M3CAD

Towards Generic Cooperative Autonomous Driving Benchmark

1University of North Texas, 2Toyota InfoTech Labs
Paper Code Dataset CARLA-to-nuScenes
Project Teaser Placeholder

Illustrations of various autonomous driving tasks using the M3CAD dataset. (a), (b), (d), and (e) show sample images captured from the vehicle's four cameras. (a) Demonstrates the path planning (PP) results, where the ego vehicle's predicted trajectory is represented by a dotted line. (b) Shows object tracking (OT) and motion forecasting (MF) results where dotted lines represent predicted trajectories of other vehicles. (c) Presents object detection (OD) results in 3D space. (f) Depicts mapping (MP) and occupancy prediction (OCC) results.

News 📰

  • May 2025: M3CAD dataset released on link.
  • May 2025: M3CAD preprint released on arXiv.

Abstract

We introduce M3CAD, a novel benchmark designed to advance research in generic cooperative autonomous driving. M3CAD comprises 204 sequences with 30k frames, spanning a diverse range of cooperative driving scenarios. Each sequence includes multiple vehicles and sensing modalities, e.g., LiDAR point clouds, RGB images, and GPS/IMU, supporting a variety of autonomous driving tasks, including object detection and tracking, mapping, motion forecasting, occupancy prediction, and path planning. This rich multimodal setup enables M3CAD to support both single-vehicle and multi-vehicle autonomous driving research, significantly broadening the scope of research in the field. To our knowledge, M3CAD is the most comprehensive benchmark specifically tailored for cooperative multi-task autonomous driving research. We evaluate the state-of-the-art end-to-end solution on M3CAD to establish baseline performance. To foster cooperative autonomous driving research, we also propose E2EC, a simple yet effective framework for cooperative driving solution that leverages inter-vehicle shared information for improved path planning. We release M3CAD, along with our baseline models and evaluation results, to support the development of robust cooperative autonomous driving systems.

Benchmark Comparison

M3CAD is comprehensively compared with existing autonomous driving benchmarks. The table highlights differences in Cooperation Type (CT), and supported tasks such as Object Detection (OD), Object Tracking (OT), Mapping (MP), Motion Forecasting (MF), Occupancy Prediction (OCC), and Path Planning (PP).

M3CAD Benchmark Comparison

UniAD Performance Evaluation

We evaluate the UniAD model on M3CAD, comparing its performance across different stages and tasks.

First Stage Performance

Comparison of UniAD's first stage performance on different benchmarks. S.:Single-vehicle tasks, C.: Cooperative tasks.

UniAD First Stage Performance

Second Stage Performance

Comparison of UniAD's second stage performance on different benchmarks. S.:Single-vehicle tasks, C.: Cooperative tasks.

UniAD Second Stage Performance

E2E Framework

This section introduces the End-to-End (E2E) framework used in our project.

E2E Framework

Video 🎬

Visual Comparison

This section visually compares the performance of various autonomous driving tasks under non-cooperative and cooperative settings. Use the buttons below to switch between different tasks (Detection, Mapping, Motion Forecasting, Planning, Occupancy) and observe the improvements achieved through cooperation.

Non-Cooperative vs. Cooperative Task Comparison

Non-Cooperative Detection
Cooperative Detection

BibTeX 🙏

@misc{zhu2025m3cad,
      title={M3CAD: Towards Generic Cooperative Autonomous Driving Benchmark}, 
      author={Morui Zhu and Yongqi Zhu and Yihao Zhu and Qi Chen and Deyuan Qu and Song Fu and Qing Yang},
      year={2025},
      eprint={2505.06746},
      archivePrefix={arXiv},
      primaryClass={cs.RO}
}