Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation
In safety-critical work environments (e.g., military, space operations), it is imperative that psycho-physical measurements do not disrupt operator's task performance. Consequently, many industries are now interested in the feasibility of implementing wearable technologies to passively assess p...
| Main Author: | |
|---|---|
| Format: | Conference Paper |
| Published: |
2020
|
| Online Access: | http://hdl.handle.net/20.500.11937/80225 |
| _version_ | 1848764181920088064 |
|---|---|
| author | Wilson, Micah |
| author_facet | Wilson, Micah |
| author_sort | Wilson, Micah |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | In safety-critical work environments (e.g., military, space operations), it is imperative that psycho-physical measurements do not disrupt operator's task performance. Consequently, many industries are now interested in the feasibility of implementing wearable technologies to passively assess psychological and physical states relevant to human performance (e.g., stress, fatigue, workload) on a continuous basis. However, studies demonstrating reliable associations between sensors and domain-relevant psycho-physical states are typically conducted in tightly controlled settings over relatively short timescales. This talk will outline the results of a field study that involved evaluating the utility of wearable sensors on board an operational Royal Australian Navy (RAN) vessel. RAN crew (n = 63) were equipped with electrocardiogram (ECG) and actigraphy (ACT) sensors, and completed psychometric testing four times per-day and a continuous activity diary. Data collection occurred over a 14 day mission, with acceptable compliance rates (> 80%). The data were used to develop a series of open-source bio-mathematical models capable of predicting subjective fatigue under different sleep schedules using full Bayesian inference. Additionally, time-domain ECG features were linked with daily diary observations, and an Artificial Neural-Network machine learning classifier was trained to detect sleep episodes. The classifier achieved a mean accuracy = 86.9%. Several challenges are discussed. |
| first_indexed | 2025-11-14T11:15:17Z |
| format | Conference Paper |
| id | curtin-20.500.11937-80225 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T11:15:17Z |
| publishDate | 2020 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-802252023-03-13T03:03:47Z Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation Wilson, Micah In safety-critical work environments (e.g., military, space operations), it is imperative that psycho-physical measurements do not disrupt operator's task performance. Consequently, many industries are now interested in the feasibility of implementing wearable technologies to passively assess psychological and physical states relevant to human performance (e.g., stress, fatigue, workload) on a continuous basis. However, studies demonstrating reliable associations between sensors and domain-relevant psycho-physical states are typically conducted in tightly controlled settings over relatively short timescales. This talk will outline the results of a field study that involved evaluating the utility of wearable sensors on board an operational Royal Australian Navy (RAN) vessel. RAN crew (n = 63) were equipped with electrocardiogram (ECG) and actigraphy (ACT) sensors, and completed psychometric testing four times per-day and a continuous activity diary. Data collection occurred over a 14 day mission, with acceptable compliance rates (> 80%). The data were used to develop a series of open-source bio-mathematical models capable of predicting subjective fatigue under different sleep schedules using full Bayesian inference. Additionally, time-domain ECG features were linked with daily diary observations, and an Artificial Neural-Network machine learning classifier was trained to detect sleep episodes. The classifier achieved a mean accuracy = 86.9%. Several challenges are discussed. 2020 Conference Paper http://hdl.handle.net/20.500.11937/80225 restricted |
| spellingShingle | Wilson, Micah Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title | Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title_full | Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title_fullStr | Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title_full_unstemmed | Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title_short | Wearable Sensors in an Extreme Work Environment: Applying Computational Modelling for Evaluation |
| title_sort | wearable sensors in an extreme work environment: applying computational modelling for evaluation |
| url | http://hdl.handle.net/20.500.11937/80225 |