Skip to content

Latest commit

 

History

42 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

Awesome_Dynamic_SLAM Awesome

GitHub stars

Interactive star history · Alternative chart

The following papers focus on SLAM in dynamic environments and life-long SLAM. In dynamic environments, there are two kinds of robust SLAM: first is detection & removal, and the second is detection & tracking. Although mapping in dynamic environments is not my focus, I will also include some interesting articles.

Vision indicates the pipeline is built with a camera. Others are the same, such as lidar, radar, sensor fusion.


2025–2026 Updates

Last updated: August 2026. Entries below link to the paper and the authors' project or code repository when available.

Dynamic SLAM and 4D Mapping

Lifelong Mapping and Localization

Dynamic Object Detection and Tracking

Related Robust Visual Odometry

Datasets and Benchmarks

  • (CVPR 2025) Wild-SLAM — dynamic monocular/RGB-D sequences with mocap ground truth plus in-the-wild iPhone sequences; evaluation and download scripts are provided with WildGS-SLAM.
  • (CVPR 2026) DROID-W Dataset — seven casually captured outdoor sequences with dynamic objects, clutter, RGB frames, and LiDAR-derived ground-truth trajectories; evaluation code.
  • (CVPR 2025) Sim4D — synthetic non-rigid 4D-SLAM benchmark with camera poses, time-varying geometry, color, and rendering scripts.
  • (IROS 2025) M3DGR — multi-sensor ground-robot benchmark with dynamic people and systematically induced visual, LiDAR, wheel, and GNSS degradation; evaluates more than 40 SLAM systems.
  • (ICRA 2025) M3DSS — multi-platform and multi-sensor SLAM dataset with event cameras, accurate ground truth, and dynamic urban driving sequences.
  • (2025) SLAM&Render — 40 synchronized RGB-D, IMU, kinematic, and pose sequences with lighting changes, occlusions, and object rearrangements; benchmark repository.
  • (CVPR 2026) Ghost-FWL — 24K mobile full-waveform LiDAR frames with billions of peak-level labels for ghost detection and removal; dataset and code.
  • (2026) MotionScape — more than 30 hours of 4K UAV video with 6-DoF trajectories and text descriptions under highly dynamic camera motion; benchmark repository (partial public release).
  • (2026) HERCULES — open-source heterogeneous UAV/UGV simulation and a kilometer-scale SLAM benchmark across dynamic desert, forest, and city environments; project page.
  • (ICRA 2025) DiTer++ — multi-robot, multi-modal, multi-session SLAM data across diverse outdoor terrain; dataset and tools.
  • (IEEE T-RO 2025) ROVER — a multi-season outdoor visual SLAM benchmark covering illumination, weather, vegetation, and structural changes; dataset and evaluation code.

Related survey papers:

Dynamic Object Detection and Removal

. “Crowd-SLAM: Visual SLAM Towards Crowded Environments Using Object Detection,” Journal of Intelligent & Robotic Systems 2021

Dynamic Object Detection and Tracking

Other Application

Citation

If you find this repository useful for your research, please consider citing it:

@misc{zhu2021awesome_dynamic_slam,
  author    = {Hu Zhu},
  title     = {{Awesome Dynamic SLAM}: A Curated List of Dynamic and Lifelong SLAM Resources},
  year      = {2021},
  publisher = {GitHub},
  journal   = {GitHub repository},
  url       = {https://github.com/zhuhu00/Awesome_Dynamic_SLAM}
}

About

Dynamic SLAM, Life-long SLAM Research(Lidar, Visual, Sensor Fusion etc.)

Topics

Resources

Stars

715 stars

Watchers

25 watching

Forks

Releases

Packages

Contributors