Driver Models for Both Human and Autonomous Vehicles

Kurt, Arda; Ozguner, Umit; Liu, Peng; Ramyar, Saina; Amsalu, Seifemichael B.; Wang, Zihao; Majd, Keyvan; Homaifar, Abdollah · 2018 · ROSA P / Ohio State University

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 final report details research conducted by The Ohio State University and North Carolina A&T State University to develop computational models of driver behavior for both human and autonomous vehicles. The primary objective was to understand how multi-agent models can inform the design of optimized autonomous systems, particularly for Advanced Driver Assistance Systems (ADAS) and pre-crash safety scenarios. The project aimed to quantify human driving behavior to improve risk assessment, trajectory prediction, and control strategies in mixed traffic environments involving human drivers, semi-autonomous vehicles, and fully autonomous convoys. The research employed two parallel methodological tracks. Ohio State University researchers utilized Hidden Markov Models (HMM) and Hybrid-State Systems (HSS) to model decision-making and trajectory prediction, specifically for lane-change and intersection approach scenarios. They developed classifiers to distinguish between normal and dangerous driving behaviors using naturalistic driving data. North Carolina A&T researchers focused on Support Vector Machines (SVM), Takagi-Sugeno fuzzy models, and personalized driving models. They also developed a simplified matrix formulation for the sensitivity analysis of HMMs to improve computational efficiency. Both groups integrated these behavioral models into control frameworks, including Model Predictive Control (MPC) for vehicle convoys and variational approaches for trajectory planning. Key findings demonstrated that HMM-based classifiers could effectively estimate driver status and predict trajectories for lane-changing vehicles, with dangerous maneuvers exhibiting larger trajectory variances than normal ones. Integrating these predictions into MPC controllers for vehicle convoys resulted in smaller velocity and spacing fluctuations compared to conservative controllers, improving stability during cut-in events. At NC A&T, SVM-based models achieved over 97% accuracy in estimating driver intentions at intersections, outperforming HMMs in generalization. Additionally, a personalized highway driving system successfully adapted to individual driver aggression levels, maintaining safety while respecting user preferences. The proposed simplified matrix formulation for HMM sensitivity analysis reduced computational complexity from quadratic to linear time, enabling more efficient real-time processing. The significance of this work lies in its contribution to the development of safer and more efficient autonomous vehicle systems. By accurately modeling human behavior and integrating these models into control algorithms, the research provides a foundation for ADAS and autonomous vehicles to anticipate human actions, reduce crash risks, and operate smoothly in mixed traffic. The personalized driving models also address the need for autonomous systems to accommodate diverse driving styles, enhancing user acceptance. Furthermore, the improved computational methods for model sensitivity analysis support the real-time implementation of complex behavioral models in vehicle control systems.

Key finding

Behavior-aware Model Predictive Control controllers reduced velocity and spacing fluctuations in vehicle convoys more effectively than conservative controllers when handling cut-in lane changes.

Methodology

modeling

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).

StageOutcomeToolModelPromptAttemptsCompleted
discover success rosap 2 2026-05-23
archive success 1 2026-05-23
extract success cached 2 2026-06-10
clean success 1 2026-06-01
chunk success 1 2026-06-01
embed success 1 2026-06-02
enrich success 1 2026-05-23
promote success 1 2026-05-23
summarize success llm qwen3.6-27b-prismaquant summ-v5 3 2026-06-10
tag success vector_similarity 19 2026-06-11
verify partial 2 2026-06-10

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

Topics

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

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).