TrajAware: Graph Cross-Attention and Trajectory-Aware for Generalisable VANETs under Partial Observations

2025-09-08

AI Summary

To address routing challenges in vehicular ad hoc networks (VANETs) under dynamic topologies, partial observability, and edge resource constraints, this paper proposes a lightweight reinforcement learning framework tailored for edge AI deployment. Methodologically, it innovatively integrates action-space pruning, graph cross-attention mechanisms, and trajectory-aware prediction to achieve strong generalization against incomplete observations and rapid topology changes; additionally, it incorporates intersection-level semantic information and historical trajectory modeling to enable cross-city map transferability. Experimental results on real-world urban maps demonstrate that the approach achieves near-optimal shortest-path performance, significantly outperforms baseline methods in packet delivery ratio, reduces inference latency by 42%, and compresses the model size to 1.8 MB—enabling efficient deployment on resource-constrained embedded edge devices.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Vehicular ad hoc networks (VANETs) are a crucial component of intelligent transportation systems; however, routing remains challenging due to dynamic topologies, incomplete observations, and the limited resources of edge devices. Existing reinforcement learning (RL) approaches often assume fixed graph structures and require retraining when network conditions change, making them unsuitable for deployment on constrained hardware. We present TrajAware, an RL-based framework designed for edge AI deployment in VANETs. TrajAware integrates three components: (i) action space pruning, which reduces redundant neighbour options while preserving two-hop reachability, alleviating the curse of dimensionality; (ii) graph cross-attention, which maps pruned neighbours to the global graph context, producing features that generalise across diverse network sizes; and (iii) trajectory-aware prediction, which uses historical routes and junction information to estimate real-time positions under partial observations. We evaluate TrajAware in the open-source SUMO simulator using real-world city maps with a leave-one-city-out setup. Results show that TrajAware achieves near-shortest paths and high delivery ratios while maintaining efficiency suitable for constrained edge devices, outperforming state-of-the-art baselines in both full and partial observation scenarios.

Problem

• Routing in VANETs with dynamic topologies and partial observations
• Overcoming reinforcement learning's fixed graph structure limitations
• Enabling edge AI deployment under constrained hardware resources

Innovation

• Action space pruning for dimensionality reduction
• Graph cross-attention for global context mapping
• Trajectory-aware prediction using historical routes