GSoC 2026 Project7 | LiDAR+IMU+GNSS Loosely Coupled SLAM Package based on PCL

[Galaxylee319]

Hi dora-rs mentors (RuPing Cen, Zhen Tian),

I am a SLAM developer currently developing a LiDAR+IMU+GNSS loosely coupled SLAM package based on PCL library, and I want to integrate it into dora-rs as a modular operator for GSoC 2026 Project7.

My core work (in progress):

1. PCL-based LiDAR point cloud preprocessing (downsampling, filtering)
2. Loosely coupled fusion of LiDAR + IMU + GNSS
3. Dora-rs dataflow operator packaging
4. Simulation-based testing (comply with dora-rs requirements)

I have read the GSoC 2026 wiki and CONTRIBUTING guide. I want to:

1. Contribute my SLAM package to dora-rs/DORA_NAV
2. Get your feedback on the technical solution
3. Submit a formal proposal for this project

Could you give me some guidance on the integration specification of dora-rs operators?

Looking forward to your reply!

[RuPingCen]

Thank you so much for your interest in the dora-rs GSoC 2026 Project 7. For the multi-sensor fusion part, you can definitely use EKF or other classical state estimation fusion algorithms to fuse IMU, LiDAR odometry, and GNSS data. This loosely coupled architecture is well-suited for the modular design of dora-rs.

Regarding test data, we suggest starting with the NCLT dataset for validation. It includes synchronized LiDAR odometry, IMU, and RTK ground truth trajectories, which will support reliable testing and comparison for your fusion module.

For result visualization and debugging, we highly recommend using Rerun — it’s fully compatible with dora-rs and ideal for visualizing point clouds, fused odometry, trajectories, and sensor data in real time, making your SLAM module’s performance intuitive and verifiable.

[RuPingCen]

The NCLT dataset can be found in this link. https://robots.engin.umich.edu/nclt/