DELO: Deep Evidential LiDAR Odometry using Partial Optimal Transport

TL;DR

DELO introduces a real-time deep learning method for LiDAR odometry that uses partial optimal transport for robust frame matching and jointly learns predictive uncertainty to enhance safety and accuracy.

Contribution

It proposes a novel deep learning framework combining partial optimal transport and uncertainty estimation for improved LiDAR odometry.

Findings

Real-time performance (~35-40ms per frame) achieved.

Competitive accuracy on KITTI dataset.

Uncertainty estimates improve pose-graph optimization.

Abstract

Accurate, robust, and real-time LiDAR-based odometry (LO) is imperative for many applications like robot navigation, globally consistent 3D scene map reconstruction, or safe motion-planning. Though LiDAR sensor is known for its precise range measurement, the non-uniform and uncertain point sampling density induce structural inconsistencies. Hence, existing supervised and unsupervised point set registration methods fail to establish one-to-one matching correspondences between LiDAR frames. We introduce a novel deep learning-based real-time (approx. 35-40ms per frame) LO method that jointly learns accurate frame-to-frame correspondences and model's predictive uncertainty (PU) as evidence to safe-guard LO predictions. In this work, we propose (i) partial optimal transportation of LiDAR feature descriptor for robust LO estimation, (ii) joint learning of predictive uncertainty while learning odometry over driving sequences, and (iii) demonstrate how PU can serve as evidence for necessary pose-graph optimization when LO network is either under or over confident. We evaluate our method on KITTI dataset and show competitive performance, even superior generalization ability over recent state-of-the-art approaches. Source codes are available.