Mixed Signals: A Diverse Point Cloud Dataset for Heterogeneous LiDAR V2X Collaboration

TL;DR

Mixed Signals introduces a large, diverse V2X point cloud dataset from multiple connected vehicles and roadside units, enabling improved perception training and benchmarking for autonomous driving systems.

Contribution

The paper presents a new comprehensive V2X dataset with diverse LiDAR configurations, detailed annotations, and benchmark results, addressing limitations of existing datasets.

Findings

Dataset contains 45.1k point clouds and 240.6k bounding boxes.

Benchmark results highlight current V2X perception method performance.

Dataset ensures reliable, aligned data across multiple viewpoints.

Abstract

Vehicle-to-everything (V2X) collaborative perception has emerged as a promising solution to address the limitations of single-vehicle perception systems. However, existing V2X datasets are limited in scope, diversity, and quality. To address these gaps, we present Mixed Signals, a comprehensive V2X dataset featuring 45.1k point clouds and 240.6k bounding boxes collected from three connected autonomous vehicles (CAVs) equipped with two different configurations of LiDAR sensors, plus a roadside unit with dual LiDARs. Our dataset provides point clouds and bounding box annotations across 10 classes, ensuring reliable data for perception training. We provide detailed statistical analysis on the quality of our dataset and extensively benchmark existing V2X methods on it. The Mixed Signals dataset is ready-to-use, with precise alignment and consistent annotations across time and viewpoints. Dataset website is available at https://mixedsignalsdataset.cs.cornell.edu/.