Map2Schedule: An End-to-End Link Scheduling Method for Urban V2V Communications

Lihao Zhang, Haijian Sun, Jin Sun, Ramviyas Parasuraman, Yinghui Ye, Rose Qingyang Hu

Abstract

Urban vehicle-to-vehicle (V2V) link scheduling with shared spectrum is a challenging problem. Its main goal is to find the scheduling policy that can maximize system performance (usually the sum capacity of each link or their energy efficiency). Given that each link can experience interference from all other active links, the scheduling becomes a combinatorial integer programming problem and generally does not scale well with the number of V2V pairs. Moreover, link scheduling requires accurate channel state information (CSI), which is very difficult to estimate with good accuracy under high vehicle mobility. In this paper, we propose an end-to-end urban V2V link scheduling method called Map2Schedule, which can directly generate V2V scheduling policy from the city map and vehicle locations. Map2Schedule delivers comparable performance to the physical-model-based methods in urban settings while maintaining low computation complexity. This enhanced performance is achieved by machine learning (ML) technologies. Specifically, we first deploy the convolutional neural network (CNN) model to estimate the CSI from street layout and vehicle locations and then apply the graph embedding model for optimal scheduling policy. The results show that the proposed method can achieve high accuracy with much lower overhead and latency.

The Urban V2V Scheduling Challenge

In dense urban environments, V2V communication is a chaotic game of interference. To maximize the "sum-rate" (the total data capacity of all active links), the network must decide which vehicle pairs should transmit and which should stay silent at any given moment.

This is traditionally a "combinatorial integer programming" problem—a complex puzzle that gets exponentially harder as more cars are added. To solve it, systems usually require perfect Channel State Information (CSI), which describes exactly how signals propagate through the environment. In high-speed city traffic, measuring CSI for every possible interfering link is too slow and overhead-heavy to be practical. Existing fast heuristics often rely only on distance, which fails in cities where buildings frequently block signals.

The Map2Schedule Framework

The authors propose a two-stage deep learning approach that bypasses the need for real-time CSI measurements by using the physical environment as a proxy for signal behavior.

The Map2Schedule architecture pipeline, featuring a dual-UNet structure for radio map generation followed by a graph embedding model for link scheduling.
  1. CSI Estimation (RadioUNet): The first stage uses a modified RadioUNet architecture. It takes a 2D city map and vehicle coordinates as input to predict "radiomaps"—visual representations of signal strength. The model is designed to avoid common CNN artifacts like "black edges" (mispredicted zero gain near buildings) to ensure the predicted CSI is physically plausible.
  2. Link Scheduling (Graph Embedding): The predicted CSI is fed into a graph neural network based on structure2vec. By treating each V2V link as a node in the graph, the model propagates interference information across the network to decide the binary state (on or off) for each link.
  3. Weighted Training: To handle the rarity of "good" links in dense interference scenarios, a Weighted Binary Cross Entropy (BCE) loss function is used, forcing the model to prioritize accuracy on high-performing links.

Experimental Evaluation

The model was evaluated using the RadioMapSeer dataset, which contains 56,000 radiomaps across 700 city layouts. Performance was compared against the FPLinQ algorithm, a high-performing but computationally expensive iterative method used as the ground truth.

Sum-rate performance comparison between Map2Schedule (M2S) and distance-based methods (DF_FP) across different link counts and distance ranges.

Key Results

  • Near-Optimal Performance: Map2Schedule achieved over 90% of the sum-rate performance of the physical-model-based baseline.
  • Massive Speedup: The system can schedule 50 V2V links in just 0.2 seconds, representing a 100x speedup over traditional iterative algorithms.
  • Superiority over Distance-based Models: While simple distance-based scheduling only achieves about 50% of potential capacity in urban settings, Map2Schedule’s map-aware approach bridges the gap to near-optimality.
  • Scalability and Transferability: The graph embedding model uses shared weights, allowing it to handle arbitrary link densities. Furthermore, models trained on short-distance links showed promising transferability to longer-distance scenarios via few-shot learning.

Conclusion and Future Work

Map2Schedule demonstrates that street layout and vehicle geometry are excellent proxies for physical channel measurements. By replacing slow measurements with deep learning inference, it enables real-time, high-capacity V2V communication. Future research will focus on expanding the model from binary "on/off" states to continuous power control and improving performance across even more diverse global cityscapes where map data may be less precise.

Summary updated

Abstract

Urban vehicle-to-vehicle (V2V) link scheduling with shared spectrum is a challenging problem. Its main goal is to find the scheduling policy that can maximize system performance (usually the sum capacity of each link or their energy efficiency). Given that each link can experience interference from all other active links, the scheduling becomes a combinatorial integer programming problem and generally does not scale well with the number of V2V pairs. Moreover, link scheduling requires accurate channel state information (CSI), which is very difficult to estimate with good accuracy under high vehicle mobility. In this paper, we propose an end-to-end urban V2V link scheduling method called Map2Schedule, which can directly generate V2V scheduling policy from the city map and vehicle locations. Map2Schedule delivers comparable performance to the physical-model-based methods in urban settings while maintaining low computation complexity. This enhanced performance is achieved by machine learning (ML) technologies. Specifically, we first deploy the convolutional neural network (CNN) model to estimate the CSI from street layout and vehicle locations and then apply the graph embedding model for optimal scheduling policy. The results show that the proposed method can achieve high accuracy with much lower overhead and latency.
BibTeX
@inproceedings{zhang2024map,
  author = {Lihao Zhang and Haijian Sun and Jin Sun and Ramviyas Parasuraman and Yinghui Ye and Rose Qingyang Hu},
  title = {{Map2Schedule: An End-to-End Link Scheduling Method for Urban V2V Communications}},
  booktitle = {ICC},
  year = {2024},
  url = {https://openreview.net/forum?id=792chLhltx},
}
Figure