Unsupervised Domain Adaptation for Self-Driving from Past Traversal Features

TL;DR

This paper introduces an unsupervised domain adaptation method for LiDAR-based self-driving object detection, utilizing repeated traversals and historical spatial features to improve detection accuracy in new environments.

Contribution

It proposes a novel adaptation framework that leverages spatial statistics from repeated traversals and a self-training process, enhancing detector performance without labeled data.

Findings

Achieves up to 20-point performance improvement in detection accuracy.

Significantly improves pedestrian and distant object detection.

Framework is detector-agnostic and effective on real-world datasets.

Abstract

The rapid development of 3D object detection systems for self-driving cars has significantly improved accuracy. However, these systems struggle to generalize across diverse driving environments, which can lead to safety-critical failures in detecting traffic participants. To address this, we propose a method that utilizes unlabeled repeated traversals of multiple locations to adapt object detectors to new driving environments. By incorporating statistics computed from repeated LiDAR scans, we guide the adaptation process effectively. Our approach enhances LiDAR-based detection models using spatial quantized historical features and introduces a lightweight regression head to leverage the statistics for feature regularization. Additionally, we leverage the statistics for a novel self-training process to stabilize the training. The framework is detector model-agnostic and experiments on real-world datasets demonstrate significant improvements, achieving up to a 20-point performance gain, especially in detecting pedestrians and distant objects. Code is available at https://github.com/zhangtravis/Hist-DA.