Social Attention Network Fused Multipatch Temporal–Variable-Dependency-Based Trajectory Prediction for Internet of Vehicles

Vehicle trajectory prediction (VTP) is important for ensuring safe decision-making and planning in Internet of Vehicles (IoV). In complex traffic scenarios, accurate and reliable trajectory prediction requires comprehensive understanding of the interaction behaviors among vehicles. However, existing methods fail to effectively capture vehicle interaction features and fully explore their potential dependencies, limiting improvements in prediction accuracy. To this end, we propose a social attention network fused multipatch temporal–variable dependency (SAN-FTVD) model to tackle the above problems. In specific, we first design a variable token embedding module (VTEM) to extract the motion state information of vehicles, which independently embeds each variable of vehicle historical data into a variable token. After that, we propose a physical informed vehicle interaction encoder (PI-VIE) to capture vehicle interaction features over continuous time. The encoder is combined with physical priors to encode vehicle interaction features based on the correlations between the variable tokens. Following that, a temporal–variable dependency fusion module (TVDFM) is proposed to extract and fuse the multipatch temporal and variable dependencies, fully exploring potential dependencies in vehicle interaction features. Numerical results demonstrate that compared with the state-of-the-art model, the proposed model reduces the average prediction root mean square error over 5-s time range by 8% and 7% on two public data sets with 75% less inference cost. Furthermore, extensive ablation experiments validate the effectiveness of the above modules in the model.

I. Introduction

Vehicle trajectory prediction (VTP) is a crucial component in the decision-making and planning processes for Internet of Vehicles (IoV) [1], [2], [3]. The precise VTP provides a reliable support for selecting optimal driving routes and identifying traffic risks in advance, ensuring safe and efficient vehicle driving [4], [5], [6]. In actual traffic scenarios, a vehicle’s future trajectory is influenced by the historical interaction behaviors between itself and surrounding vehicles. For example, when a vehicle intends to change lanes, lane change maneuvers will only be performed if the vehicle maintains a safe distance from surrounding vehicles within an observable period. This requires capturing the interaction features of vehicles over continuous time when predicting their trajectories [7]. However, with the development of modern transportation, the increase in the number of vehicles and the diversification of road conditions have made vehicle interaction behaviors more complex, posing greater challenges for VTP [8], [9].