V2VLoc: Robust GNSS-Free Collaborative Perception via LiDAR Localization

TL;DR

This paper introduces V2VLoc, a GNSS-free collaborative perception framework using LiDAR localization, featuring a lightweight pose generator and a confidence-aware transformer, validated on a new simulation dataset and real-world data.

Contribution

The paper presents a novel GNSS-free localization framework with a new dataset, including a Pose Generator with Confidence and a Pose-Aware Spatio-Temporal Alignment Transformer.

Findings

Achieves state-of-the-art performance in GNSS-denied environments.

Demonstrates effectiveness of PASTAT on real-world datasets.

Provides a versatile simulation dataset for localization and detection.

Abstract

Multi-agents rely on accurate poses to share and align observations, enabling a collaborative perception of the environment. However, traditional GNSS-based localization often fails in GNSS-denied environments, making consistent feature alignment difficult in collaboration. To tackle this challenge, we propose a robust GNSS-free collaborative perception framework based on LiDAR localization. Specifically, we propose a lightweight Pose Generator with Confidence (PGC) to estimate compact pose and confidence representations. To alleviate the effects of localization errors, we further develop the Pose-Aware Spatio-Temporal Alignment Transformer (PASTAT), which performs confidence-aware spatial alignment while capturing essential temporal context. Additionally, we present a new simulation dataset, V2VLoc, which can be adapted for both LiDAR localization and collaborative detection tasks. V2VLoc comprises three subsets: Town1Loc, Town4Loc, and V2VDet. Town1Loc and Town4Loc offer multi-traversal sequences for training in localization tasks, whereas V2VDet is specifically intended for the collaborative detection task. Extensive experiments conducted on the V2VLoc dataset demonstrate that our approach achieves state-of-the-art performance under GNSS-denied conditions. We further conduct extended experiments on the real-world V2V4Real dataset to validate the effectiveness and generalizability of PASTAT.