Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Comparative Evaluation of RGB-D SLAM Methods for Humanoid Robot Localization and Mapping

Amirhosein Vedadi, Aghil Yousefi-Koma, Parsa Yazdankhah, Amin Mozayyan

TL;DR

The paper addresses the challenge of robust localization and mapping for humanoid robots by benchmarking three RGB-D SLAM systems—RTAB-Map, ORB-SLAM3, and OpenVSLAM—on the SURENA-V platform using a RealSense D435. It provides a detailed comparison of localization accuracy, loop-closure behavior, and mapping outputs, revealing that ORB-SLAM3 achieves the lowest Absolute Trajectory Error (ATE) while RTAB-Map offers the most versatile mapping outputs (dense maps, OctoMap, and occupancy grids). OpenVSLAM demonstrates loop-closure-based relocalization but can struggle with odometry in feature-scarce or high-motion scenarios. The results inform practitioners about trade-offs between localization precision and mapping richness, guiding algorithm choice for humanoid autonomy depending on task priorities.

Abstract

In this paper, we conducted a comparative evaluation of three RGB-D SLAM (Simultaneous Localization and Mapping) algorithms: RTAB-Map, ORB-SLAM3, and OpenVSLAM for SURENA-V humanoid robot localization and mapping. Our test involves the robot to follow a full circular pattern, with an Intel RealSense D435 RGB-D camera installed on its head. In assessing localization accuracy, ORB-SLAM3 outperformed the others with an ATE of 0.1073, followed by RTAB-Map at 0.1641 and OpenVSLAM at 0.1847. However, it should be noted that both ORB-SLAM3 and OpenVSLAM faced challenges in maintaining accurate odometry when the robot encountered a wall with limited feature points. Nevertheless, OpenVSLAM demonstrated the ability to detect loop closures and successfully relocalize itself within the map when the robot approached its initial location. The investigation also extended to mapping capabilities, where RTAB-Map excelled by offering diverse mapping outputs, including dense, OctoMap, and occupancy grid maps. In contrast, both ORB-SLAM3 and OpenVSLAM provided only sparse maps.

Comparative Evaluation of RGB-D SLAM Methods for Humanoid Robot Localization and Mapping

TL;DR

The paper addresses the challenge of robust localization and mapping for humanoid robots by benchmarking three RGB-D SLAM systems—RTAB-Map, ORB-SLAM3, and OpenVSLAM—on the SURENA-V platform using a RealSense D435. It provides a detailed comparison of localization accuracy, loop-closure behavior, and mapping outputs, revealing that ORB-SLAM3 achieves the lowest Absolute Trajectory Error (ATE) while RTAB-Map offers the most versatile mapping outputs (dense maps, OctoMap, and occupancy grids). OpenVSLAM demonstrates loop-closure-based relocalization but can struggle with odometry in feature-scarce or high-motion scenarios. The results inform practitioners about trade-offs between localization precision and mapping richness, guiding algorithm choice for humanoid autonomy depending on task priorities.

Abstract

In this paper, we conducted a comparative evaluation of three RGB-D SLAM (Simultaneous Localization and Mapping) algorithms: RTAB-Map, ORB-SLAM3, and OpenVSLAM for SURENA-V humanoid robot localization and mapping. Our test involves the robot to follow a full circular pattern, with an Intel RealSense D435 RGB-D camera installed on its head. In assessing localization accuracy, ORB-SLAM3 outperformed the others with an ATE of 0.1073, followed by RTAB-Map at 0.1641 and OpenVSLAM at 0.1847. However, it should be noted that both ORB-SLAM3 and OpenVSLAM faced challenges in maintaining accurate odometry when the robot encountered a wall with limited feature points. Nevertheless, OpenVSLAM demonstrated the ability to detect loop closures and successfully relocalize itself within the map when the robot approached its initial location. The investigation also extended to mapping capabilities, where RTAB-Map excelled by offering diverse mapping outputs, including dense, OctoMap, and occupancy grid maps. In contrast, both ORB-SLAM3 and OpenVSLAM provided only sparse maps.
Paper Structure (11 sections, 1 equation, 5 figures, 2 tables)

This paper contains 11 sections, 1 equation, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Visual Slam Methods
  3. RTAB-Map
  4. ORB-SLAM3
  5. OpenVSLAM
  6. Experiments And Results
  7. System Overview
  8. Evaluation of Localization
  9. Evaluation in Challenging Settings and Loop Closure
  10. Evaluation of Mapping
  11. Conclusion

Figures (5)

  • Figure 1: (a) World, base, and camera co-ordinate frames of SLAM algorithms (b) Intel® RealSense™ D435 RGB-D camera placement on the SURENA-V humanoid robot
  • Figure 2: Localization outputs of algorithms in comparison with ground truth marker positions in x and y directions
  • Figure 3: Comparison of odometry output of algorithms. As shown, RTAB-Map has successfully tracked the robot's position; however, the other two methods lost track of the robot just after the number of features were diminished in the environment (A). OpenVSLAM, on the other hand was able to relocalize itself in the map after detecting a loop closure (B).
  • Figure 4: Output maps of investigated methods (a) Dense map of RTAB-Map algorithm (b) Sparse map of OpenVSLAM algorithm (c) Sparse map of ORB-SLAM3 algorithm
  • Figure 5: RTAB-Map outputs (a) occupancy grid (b) bird's eye view Dense map