Publication

Abstract : This paper presents the design and implementation of a visual SLAM (Simultaneous Localization and Mapping) loop closure system for autonomous navigation using BRISK (Binary Robust Invariant Scalable Keypoints) feature descriptors on a TortoiseBot platform. The proposed system integrates visual odometry and IMU data to construct a pose graph, where loop closure constraints are established through keyframe matching based on BRISK features. The entire pipeline is developed using MATLAB and deployed on a ROS2-enabled mobile robot equipped with a monocular camera and 9-axis IMU. Loop closures are detected by comparing current frames with historical keyframes, enabling pose graph optimization that corrects accumulated drift and improves localization accuracy. Experimental validation in both simulated and real indoor environments show improved path alignment and robustness. The method is computationally efficient, making it suitable for low-cost robots in indoor navigation tasks. The approach also lays the groundwork for future enhancements involving dynamic obstacle handling and collaborative SLAM.