Vision/Odometer Autonomous Navigation Based on RatSLAM for Land Vehicles

Abstract

This paper discusses the design of computer vision and odometer integrated autonomous navigation algorithm based on RatSLAM system for land vehicles. The research aims at testing the effectiveness of a vision/odometer bio-inspired navigation approach for land vehicle applications. Images and odometer measurements are inputted into the RatSLAM system. Odometer data can be used to improve the robustness of RatSLAM system significantly. The proposed system has been verified on real vision/odometer data collected in land vehicle tests. Campus round road experiments illustrate that the algorithm is effective and robust.