Snake-SLAM

Abstract. Snake-SLAM is a scalable visual inertial SLAM system for autonomous navigation in low-power aerial devices. The tracking front-end features map reuse, loop closing, relocalization, and supports monocular, stereo, and RGBD input. The keyframes are reduced by a graph-based simplification approach and further refined using a novel deferred mapping stage to ensure a sparse yet accurate global map. The optimization back-end decouples IMU state estimation from visual bundle adjustment and solves them separately in two simplified sub problems. This greatly reduces computational complexity and allows Snake-SLAM to use a larger local window size than existing SLAM methods. Our system implements a novel multi-stage VI initialization scheme, which uses gyroscope data to detect visual outliers and recovers metric velocity, gravity, and scale. We evaluate Snake-SLAM on the EuRoC dataset and show that it outperforms all other approaches in efficiency while also achieving state-of-the-art tracking accuracy.

Prerequisites

Download and install the following libraries before continuing with the build instructions.

Build Instructions (Ubuntu 20.04, Cuda 11.1)

cd Snake-SLAM
git submodule update --init --recursive

export CXX=clang++-10
export CUDAHOSTCXX=g++-9

mkdir build
cd build
cmake ..
make -j8

Run Snake-SLAM on the EuRoC dataset

Open the file configs/euroc.ini
Update the line [Dataset] -> dir=XX
For example: dir = /ssd2/slam/euroc/MH_01/mav0
Run Snake-SLAM with

cd Snake-SLAM
./build/bin/snake_slam configs/euroc.ini

Notes on Efficiency

At the time of release, Snake-SLAM is the most efficient VI-SLAM system. This can be validated by adjusting the Dataset-playback_fps in the config file.

By default we have set async = false in the config file and which disables all non-deterministic parallelism. Maximum performance will be only be achieved by setting async = true. However this can reduce tracking accuracy, if the playback_fps is unreasonably high. Our recommendation therefore is:

Debugging, Testing, Development

async = false
playback_fps = 200

Deployment, Real-world Usage

async = true
playback_fps = 30

Additional Information

All trajectories used in the paper are included here in euroc_mono_25_runs_each.zip and euroc_stereo_25_runs_each.zip
Feel free to use these trajectories for further analysis or recompute them using the config file configs/euroc.ini
The Snake-SLAM source code is released under the MIT License. You are allowed to use the code in commercial or non-commercial projects, however a reference to this repository and the respective paper (see below) must be made.

Publication

Rückert, Darius, and Marc Stamminger. "Snake-SLAM: Efficient Global Visual Inertial SLAM using Decoupled Nonlinear Optimization." Proceedings of the 2021 International Conference on Unmanned Aircraft Systems (ICUAS), Athen 2021.

https://ieeexplore.ieee.org/abstract/document/9476760

@INPROCEEDINGS{9476760,
  author={Rückert, Darius and Stamminger, Marc},
  booktitle={2021 International Conference on Unmanned Aircraft Systems (ICUAS)}, 
  title={Snake-SLAM: Efficient Global Visual Inertial SLAM using Decoupled Nonlinear Optimization}, 
  year={2021},
  volume={},
  number={},
  pages={219-228},
  doi={10.1109/ICUAS51884.2021.9476760}}