Context-aware techniques for real-time decision making in autonomous mobile robots

Abstract

Autonomous robots have become increasingly popular in recent years. Its services have been utilised in various industries such as hospitality, dining, transport & logistics, healthcare and more. In order for us to make better use of them, we have to understand how these robots work. These robots employ the use of Simultaneous Localization and Mapping (SLAM) to navigate through their environment and accomplish tasks. To comprehend the complexities of SLAM algorithms, this paper focuses on the concepts behind VSLAM, the implementation of a select few algorithms and the comparative analysis on the accuracy of these algorithms to identify which one is best suited to be installed in the Delta-NTU Lab’s Carter robot. The methodology used in this paper evaluates the performance of selected algorithms against a publicly available dataset, the EuRoC MAV Dataset, and goes on to test the selected algorithm, ORB-SLAM3, on the data collected by the Carter robot.