Cognitive Attention Models for Driver Engagement in Intelligent and Semi-autonomous Vehicles

Lee, John D.; Fisher, Donald; Homaifar, Abdollah; Woods, David · 2018 · ROSA P / Ohio State University. Crash Imminent Safety (CrIS) University Transportation Center

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 accurately estimating the safety benefits of forward collision warning systems (FCWS) in preventing rear-end collisions, which account for 32.4% of all crashes in the United States. The research investigates whether traditional driver models, which assume that behavioral parameters like reaction time, jerk, and deceleration are independent, provide valid safety estimates. The authors hypothesize that these parameters are interdependent due to parallel cognitive processing, and that modeling these associations using copula functions (dependent-stage models) yields more accurate predictions than independent-stage models. The study utilized data from a driving simulator experiment conducted at the National Advanced Driving Simulator. Forty-eight participants engaged in two rear-end collision scenarios: imminent collision with a stopped lead vehicle and with a decelerating lead vehicle. Participants were assigned to either a baseline no-warning condition or a warning condition (visual/auditory or haptic). Driver behavior parameters—brake reaction time, jerk, and maximum deceleration—were extracted from 64 valid events. The researchers developed eight driver models: four independent-stage models sampling parameters from independent distributions, and four dependent-stage models using trivariate copulas to capture the joint distribution of these parameters. These models were applied in counterfactual simulations to estimate crash risk and severity, measured by delta-velocity and Maximum Abbreviated Injury Scale (MAIS) probabilities. The results demonstrated that assumptions of independence between driver model parameters are invalid, as significant associations exist among reaction time, jerk, and deceleration. Counterfactual simulations revealed that copula-based dependent-stage models predicted larger reductions in estimated crash risk and severity compared to independent-stage models. Furthermore, the differential impact of these models varied across different initial event kinematics, indicating that the choice of modeling approach significantly influences safety benefit estimates. The dependent models provided a more nuanced representation of driver behavior, capturing the co-occurrence of behavioral outcomes that independent models oversimplify. The significance of this work lies in its implication for the design and evaluation of vehicle safety systems. The findings suggest that assuming parametric independence in driver models is an oversimplification that may lead to inaccurate safety benefit estimations. By adopting copula-based dependent-stage models, researchers and engineers can achieve more valid and generalizable estimates of FCWS effectiveness. This approach has broader implications for designing vehicle automation algorithms and improving driver safety, particularly in semi-autonomous vehicles where understanding the interdependence of cognitive and behavioral responses is critical for effective driver re-engagement and collision avoidance.

Key finding

Copula-based dependent-stage driver models predict larger reductions in estimated crash risk and severity than independent-stage models.

Methodology

simulator

Sample size: 48

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 bulk_ingest_rosap on 2026-05-23 (6 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.

• situational awareness
• anticipation
• braking response
• mental model of traffic
• automation surprise
• telematics crash prediction

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