Hybrid Perception in Autonomous Mobile Robots: Integrating 2D and 3D SLAM with ROS2 for Enhanced Navigation and Mapping

The work is moving towards the integration of 2D and 3D Simultaneous Localization and Mapping techniques into autonomous mobile robots using the ROS2 framework. We improve perception and navigation abilities by combining 2D LiDAR capabilities with dual front-facing depth cameras. An approach, consisting of SLAM Toolbox for 2D mapping, RTAB-Map for rich 3D reconstruction, and NAV2 for robust autonomous navigation, is adopted. We introduce a custom URDF model that can effectively combine planar and volumetric data to perform in various environments at structured indoor environments and unstructured outdoor settings. The implementation of RQT graphs that provide details of how nodes interact within the ROS2 ecosystem is demonstrated. The obtained results depict substantially better spatial understanding and adaptability, providing a nice basis for further improvements in mobile robotics within operational scenarios of varying complexity. It is quite efficient to combine the synergy between 2D and 3D SLAM techniques: maps are optimized in terms of navigation efficiency; maps provide context with the environment. It hence opens new perspectives for autonomous systems in complex and dynamic environments. This paper investigates opportunities for fusing 2D and 3D data in real- time for better performance in navigation and object interaction tasks.