TRAJECTOGRAPHY ESTIMATION FOR A SMART POWERED WHEELCHAIR ORB-SLAM2 VS RTAB-MAP A PREPRINT

This work is part of the ADAPT project relating to the implementation of a trajectography functionality that aims to measure the path travelled by a patient during the clinical trials. This system (hardware and software) must be reliable, quickly integrable, low-cost and real-time. Therefore, our choices have been naturally made towards visual SLAM-based solutions coupled with an Intel real-sense consumer sensors. This paper is a comparison of two well-known visual SLAM algorithms in the scientific community, ORB-SLAM2 and RTAB-Map, evaluated in different path configurations. The added value of our work lies in the accurate estimation of the trajectories achieved through the use of a VICON motion capture system.