Self-supervised Learning of LiDAR Odometry for Robotic Applications (ICRA 2021 Presentation)

Presentation for the IEEE International Conference on Robotics and Automation (ICRA) 2021
Julian Nubert, Shehryar Khattak and Marco Hutter
Paper: arxiv.org/pdf/...
Code: github.com/leg...
Abstract:
Reliable robot pose estimation is a key building block of many robot autonomy pipelines, with LiDAR localization being an active research domain. In this work, a versatile self-supervised LiDAR odometry estimation method is presented, in order to enable the efficient utilization of all available LiDAR data while maintaining real-time performance. The proposed approach selectively applies geometric losses during training, being cognizant of the amount of information that can be extracted from scan points. In addition, no labeled or ground-truth data is required, hence making the presented approach suitable for pose estimation in applications where accurate ground-truth is difficult to obtain. Furthermore, the presented network architecture is applicable to a wide range of environments and sensor modalities without requiring any network or loss function adjustments. The proposed approach is thoroughly tested for both indoor and outdoor real-world applications through a variety of experiments using legged, tracked and wheeled robots, demonstrating the suitability of learning-based LiDAR odometry for complex robotic applications.

[@RonLWilson]

This was very interesting! Best of luck on your follow up work!

[@daegyulee7976]

Brilliant approach and good output. If we progress this research, we can replace conventional registration-approach, i.e. ICP, NDT, Fast-GICP…, with leaning-based-approach. Thank you for sharing your work. I expect I can do the relative work based on your approach.