Systematic Literature Review on The Application of Data Mining Methods to Monitor the Operator’s Functional State in Human-Machine Systems

Abstract

Background and aims: Continuous monitoring of operator functional state, is one of the most important topics, and the data mining methods are considered as a suitable tool for providing the performance evaluation models. However, there has been no comprehensive study on the use of data mining methods in this field so far, and in so doing, the aim of the present article, was to systematically review the role and importance of the data mining methods to monitor the operator functional state in human-machine systems. Methods: A total of 86 published articles that evaluated the operator functional state, were reviewed in five databases. All articles were analyzed in four groups related to the operator functional state, three critical safety systems, and three types of data mining techniques. The operator functional state, was also assessed through methods of physical measurement, psychophysiological measurement, task-performance indicators, and subjective judgment. Results: Most of the data mining models were related to the field of road and air transportation, which are mainly focused on fatigue and the task-performance indicators. The support vector machine and neural network models were the most frequently used data mining methods. The results showed, that most studies were performed on fatigue models, among the functional states where mainly physical measurements were used; however, the psychophysiological measurements were the most frequently applied measuring method for the mental workload models and the task performance indicators. Conclusion: A comprehensive evaluation of the data mining methods and the parameters used in these models to assess the operator functional state, will identify the research gaps in this area, and results in conducting more extensive studies to improve the human performance.