SemAgent: Semantic-Driven Agentic AI Empowered Trajectory Prediction in Vehicular Networks

TL;DR

The paper tackles inefficient, context-poor trajectory prediction in V2X by marrying semantic communication with Agentic AI. It introduces a multi-agent framework with RSU-based feature extraction and semantic analysis for V2I, and prompts-guided, semantically encoded V2V collaboration that fuses local history with neighbors’ predictions. Through extensive experiments on the NGSIM US-101 dataset, the approach achieves notable improvements in ADE, RMSE, and FDE, including up to 47.5% gains under low SNR. The work demonstrates reduced transmission overhead and robust, context-aware trajectory forecasting in dynamic traffic scenarios, highlighting practical benefits for autonomous driving and smart transportation systems.

Abstract

Efficient information exchange and reliable contextual reasoning are essential for vehicle-to-everything (V2X) networks. Conventional communication schemes often incur significant transmission overhead and latency, while existing trajectory prediction models generally lack environmental perception and logical inference capabilities. This paper presents a trajectory prediction framework that integrates semantic communication with Agentic AI to enhance predictive performance in vehicular environments. In vehicle-to-infrastructure (V2I) communication, a feature-extraction agent at the Roadside Unit (RSU) derives compact representations from historical vehicle trajectories, followed by semantic reasoning performed by a semantic-analysis agent. The RSU then transmits both feature representations and semantic insights to the target vehicle via semantic communication, enabling the vehicle to predict future trajectories by combining received semantics with its own historical data. In vehicle-to-vehicle (V2V) communication, each vehicle performs local feature extraction and semantic analysis while receiving predicted trajectories from neighboring vehicles, and jointly utilizes this information for its own trajectory prediction. Extensive experiments across diverse communication conditions demonstrate that the proposed method significantly outperforms baseline schemes, achieving up to a 47.5% improvement in prediction accuracy under low signal-to-noise ratio (SNR) conditions.

Figures (9)

• Figure 1: V2I communication framework.
• Figure 2: Illustration of the V2I prediction scenario.
• Figure 3: V2V communication framework.
• Figure 4: Illustration of the V2V prediction scenario.
• Figure 5: V2I network diagram. The RSU transmits trajectory-related information within its coverage via semantic communication. The target vehicle performs feature extraction and semantic analysis and then predicts its future trajectory by integrating its own historical trajectory data.