A Taxonomy and Review of Algorithms for Modeling and Predicting Human Driver Behavior
DOI: 10.1109/jproc.2025.3617487
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the challenge of modeling human driver behavior, a critical requirement for autonomous vehicle planning, safety validation, and causal inference. The authors identify that existing literature is fragmented, with prior reviews focusing narrowly on motion prediction. To unify the field, the paper proposes a comprehensive taxonomy and review of algorithms covering four core tasks: state estimation, intent estimation, trait estimation, and motion prediction. The motivation stems from the stochastic, high-dimensional, and interactive nature of human driving, which complicates prediction for autonomous systems. The authors establish a unifying mathematical framework based on the Partially Observable Stochastic Game (POSG). This framework models multi-agent traffic as a system where agents possess physical states (position, velocity) and internal states (intentions, traits, mental models). Within this structure, driver modeling tasks are cast as inference problems. The review analyzes over 200 papers, classifying them based on the specific tasks they address and key algorithmic attributes. The taxonomy distinguishes between core tasks—such as inferring current physical states or predicting future trajectories—and auxiliary tasks like risk estimation and anomaly detection. The analysis focuses on fundamental model attributes, including architecture, training methods, theoretical underpings, scope, and evaluation metrics, rather than providing quantitative performance comparisons. The findings provide a structured classification of existing models across the four core tasks. For state estimation, the review identifies that while few papers explicitly address this foundational task, methods typically involve approximate Bayesian filters, such as Kalman or particle filters, often enhanced by environmental structure. For intention estimation, the authors categorize models by their intention space (e.g., route, configuration, or special maneuvers), hypothesis representation (e.g., discrete probability distributions or particle distributions), and inference paradigms. The paper highlights how models differ in defining behavior modes, such as lane changes or merging gaps, and how they represent uncertainty regarding driver intentions. The review also details approaches for trait estimation and motion prediction, mapping specific algorithms to their respective positions within the proposed taxonomy. The significance of this work lies in its provision of a common mathematical framework and a comprehensive taxonomy that facilitates understanding of the driver modeling landscape. By unifying disparate literature under the POSG framework, the paper helps researchers identify meaningful connections between existing models and clarifies the relationships between upstream tasks (intent and trait estimation) and downstream tasks (motion prediction). The authors conclude by identifying open research opportunities, aiming to guide future development in creating more robust and interpretable models of human driving behavior for autonomous systems.
Key finding
The study provides a comprehensive taxonomy and review of over 200 algorithms for modeling human driver behavior, unified under a partially observable stochastic game framework.
Methodology
review
Sample size: 200
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 author_sweep_intake on 2026-05-28.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-28 |
| archive | success | canonical_url | — | — | 7 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
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.
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).
- Theoretical Contribution: computational model, theory or model, conceptual framework