Trust Evaluation of Intelligent Connected Vehicles Based on GNN: Multi-feature Fusion and Temporal Modeling

In the realm of intelligent connected vehicles, trust evaluation is critical for maintaining communication security and system reliability. However, existing GNN-based methods struggle to capture the dynamic evolution of trust relationships. To address this challenge, we propose TrustGA, which integrates vehicle context, network structure, and interaction features, and uses temporal snapshot analysis to model trust relationships over time. TrustGA adopts a hierarchical architecture—comprising a snapshot input layer, a spatial feature aggregation layer, an LSTM-based temporal modeling layer, and a trust prediction layer—and fuses multidimensional features to enhance trust evaluation accuracy. Experimental results on a simulated dataset demonstrate that TrustGA outperforms existing GNN-based trust evaluation models.