Human Contributions to Safety Data Testbed Flight Simulation Study: Data Methods, Processing, and Quality

Fettrow, Tyler; Stephens, Chad; Prinzel, Lance; Holbrook, Jon; Ballard, Kathryn; Bastami, Sepehr; Stewart, Michael; Kiggins, Daniel · 2025 · Crossref

DOI: 10.1038/s41597-025-05336-7

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 describes the methodology, data processing, and quality assurance for the System-Wide Safety Operations and Technologies for Enabling Resilient In-Time Assurance (SOTERIA) study. The research addresses a critical gap in aviation safety science: the lack of data on routine, successful pilot performance. While rare human errors are extensively studied, the majority of actions flight crews take to maintain safe operations remain unmeasured. The primary research question investigates how commercial airline pilots manage routine contingencies during Area Navigation (RNAV) arrivals, aiming to quantify resilient behavior and proactive safety contributions rather than just failures. The study involved 24 commercial airline pilots (15 men, 9 women; mean age 49.1) operating in a high-fidelity, motion-based Boeing 737-800 simulator at NASA Langley Research Center. Participants performed scenarios replicating real-world challenges at Charlotte Douglas International Airport (KCLT), including traffic compression, convective weather, energy management issues, unanticipated tailwinds, autoflight failures, wake turbulence encounters, and communication errors. Data collection was multimodal and minimally intrusive. Psychophysiological measures included electroencephalography (EEG) and electrocardiography (ECG) sampled at 256 Hz via a B-Alert ABM X10 device, and galvanic skin response, skin temperature, and heart rate sampled via an Empatica E4 smartwatch. Eye tracking was recorded at 60 Hz using Smarteye Pro DX. All scenarios were video and audio recorded, with systems time-synchronized via Network Time Protocol. Subjective data were collected using NASA-TLX, Situation Awareness Rating Technique (SART), and custom resilience assessments. Post-simulation, pilots completed retrospective think-aloud exercises and event report drafts. The paper details the rigorous data management and quality control processes. Raw data from twelve crews (one incomplete) were stored in a structured repository, with files organized by crew, scenario, and data type. Preprocessing involved converting raw logs to CSV formats and mapping timestamps to specific scenarios using manual notes. Quality assurance included automated Python scripts to verify file integrity during transfers from local drives to NASA’s Box repository and subsequently to an AWS S3 bucket for public access. The authors report on data loss due to network drops and storage limitations, noting that specific files were overwritten or re-synced to ensure consistency. The dataset includes synchronized psychophysiological signals, flight dynamics, pilot inputs, video recordings, and subjective survey responses. The significance of this work lies in the creation of a comprehensive, publicly available dataset that enables the quantification of resilient performance in commercial aviation. By providing access to both physiological and behavioral data from routine operations, the SOTERIA dataset supports a paradigm shift from error-centric safety models to those that value proactive, positive safety contributions. This resource allows researchers to analyze how pilots anticipate, monitor, and respond to disturbances, fostering new insights into human factors and resilience engineering. The open availability of the data and code encourages broader scientific contribution to understanding and enhancing safety in complex operational environments.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

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

• naturalistic crash near crash

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: self report data
• Methodological Resource: tool software, dataset resource