DriveX 2026 – Foundation Models for Autonomous Driving

Introduction

The 6th edition of the DriveX Workshop focuses on how foundation models and cooperative systems can redefine perception, prediction, planning, and decision-making for autonomous driving and intelligent transportation infrastructure.

Traditional single-vehicle pipelines have achieved impressive progress in 3D detection and tracking, yet they remain constrained by limited viewpoints, occlusions, and domain shifts. Cooperative driving systems, powered by V2X communication and roadside/edge intelligence, extend sensing range, enrich scene context, and enable shared representations across vehicles and infrastructure.

In parallel, foundation models, including vision, vision-language, and multi-modal large models, unlock powerful generalization capabilities: open-vocabulary understanding, scalable pretraining, zero-shot adaptation, and interpretable reasoning about complex road scenes. Emerging end-to-end and agentic systems such as large driving models promise unified perception-to-control frameworks but raise new questions in trustworthiness, reliability, calibration, and evaluation at urban scale.

DriveX 2026 convenes researchers and practitioners from computer vision, robotics, communications, transportation, AI safety, and policy to:

Design foundation models natively aware of cooperative perception and real-world V2X constraints.
Bridge large-scale datasets, simulation, and deployment for multi-agent coordination.
Discuss standards, benchmarks, and open challenges for safe, human-centric cooperative autonomy.

Topics of Interest

Foundation models and world models for cooperative autonomous driving and intelligent transportation systems
Vision-Language(-Action) Models for traffic scene understanding and explanation
LLM- and agent-based support for perception, prediction, planning, and V2X coordination
Cooperative perception, V2X communication, and infrastructure-assisted driving
3D/4D Gaussian Splatting for 3D Scene Reconstruction
Datasets, benchmarks, and evaluation protocols for foundation models and cooperative perception
3D scene understanding and reasoning
End-to-end pipelines and large driving models for multi-agent decision-making
Safety, robustness, uncertainty estimation, and open-set reasoning in cooperative systems
Vehicle-to-Infrastructure (V2I) and Vehicle-to-Everything (V2X) interaction and standards

Schedule (Tentative)

Time	Session
08:00 – 08:10	Opening Remarks – Welcome & Workshop Overview
08:10 – 08:30	Opening Keynote Keynote
08:30 – 09:00	Keynotes 1 - Prof. Dr. Jiaqi Ma (UCLA) Keynote
09:00 – 09:30	Keynotes 2 - TODO Keynote
09:30 – 10:00	Keynotes 3 - TODO Keynote
10:00 – 10:30	Coffe Break
10:30 – 11:00	Keynote 4 - TODO Keynote
11:00 – 11:30	Keynote 5 - TODO Keynote
11:30 – 12:00	Panel I Panel
12:00 – 13:00	Lunch Break & Networking
13:00 – 13:30	Keynotes 6 - TODO Keynote
13:30 – 14:00	Keynotes 7 - TODO Keynote
14:00 – 14:30	Keynotes 8 - TODO Keynote
14:30 – 15:00	Keynotes 9 - TODO Keynote
15:00 – 15:30	Coffee Break
15:30 – 16:00	Keynotes 10 - TODO Keynote
16:00 – 16:30	Panel II Panel
16:30 – 17:30	Oral Presentations 1–6 Oral
17:30 – 18:00	Poster Session II & Coffee Break
18:00 – 18:10	Awards Ceremony – Best Paper, Poster, Keynote
18:10 – 18:20	Awards Ceremony – Challenge Winners
18:20 – 18:30	Closing Remarks & Group Photo
19:00 – 21:00	Workshop Reception & Networking Reception

Final schedule, room allocation, and speaker order will be announced closer to the workshop date.

Invited Keynote Speakers

Prof. Dr. Daniel Cremers

Technical University of Munich, Germany

Dr. Mingxing Tan

Waymo, USA

Dr. Jamie Shotton

Wayve, Vancouver

Prof. Dr. Marco Pavone

Stanford University & NVIDIA, USA

Prof. Dr. Angela Dai

Technical University of Munich, Germany

Prof. Dr. Manmohan Chandraker

University of California San Diego (UCSD), USA

Prof. Dr. Jiaqi Ma

University of California Los Angeles (UCLA), USA

Paper Track

DriveX 2026 invites high-quality contributions on foundation models, V2X-based cooperative perception, large driving models, and related topics outlined above.

We welcome:

Novel full papers (max. 8 pages, excluding references) for publication in the official proceedings.
4-page extended abstracts or 8-10 page versions of recently published work (non-archival, not included in proceedings).

Submissions must follow the official ECCV 2026 style: LaTeX or Typst.

📘 Archival Track (Proceedings Track)

Submit Now

Paper Type: Full papers (max. 8 pages, excluding references)
Included in Proceedings: Yes (archival)
Submission Opens: April 15, 2026
Submission Deadline: June 15, 2026 (23:59 PST)
Notification to Authors: June 30, 2026
Camera-Ready Deadline: July 15, 2026

📗 Non-Archival Track

Submit Now

Paper Type: 4-page extended abstracts or 8-10 page versions of recently published work
Included in Proceedings: No (non-archival)
Submission Opens: April 15, 2026
Submission Deadline: July 1, 2026 (23:59 PST)
Notification to Authors: July 15, 2026
Camera-Ready Deadline: July 31, 2026

Paper Awards

🏆 Best Paper Award: $300
🏆 Best Paper Runner-Up Award: $200
🏆 Best Application Paper Award: $100
🏆 Best Poster Award Certificate
🏆 Best Keynote Presentation Award Certificate

Challenge Awards

🏆 TUMTraf-V2X Challenge Winner Award: $100
🏆 doScenes Challenge Winner Award: $100
🏆 MDrive Challenge Winner Award: $100

Second and third place of each challenge will receive an award certificate.

DriveX Grand Challenge

Track 1 TUMTraf-V2X Challenge Apr 15 - July 1 Track 2 doScenes Challenge Apr 15 - July 1 Track 3 MDrive Challenge Apr 15 - July 1 Track 4 nuReasoning Challenge Apr 15 - July 1 Track 5 HetroD Challenge Apr 15 - July 1

V2I-Based Cooperative Perception
Infrastructure–vehicle fusion using TUMTraf-V2X. Focus on cooperative 3D detection and tracking with infrastructure-mounted LiDAR, radar, and cameras, emphasizing occlusion handling, long-range awareness, and reliability under real-world conditions.
Natural Language Instruction for Human Interaction and Vision-Language Navigation
Built upon doScenes. Participants design models for natural language instructions and visual language navigation to facilitate research on human-vehicle instruction interactions.
Multi-Agent Reasoning
Using MDrive teams explore a cooperative driving benchmark for end-to-end closed-loop multi-agent systems.
Reasoning-intensive Long-Tail Driving Scenarios
The nuReasoning Challenge aims to develop autonomous driving agents that can robustly handle rare interactions and complex causal dependencies, going beyond average-case perception and planning
Traffic Simulation in Heterogeneous Traffic
The HetroD Challenge focuses on multi-agent traffic simulation in highly heterogeneous urban environments, based on HetroD dataset.

Competition Timeline

Competition Announcement: Apr 13, 2026
Competition Start: Apr 15, 2026 (10 weeks)
Submission Deadline: July, 1 (23:59 PST)
Notification to Participants: July 7, 2026

Top-performing teams will be invited to present at the workshop and will receive money prizes ($100) and award certificates. Detailed rules, baselines, and submission instructions are available on the official challenge page.