Omni-Directional Multi-Sensor Map Building System
Loading...
Authors
Saoudi, Anass
El-Sayed, Ahmed
Issue Date
2025-04-04
Type
Other
Language
en_US
Keywords
Multi sensor fusion , Autonomous navigation system , Real time SLAM
Alternative Title
Abstract
Autonomous navigation in new environments remains a challenge for modern vehicles due to conventional locomotion constraints. This research introduces an Omni-Directional Multi-Sensor Map Building System that integrates LiDAR and an Intel RealSense RGB-D camera with advanced SLAM techniques to enhance mapping accuracy, perception, and obstacle detection. A novel data fusion strategy combines depth and point cloud data, improving resolution and environmental awareness. Designed for challenging conditions, the system ensures accurate localization even in feature-sparse, poorly lit, or reflective environments. Powered by a Jetson Xavier NX and ROS, it processes sensor data in real time and controls omnidirectional motion via Mecanum wheels, enabling smooth navigation over complex terrains. Extensive testing validates its ability to generate high-resolution maps and maintain localization in scenarios where traditional SLAM systems struggle. This work highlights the advantages of multi-sensor integration for robust and reliable autonomous navigation.
Description
UB Rise 2025
Department of Electrical and Computer Engineering