A life-long SLAM approach using adaptable local maps based on rasterized LIDAR images

Waqas Ali, Peilin Liu, Rendong Ying, Zheng Gong

Abstract

Most real-time autonomous robot applications require a robot to traverse through a dynamic space for a long time. In some cases, a robot needs to work in the same environment. Such applications give rise to the problem of a life-long SLAM system. Life-long SLAM presents two main challenges: tracking should not fail in a dynamic environment, and there is a need for a robust and efficient mapping strategy. The system should update maps with new information while keeping track of older observations. However, mapping for a long time can require high computational resources. In this paper, we propose a solution to the problem of life-long SLAM. We represent the global map as a set of rasterized images of local maps along with a map management system responsible for updating local maps and keeping track of older values. We also present an efficient approach using the bag of visual words method for loop closure detection and relocalization. We evaluate the performance of our system on the KITTI dataset and an indoor dataset. Our loop closure system reported recall and precision of above 90 percent. The computational cost of our system is much lower compared to state-of-the-art methods, even for long-term operation.

Introduction

Standard SLAM (Simultaneous Localization and Mapping) algorithms often face a "memory explosion" problem. As a robot operates over extended periods or across multiple sessions in the same area, the map grows indefinitely, leading to massive computational requirements for tracking and loop closure. Furthermore, real-world environments are dynamic—furniture moves, people walk by, and seasons change. A "life-long" SLAM system must be able to update its map with new information while gracefully "forgetting" or culling redundant and outdated observations to keep the system responsive and efficient.

Methodology

The proposed system architecture bridges the gap between 3D LIDAR point clouds and efficient 2D image processing by representing 3D data as a collection of 2D rasterized images.

System architecture overview showing tracking, map management, and loop closure detection modules.

Rasterization and Local Map Representation

Instead of storing raw 3D points, the system projects LIDAR data into 2D grayscale images. The pixel intensity represents the elevation ($z$-value), effectively creating a "top-down" visual summary of a local area. The global map is managed as a graph of these rasterized local maps rather than one giant point cloud.

Conceptual diagram showing the relationship between keyframe nodes, 3D map points, and the generated rasterized image of a local map.

Loop Closure via Bag of Visual Words (BoW)

Typically used in visual SLAM with cameras, the authors apply the Bag of Visual Words (BoW) method to their rasterized LIDAR images. They extract ORB features from these images to build a searchable database. This allows the robot to quickly "query" the database to see if its current view matches a previous local map (loop closure), which is orders of magnitude faster than 3D point cloud matching.

Map Management and Culling

To prevent the map from growing indefinitely, the system includes a culling algorithm. It removes redundant local maps based on distance traveled and updates existing ones when the robot revisits an area, ensuring the "online" map remains lightweight.

Key Results and Insights

The researchers evaluated their approach using the KITTI Odometry Dataset and a custom Indoor Lab Dataset.

  • High Accuracy: The loop closure system achieved recall and precision rates above 90% on the KITTI dataset.
  • Low Computational Footprint: In comparisons with popular frameworks like ORB-SLAM2 and LOAM, this method showed a massive reduction in resource usage. CPU usage remained consistently under 10%, while memory usage stayed below 5-8%, whereas some state-of-the-art systems consumed up to 80% of available memory as the run progressed.
  • Efficiency through Sparsity: By only using a limited number of local map images to build the BoW database—rather than every single keyframe—the search space for loop closure is dramatically reduced without sacrificing accuracy.

Conclusion and Future Work

This work demonstrates that for many navigation tasks, a full 3D point cloud is overkill. By rasterizing the data, the authors leverage highly optimized image-matching algorithms to achieve efficient long-term SLAM. Future research could involve addressing vertical complexity in multi-story environments and integrating semantic labels (e.g., "car," "tree") to further improve the robustness of the map management system.

Summary updated

Abstract

Most real-time autonomous robot applications require a robot to traverse through a dynamic space for a long time. In some cases, a robot needs to work in the same environment. Such applications give rise to the problem of a life-long SLAM system. Life-long SLAM presents two main challenges i.e. the tracking should not fail in a dynamic environment and the need for a robust and efficient mapping strategy. The system should update maps with new information; while also keeping track of older observations. But, mapping for a long time can require higher computational requirements. In this paper, we propose a solution to the problem of life-long SLAM. We represent the global map as a set of rasterized images of local maps along with a map management system responsible for updating local maps and keeping track of older values. We also present an efficient approach of using the bag of visual words method for loop closure detection and relocalization. We evaluate the performance of our system on the KITTI dataset and an indoor dataset. Our loop closure system reported recall and precision of above 90 percent. The computational cost of our system is much lower as compared to state-of-the-art methods. Our method reports lower computational requirements even for long-term operation.
BibTeX
@article{ali2021life,
  author = {Waqas Ali and Peilin Liu and Rendong Ying and Zheng Gong},
  title = {{A life-long SLAM approach using adaptable local maps based on rasterized LIDAR images}},
  journal = {IEEE Sensors Journal},
  year = {2021},
  doi = {10.1109/jsen.2021.3100882},
}
Figure