An explorative neural networks-enabled approach to predict stress perception of traffic control operators in dynamic working scenarios

Abstract

Traffic control operators are usually confronted with a high level of stress. Minimizing stress in traffic control operations has always been the main objective of redesigning management systems, updating procedures, and developing training programs. However, there is a lack of methods that are able to feed stress analysis at the design stage and bring results that can be used in decision-making. In an effort to facilitate the proactive management of stress, we introduced a stress prediction model which can predict the operator’s stress perception of a working scenario based on task, human, and environmental information. First, a framework of scenario-based stress investigation was proposed for data collection. Second, a neural networks-enabled novel model for stress prediction was developed based on the Job Demands-Resources(JD-R) theory. Third, a personalization module and a Mixture-of-Expert (MoE) module were introduced into the prediction model to overcome the problem of individual differences. A case study of vessel traffic management showed that the model can predict an individual’s stress perception of a working scenario with a mean absolute error of 0.33. A comparative study was conducted to validate the effectiveness of proposed model.