Human Performance Models of Pilot Behavior
DOI: 10.1177/154193120504901202
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 reviews the development and application of five distinct Human Performance Models (HPMs) to predict pilot behavior and error in aviation safety-critical scenarios. Motivated by the fact that over two-thirds of aircraft accidents are attributed to pilot error, the research addresses the difficulty of studying rare, low-probability errors in field or laboratory settings. The study aims to demonstrate how HPMs can uncover "latent design flaws" and predict error vulnerabilities early in the design cycle, particularly regarding new flight deck technologies. Five modeling teams, selected by the NASA Aviation Safety and Security Program, developed models to address two specific problems: taxiway navigation errors during surface operations and pilot performance during instrument approaches with and without Synthetic Vision Systems (SVS). The methodology involved integrating cognitive models with high-fidelity external environment simulations. Each team utilized empirical data from Human-In-The-Loop (HITL) simulations conducted at NASA Ames Research Center to populate, develop, and validate their models. For surface operations, data included taxi speeds, navigation errors, communications, and workload metrics from trials comparing baseline operations with a prototypical Taxiway Navigation and Situation Awareness (T-NASA) system. For approach and landing, data included eye movements and control responses from part-task simulations involving SVS. The five modeling frameworks employed were: Attention-Situation Awareness (A-SA), which focuses on visual attention allocation and belief updating; ACT-R Version 5.0, a closed-loop cognitive architecture modeling fine-temporal behaviors; Air-MIDAS, which uses working memory limits and heuristics; D-OMAR, an event-based simulator focusing on procedural and habit-based errors; and an integrated IMPRINT/ACT-R approach combining task network simulation with cognitive agent modeling. The results demonstrated that all five models successfully predicted errors and error vulnerabilities arising from situation awareness degradation, memory interference, airport layout complexities, and workload. Specifically, the models identified that taxiway navigation errors were significantly mitigated by the T-NASA display system. Regarding SVS, the models revealed that while SVS did not adversely affect flight safety, it altered pilot scanning behavior. Pilots tended to fixate on redundant information provided by SVS, reducing attention to traditional displays and causing small delays in action initiation during the approach phase. However, no performance degradation was observed in landing or go-around phases. The A-SA model accounted for 85% of the variance in scanning behavior, while ACT-R and Air-MIDAS models highlighted the critical role of environmental fidelity and closed-loop interactions in predicting behavior. The significance of this work lies in validating HPMs as effective tools for the aerospace community to evaluate new technologies and procedures without expensive hardware prototyping. The study concludes that different modeling architectures offer distinct advantages: A-SA and ACT-R are superior for analyzing visual attention allocation, while Air-MIDAS and D-OMAR are better suited for modeling multiple operator interactions and crew dynamics. The findings underscore that HPMs can effectively identify latent design flaws, optimize display information allocation to maintain efficient scan patterns, and predict the impact of technological aids on human performance, thereby enhancing aviation safety.
Key finding
Human performance models successfully predicted pilot errors and scanning behavior changes during taxi and approach tasks, demonstrating their utility in identifying latent design flaws and evaluating the impact of advanced cockpit displays.
Methodology
simulation_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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-28 |
| archive | success | unpaywall | — | — | 1 | 2026-06-04 |
| 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 | crossref | — | — | 2 | 2026-06-04 |
| 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