RESUMEN
INTRODUCTION: It is critical to develop and implement lab-based computer experiments that simulate real-world tasks in order to characterize operational requirements and challenges or identify potential solutions. Achieving a high degree of laboratory control, operational generalizability, and ease-of-use for a task is challenging, often leading to the development of tasks that can satisfy some facets but not all. This can result in insufficient solutions that leave real-world stakeholders with unsolved problems. MATERIALS AND METHODS: This issue is addressed using a customized passive sonar simulator application that provides extensive researcher control over the design and manipulation of a sonar task; a visual appearance and cognitive demand similar to a true submarine-based sonar task; and a convenient and short training routine for sonar novices. The task requires participants to watch for multiple signal sources of varying appearance and salience and subsequently classify these signals into their respective categories. RESULTS: The current study investigated the effects of stimulus signal strength and signal density on sonar task performance-including metrics of classification accuracy, classification confidence, and response times-finding an interaction between signal density and signal strength that resulted in greater performance errors with high signal density at the weakest signal strength. CONCLUSIONS: The lab-based sonar application provides new possibilities for research, not limited to signal intensity and signal density but also through the manipulation of parameters such as the number of unique targets, target appearance, and task duration. This application may illuminate the operational demands that each of these factors may have on operator behavior within the dynamic tasks.
Asunto(s)
Análisis y Desempeño de Tareas , Humanos , Masculino , Adulto , Femenino , Carga de Trabajo/psicología , Carga de Trabajo/normas , Simulación por Computador/normasRESUMEN
High risk professions, such as pilots, police officers, and TSA agents, require sustained vigilance over long periods of time and/or under conditions of little sleep. This can lead to performance impairment in occupational tasks. Predicting impaired states before performance decrement manifests is critical to prevent costly and damaging mistakes. We hypothesize that machine learning models developed to analyze indices of eye and face tracking technologies can accurately predict impaired states. To test this we trained 12 types of machine learning algorithms using five methods of feature selection with indices of eye and face tracking to predict the performance of individual subjects during a psychomotor vigilance task completed at 2-h intervals during a 25-h sleep deprivation protocol. Our results show that (1) indices of eye and face tracking are sensitive to physiological and behavioral changes concomitant with impairment; (2) methods of feature selection heavily influence classification performance of machine learning algorithms; and (3) machine learning models using indices of eye and face tracking can correctly predict whether an individual's performance is "normal" or "impaired" with an accuracy up to 81.6%. These methods can be used to develop machine learning based systems intended to prevent operational mishaps due to sleep deprivation by predicting operator impairment, using indices of eye and face tracking.