DeepLO: Geometry-Aware Deep LiDAR Odometry

TL;DR

DeepLO introduces a geometry-aware deep learning framework for LiDAR odometry that integrates ICP and supports both supervised and unsupervised training, achieving high accuracy on benchmark datasets.

Contribution

It is the first to combine ICP with deep learning for LiDAR odometry, enabling flexible training modes and improved pose estimation accuracy.

Findings

Outperforms existing LiDAR odometry methods on KITTI and Oxford datasets.

Supports both supervised and unsupervised training modes.

Demonstrates high accuracy and efficiency in pose estimation.

Abstract

Recently, learning-based ego-motion estimation approaches have drawn strong interest from studies mostly focusing on visual perception. These groundbreaking works focus on unsupervised learning for odometry estimation but mostly for visual sensors. Compared to images, a learning-based approach using Light Detection and Ranging (LiDAR) has been reported in a few studies where, most often, a supervised learning framework is proposed. In this paper, we propose a novel approach to geometry-aware deep LiDAR odometry trainable via both supervised and unsupervised frameworks. We incorporate the Iterated Closest Point (ICP) algorithm into a deep-learning framework and show the reliability of the proposed pipeline. We provide two loss functions that allow switching between supervised and unsupervised learning depending on the ground-truth validity in the training phase. An evaluation using the KITTI and Oxford RobotCar dataset demonstrates the prominent performance and efficiency of the proposed method when achieving pose accuracy.