VISTA-DZ: Visual Semantic Trajectory Adaptation for Personalized Dilemma Zone Prediction

2026-06-28

AI Summary

This study addresses the challenge of predicting both the stop/go decisions and their timing for drivers in the dilemma zone at signalized intersections, where individual behavioral differences complicate accurate forecasting. To this end, the authors propose a novel approach that integrates visual semantic trajectory representations with a vision-language model. Historical trajectories are transformed into fine-grained visual semantic depictions to construct personalized driver behavior profiles, which condition a dual-output prediction network to jointly forecast individualized decisions and their execution times. The architecture employs a bidirectional GRU encoder, driver-conditioned multi-head cross-attention, and FiLM modulation to effectively fuse temporal and semantic trajectory features. Evaluated on the SDZ dataset and a newly collected FDZ dataset, the model achieves 93.26% in-domain simulation accuracy and 90.22% average accuracy on held-out drivers, demonstrating strong zero-shot transfer capability from simulation to real-world scenarios.

Abstract

Driver decision making in the dilemma zone at signalized intersections is safety critical, as vehicles approaching a yellow signal must decide whether to stop or proceed within limited time and distance margins. Accurate prediction of both stop-go decisions and decision timing is important for adaptive signal control, advanced driver assistance systems, and human-centered intelligent transportation applications. However, dilemma zone behavior is strongly driver dependent. Similar approach trajectories may lead to different decisions across drivers because of differences in risk preference, braking habit, and decision threshold. Existing personalized models often rely on handcrafted scalar descriptors, which provide useful but limited summaries of individual behavior. This paper proposes VISTA-DZ, a semantic-profile-conditioned framework for personalized stop-go and decision-time prediction. Historical trajectories are converted into visual representations, interpreted by a vision-language model to generate behavioral profiles, and encoded as semantic embeddings to condition a dual-output prediction network. The final model combines a bidirectional GRU encoder, driver-conditioned multi-head cross-attention, and Feature-wise Linear Modulation for temporal evidence selection and feature adaptation. Experiments on the SDZ dataset and a newly collected FDZ dataset show that VISTA-DZ outperforms trajectory-only and handcrafted personalization baselines, achieving 93.26% in-domain simulation accuracy and 90.22% mean accuracy across 20 held-out simulation drivers. Cross-domain results further show feasible zero-shot simulation-to-real transfer and better real-world generalization when simulation data are combined with limited field data.

Problem

• dilemma zone
• personalized prediction
• stop-go decision
• driver behavior
• signalized intersections

Innovation

• visual semantic trajectory
• personalized prediction
• vision-language model
• dilemma zone
• feature-wise linear modulation