# Defer to Plan: Adaptive Multi-Agent Fusion for End-to-End V2X Driving

arXiv: https://arxiv.org/abs/2607.19774 · HTML: https://mlsanity.com/paper/2607.19774/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Cooperative and V2X driving, one of 9 directions in End-to-end autonomous driving.
- Builds on 20 of its 24 references that are in the map.
- First posted on arXiv 22 Jul 2026.

## Placed in

- End-to-end autonomous driving / [Cooperative and V2X driving](https://mlsanity.com/driving/cooperative-and-v2x-driving/)

## Builds on (most cited first, 20 of 20)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Planning-oriented Autonomous Driving](https://mlsanity.com/paper/2212.10156/) (2022)
- [Lift, Splat, Shoot: Encoding Images From Arbitrary Camera Rigs by Implicitly Unprojecting to 3D](https://mlsanity.com/paper/2008.05711/) (2020)
- [Multi-Modal Fusion Transformer for End-to-End Autonomous Driving](https://mlsanity.com/paper/2104.09224/) (2021)
- [V2X-ViT: Vehicle-to-Everything Cooperative Perception with Vision Transformer](https://mlsanity.com/paper/2203.10638/) (2022)
- [V2VNet: Vehicle-to-Vehicle Communication for Joint Perception and Prediction](https://mlsanity.com/paper/2008.07519/) (2020)
- [Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline](https://mlsanity.com/paper/2206.08129/) (2022)
- [COOPERNAUT: End-to-End Driving with Cooperative Perception for Networked Vehicles](https://mlsanity.com/paper/2205.02222/) (2022)
- [DriveMoE: Mixture-of-Experts for Vision-Language-Action Model in End-to-End Autonomous Driving](https://mlsanity.com/paper/2505.16278/) (2025)
- [TUMTraf V2X Cooperative Perception Dataset](https://mlsanity.com/paper/2403.01316/) (2024)
- [End-to-End Autonomous Driving through V2X Cooperation](https://mlsanity.com/paper/2404.00717/) (2024)
- [V2XPnP: Vehicle-to-Everything Spatio-Temporal Fusion for Multi-Agent Perception and Prediction](https://mlsanity.com/paper/2412.01812/) (2024)
- [Towards Collaborative Autonomous Driving: Simulation Platform and End-to-End System](https://mlsanity.com/paper/2404.09496/) (2024)
- [HeightFormer: Learning Height Prediction in Voxel Features for Roadside Vision Centric 3D Object Detection via Transformer](https://mlsanity.com/paper/2503.10777/) (2025)
- [HeatV2X: Scalable Heterogeneous Collaborative Perception via Efficient Alignment and Interaction](https://mlsanity.com/paper/2511.10211/) (2025)
- [PillarMamba: Learning Local-Global Context for Roadside Point Cloud via Hybrid State Space Model](https://mlsanity.com/paper/2505.05397/) (2025)
- [RoadFormer : Local-Global Feature Fusion for Road Surface Classification in Autonomous Driving](https://mlsanity.com/paper/2506.02358/) (2025)
- [RoadMamba: A Dual Branch Visual State Space Model for Road Surface Classification](https://mlsanity.com/paper/2508.01210/) (2025)

## Built on by (most cited first, 1 of 1)

- [CoVLM-Bench: A Real-World Benchmark for Cooperative Driving Question Answering and Planning](https://mlsanity.com/paper/2609.35823/) (2026)

## Abstract (arXiv)

Vehicle-to-everything-aided autonomous driving (V2X-AD) significantly enhances driving performance through information sharing. However, existing collaborative perception methods only optimize module-level perception capabilities and fail to effectively serve the ultimate planning and control tasks. We propose an end-to-end collaborative driving system that directly optimizes planning task performance. The system employs MotionNetwork to fuse historical temporal information, utilizes attention mechanisms to efficiently compress spatial features into compact tokens, and adaptively fuses multi-agent features through an autoregressive decoder. Additionally, we introduce Mixture-of-Experts (MoE) architecture to enhance the model's representation capacity for heterogeneous features. Experiments demonstrate that our method achieves a driving score of 79.72, surpassing the state-of-the-art CoDriving baseline (77.15) by 3.33% in closed-loop evaluation while maintaining communication efficiency.