Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Ground-Fusion: A Low-cost Ground SLAM System Robust to Corner Cases

Jie Yin, Ang Li, Wei Xi, Wenxian Yu, Danping Zou

TL;DR

The paper tackles the challenge of robust, low-cost SLAM for ground vehicles in diverse indoor and outdoor environments. It introduces Ground-Fusion, a tightly coupled fusion of RGB-D, IMU, wheel odometry, and GNSS within a sliding-window factor-graph to provide accurate localization with real-time dense color mapping. Key contributions include an adaptive initialization with stationary/visual/dynamic modes, explicit anomaly-detection and handling for wheel, vision, and GNSS faults, and a new multi-sensor ground robot dataset. Experiments on Openloris-Scene and Ground-Challenge demonstrate improved robustness in corner cases and GNSS-challenged outdoor scenarios, highlighting the method's practicality and potential impact.

Abstract

We introduce Ground-Fusion, a low-cost sensor fusion simultaneous localization and mapping (SLAM) system for ground vehicles. Our system features efficient initialization, effective sensor anomaly detection and handling, real-time dense color mapping, and robust localization in diverse environments. We tightly integrate RGB-D images, inertial measurements, wheel odometer and GNSS signals within a factor graph to achieve accurate and reliable localization both indoors and outdoors. To ensure successful initialization, we propose an efficient strategy that comprises three different methods: stationary, visual, and dynamic, tailored to handle diverse cases. Furthermore, we develop mechanisms to detect sensor anomalies and degradation, handling them adeptly to maintain system accuracy. Our experimental results on both public and self-collected datasets demonstrate that Ground-Fusion outperforms existing low-cost SLAM systems in corner cases. We release the code and datasets at https://github.com/SJTU-ViSYS/Ground-Fusion.

Ground-Fusion: A Low-cost Ground SLAM System Robust to Corner Cases

TL;DR

The paper tackles the challenge of robust, low-cost SLAM for ground vehicles in diverse indoor and outdoor environments. It introduces Ground-Fusion, a tightly coupled fusion of RGB-D, IMU, wheel odometry, and GNSS within a sliding-window factor-graph to provide accurate localization with real-time dense color mapping. Key contributions include an adaptive initialization with stationary/visual/dynamic modes, explicit anomaly-detection and handling for wheel, vision, and GNSS faults, and a new multi-sensor ground robot dataset. Experiments on Openloris-Scene and Ground-Challenge demonstrate improved robustness in corner cases and GNSS-challenged outdoor scenarios, highlighting the method's practicality and potential impact.

Abstract

We introduce Ground-Fusion, a low-cost sensor fusion simultaneous localization and mapping (SLAM) system for ground vehicles. Our system features efficient initialization, effective sensor anomaly detection and handling, real-time dense color mapping, and robust localization in diverse environments. We tightly integrate RGB-D images, inertial measurements, wheel odometer and GNSS signals within a factor graph to achieve accurate and reliable localization both indoors and outdoors. To ensure successful initialization, we propose an efficient strategy that comprises three different methods: stationary, visual, and dynamic, tailored to handle diverse cases. Furthermore, we develop mechanisms to detect sensor anomalies and degradation, handling them adeptly to maintain system accuracy. Our experimental results on both public and self-collected datasets demonstrate that Ground-Fusion outperforms existing low-cost SLAM systems in corner cases. We release the code and datasets at https://github.com/SJTU-ViSYS/Ground-Fusion.
Paper Structure (12 sections, 11 equations, 5 figures, 5 tables)

This paper contains 12 sections, 11 equations, 5 figures, 5 tables.

Table of Contents

  1. INTRODUCTION
  2. Related work
  3. Initialization of a multi-sensor SLAM system
  4. Corner cases
  5. Ground robot datasets
  6. Methodology
  7. Adaptive initialization
  8. Multi-sensor state estimator with corner case handling
  9. Experiments
  10. Benchmark tests
  11. Outdoor Tests
  12. Conclusion

Figures (5)

  • Figure 1: The system adopts an adaptive initialization strategy based on the robot's motion state. Potential sensor faults will be detected and handled accordingly. Real-time dense color mapping is supported to facilitate navigation tasks.
  • Figure 2: Estimated and ground-truth (GT) trajectories on part of sample sequences are visualized on the x-y plane.
  • Figure 3: The relative pose errors (m) of each method and the number of effective feature points over time on some visual challenging sequences are plotted.
  • Figure 4: (a) Wheel anomaly analysis and (b) Trajectory of different methods in the $Anomaly$ sequence.
  • Figure 5: (a) ATE RMSE(m) of SLAM systems on wheel anomaly sequences (b) The solid lines denote the value of each method, and dashed lines denote their corresponding thresholds. Grey shading denotes areas where at least two stationary conditions are satisfied.