MODALSIM94: A Multimodal Physiological and Performance Dataset from Simulated Flight Workload Tasks

Abstract

MODALSIM94 is a multimodal dataset collected from 94 participants performing simulated flight workload tasks in a fixed-base research flight simulator. The dataset was designed to support research on cognitive workload assessment, pilot monitoring, adaptive automation, and human–machine interaction in aviation-relevant environments. Participants completed three five-minute flight scenarios corresponding to low, medium, and high workload conditions. Workload was manipulated through a secondary N-back working-memory task, with no N-back task in the low workload condition, a 1-back task in the medium workload condition, and a 2-back task in the high workload condition. The dataset includes synchronised physiological, behavioural, performance, subjective, and demographic data. Physiological measures comprise cardiac activity from an eSENSE Pulse sensor, respiration data from an eSENSE Respiratory sensor, and eye-tracking data from Dikablis 3 goggles. Flight and performance data include aircraft dynamics, control-related variables, trajectory information, tunnel-deviation metrics, and N-back task responses, accuracy, and reaction times. Subjective measures include NASA-TLX workload ratings, SART situation-awareness ratings, Karolinska Sleepiness Scale scores, Samn–Perelli fatigue ratings, and participant demographic information. All data are provided in open CSV/TXT formats and organised by modality. Each modality preserves its native temporal resolution, while participant identifiers, scenario labels, and elapsed-time references from scenario onset support temporal alignment across data streams. MODALSIM94 can be used for workload classification and regression, multimodal physiological modelling, flight-performance analysis, personalisation studies, and cross-domain research on cognitive-state monitoring.