UnDeepLIO: Unsupervised Deep Lidar-Inertial Odometry

TL;DR

UnDeepLIO introduces an unsupervised deep learning framework for lidar odometry that uniquely incorporates IMU data, utilizing siamese LSTMs and attention modules to estimate vehicle pose without supervision.

Contribution

The paper presents a novel unsupervised lidar odometry method that integrates IMU data and employs attention mechanisms for residual pose estimation, which is a new approach in deep lidar odometry.

Findings

Achieves comparable performance to state-of-the-art methods on KITTI benchmark.

Introduces an unsupervised loss function based on voxelized point clouds.

Effectively combines IMU data with deep learning for odometry estimation.

Abstract

Extensive research efforts have been dedicated to deep learning based odometry. Nonetheless, few efforts are made on the unsupervised deep lidar odometry. In this paper, we design a novel framework for unsupervised lidar odometry with the IMU, which is never used in other deep methods. First, a pair of siamese LSTMs are used to obtain the initial pose from the linear acceleration and angular velocity of IMU. With the initial pose, we perform the rigid transform on the current frame and align it closer to the last frame. Then, we extract vertex and normal features from the transformed point clouds and its normals. Next a two-branches attention modules are proposed to estimate residual rotation and translation from the extracted vertex and normal features, respectively. Finally, our model outputs the sum of initial and residual poses as the final pose. For unsupervised training, we introduce an unsupervised loss function which is employed on the voxelized point clouds. The proposed approach is evaluated on the KITTI odometry estimation benchmark and achieves comparable performances against other state-of-the-art methods.