Evaluation of Lightweight Local Descriptors for Level Ground Navigation with Monocular SLAM

Mobile robots play an important role in Ambient Assisted Living (AAL) by supporting or guiding people with reduced mobility to move in an indoor environment. Visual SLAM algorithms have become an important component of such robots by largely reducing the cost of tracking components. These AAL robots represent a typical situation in which robots move on level ground with merely in-plane navigation tasks. In order to find an optimized configuration of monocular SLAM systems in level ground navigation scenarios, we compared different lightweight local descriptors (LDB, BRIEF and ORB) by evaluating their influence on system performance based on the framework of ORB-SLAM. The results indicate that BRIEF outperforms others in metrics like time and trajectory accuracy, while LDB provides best descriptor matching quality. To conclude, BRIEF would be preferred for indoor level ground navigation with a monocular SLAM system, and LDB can be used instead if matching quality is the primary concern.