Understanding Driver Cognition and Decision-Making Behaviors in High-Risk Scenarios: A Drift Diffusion Perspective

Huang, Heye; Li, Zheng; Cheng, Hao; Wang, Haoran; Jiang, Junkai; Li, Xiaopeng; Zgonnikov, Arkady · 2025 · arXiv

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of modeling human driver cognition and decision-making in high-risk scenarios to improve interactions between autonomous vehicles (AVs) and human drivers. Existing models, such as behavioral simulations or game-theoretic approaches, often fail to capture the complex, multidimensional nature of risk cognition or lack interpretability. To bridge this gap, the authors propose a cognition-decision framework that integrates individual variability in risk perception with common decision-making mechanisms using the Drift Diffusion Model (DDM). The study employs a driving simulator experiment involving 58 licensed drivers. Participants navigated three high-risk scenarios: cut-in (lateral risk), rear-end collision (longitudinal risk), and lane-changing (multidimensional risk). The experimental design captured key behavioral metrics, including cognitive reaction time, brake reaction time, maximum deceleration, and steering angles. The authors developed two primary models: a risk sensitivity model based on a multivariate Gaussian distribution to quantify individual differences in risk perception, and a DDM to simulate the dynamic accumulation of evidence leading to a decision (brake vs. steer). The DDM parameters—drift rate, decision boundary, initial bias, and non-decision time—were formulated as functions of vehicle kinematics (speed, distance, time headway) and adjusted by the driver’s specific risk sensitivity level. Results indicate that the proposed framework accurately predicts driver behavior across all scenarios. The multivariate Gaussian model successfully characterized individual differences, showing that higher risk levels elicited stronger braking and steering responses with significant variability among drivers. The DDM, calibrated using differential evolution optimization, closely matched empirical cumulative response time probabilities. Specifically, the model demonstrated that higher risk sensitivity increases the drift rate (faster information accumulation) and lowers the decision boundary (requiring less evidence for a decision), while also shifting initial bias toward evasive steering. Comparative analysis showed that the DDM more precisely captures human cognitive processes and adaptive decision-making than traditional models like IDM, Gipps, or MOBIL. The significance of this work lies in providing a theoretically grounded, interpretable framework for modeling human driving behavior in complex environments. By quantifying how individual risk cognition influences decision thresholds and speeds, the model offers critical insights for designing AV systems that can better predict and adapt to human driver behaviors, thereby enhancing safety in mixed traffic environments.

Key finding

A drift-diffusion cognition–decision model with individualised risk-sensitivity parameters predicts driver emergency-maneuver decisions in high-risk scenarios more accurately than IDM/Gipps/MOBIL baselines.

Methodology

lab_experiment

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via discover_arxiv on 2026-05-04 (5 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• risk taking
• decision making risk perception
• mental model of traffic
• situational awareness
• braking response
• gap acceptance

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: behavioral performance data
• Theoretical Contribution: computational model, theory or model