Design of a vision-based autonomous robot for street navigation

Abstract

In this study we present the design of an autonomous mobile robot that navigates indoor and outdoor environments using computer vision. System's hardware, software and simulation infrastructures are explained and autonomous navigation algorithms used to reach a given target are described. Successful experimental results are obtained for computer vision algorithms like sidewalk following, visual localization and sidewalk detection from satellite maps, crosswalk detection and path planning. We envision that in the future such a system can be a basis for designing advanced robot systems that help people in their daily lives.