localization algorithm in real-time.