top of page
Take your time exploring the contents
Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
RTAB-MAP SLAM
Date
October 2023 - December 2023
- Project Overview
Implemented and evaluated RTAB-Map (Real-Time Appearance-Based Mapping) SLAM algorithm across three distinct environments using ROS (Robot Operating System). The project demonstrated practical applications of simultaneous localization and mapping (SLAM) techniques using various sensor configurations and environmental conditions.
Key Implementations
Developer Dataset: Successfully implemented SLAM using a combination of RGB-D camera, LiDAR, IMU, and wheel odometry sensors in an indoor corridor environment
Northeastern Outdoor Dataset: Adapted RTAB-Map for autonomous vehicle applications using stereo camera data collected from the NUANCE autonomous car on Newbury Street, Boston
Northeastern Indoor Dataset: Deployed SLAM using iPhone LiDAR and camera sensors in Snell Library basement, achieving successful 3D mapping and loop closure detection
Technical Highlights
Integrated multiple sensor inputs including stereo cameras, LiDAR, IMU, and wheel odometry
Implemented real-time loop closure detection and graph optimization
Developed solutions for visual odometry and feature matching using SURF (Speeded Up Robust Features)
Generated and optimized 2D projections, occupancy grid maps, and 3D point cloud representations
Modified and debugged ROS launch files and handled package dependencies
Results & Analysis
Successfully generated accurate 3D maps with loop closure detection in indoor environments
Analyzed system performance under various lighting conditions and environmental challenges
Evaluated mapping accuracy and identified key factors affecting SLAM performance
Documented comparative analysis between indoor and outdoor implementations
Technologies Used
ROS (Robot Operating System)
RTAB-Map
C++
Visual Odometry
Point Cloud Processing
3D Mapping
Sensor Fusion
bottom of page