Detecting Soldiers’ Fatigue Using Eye-Tracking Glasses: Practical Field Applications and Research Opportunities

Abstract

Introduction Objectively determining soldiers’ fatigue levels could help prevent injuries or accidents resulting from inattention or decreased alertness. Eye-tracking technologies, such as optical eye tracking (OET) and electrooculography (EOG), are often used to monitor fatigue. Eyeblinks—especially blink frequency and blink duration—are known as easily observable and valid biomarkers of fatigue. Currently, various eye trackers (i.e., eye-tracking glasses) are available on the market using either OET or EOG technologies. These wearable eye trackers offer several advantages, including unobtrusive functionality, practicality, and low costs. However, several challenges and limitations must be considered when implementing these technologies in the field to monitor fatigue levels. This review investigates the feasibility of eye tracking in the field focusing on the practical applications in military operational environments. Materials and Method This paper summarizes the existing literature about eyeblink dynamics and available wearable eye-tracking technologies, exposing challenges and limitations, as well as discussing practical recommendations on how to improve the feasibility of eye tracking in the field. Results So far, no eye-tracking glasses can be recommended for use in a demanding work environment. First, eyeblink dynamics are influenced by multiple factors; therefore, environments, situations, and individual behavior must be taken into account. Second, the glasses’ placement, sunlight, facial or body movements, vibrations, and sweat can drastically decrease measurement accuracy. The placement of the eye cameras for the OET and the placement of the electrodes for the EOG must be chosen consciously, the sampling rate must be minimal 200 Hz, and software and hardware must be robust to resist any factors influencing eye tracking. Conclusion Monitoring physiological and psychological readiness of soldiers, as well as other civil professionals that face higher risks when their attention is impaired or reduced, is necessary. However, improvements to eye-tracking devices’ hardware, calibration method, sampling rate, and algorithm are needed in order to accurately monitor fatigue levels in the field.

Examining the Landscape of Cognitive Fatigue Detection: A Comprehensive Survey

“…Cognitive fatigue arises from a complex interplay of factors, encompassing physical, social, environmental, and emotional influences [199]. While research has explored individual features like heart rate [200,201], eye movement [202], and EEG signals [203,204], a holistic approach integrating diverse features remains elusive. Datasets like Cogbeacon [205] showcase the multi-faceted nature of fatigue, but pinpointing individual triggers proves challenging due to inter-personal differences.…”

Eye tracking in extreme environments: from invention to new frontiers of human-machine teaming

“…Dawn Kernagis and a comprehensive literature review helped pinpoint critical oculometric and pupillometry measures utilized in medical diagnostics to detect neuro-cognitive and physiological states. This revealed a number of oculometric and pupillometry measures which have been used in medical diagnostics and scientific literature to detect neuro-cognitive and physiological states [4][5][6]7,12,14,15,16,[17][18][19]22,25,26]. These measures are outlined in the table below (Table 1).…”

A CNN-Based Wearable System for Driver Drowsiness Detection

“…These issues can be easily overcome by head-mounted devices, which offer a plug-and-play solution for existing cars that do not have cameras built in. In this line of research work, head-mounted eye-tracking glasses have long been used to examine driver drowsiness both in the lab and in natural driving scenarios [ 29 ]. Using SMI (SensoMotoric Instruments) eye-tracking glasses, Gao et al [ 30 ] used a support vector machine (SVM) to analyze PERCLOS [ 31 ] to evaluate driving fatigue with time series data.…”