Workload categorization for hazardous industries: the semantic modelling of multi-modal physiological data

The forestry industry is one of the most hazardous industries in New Zealand, and the physical and cognitive fatigue of forestry workers has been shown to contribute to this. Physical and cognitive fatigue can be exacerbated by prolonged physical and cognitive workload. As such, we propose that the identification and mitigation of fatigue factors could reduce the risk of incidents and injuries in hazardous work environments. This paper introduces a semantic model for workload categorization. The model takes as input a set of multi-modal physiological measurements and uses parallel processing, complex event processing, and rule-based modeling to categorize a series of workloads (resting, cognitive workload, and physical workload). The model has undergone a set of evaluations, including categorization accuracy and performance. The model has been tested under three scenarios: when a participant is resting and refraining from any physically or mentally demanding tasks; when a participant is undertaking a cognitively intensive task; and when a participant is walking, jogging, and running. The study has been conducted with participants between the ages of 22 and 39 and has shown an average accuracy of 89% for resting workload, 76% for cognitive workload, and 97% for physical workload. Finally, in this paper, we discuss the application and extension of this model to predict fatigue in hazardous industries. The work described in this paper contributes to a larger research project centered on investigating technology uses in hazardous work environments.
Journal Article
Type of thesis
This is an author’s accepted version of an article published in Future Generation Computer Systems. © 2023 Elsevier B.V.