Assessing Impacts of Age on Cognitive and Physiological Performance in Construction Activities with Falling Risks Using Wearable Sensors

Abstract

The construction industry is among the most hazardous globally, with falling risks being a leading cause of safety violations. These risks are further influenced by age-related differences in cognitive and physiological capacities. Yet current safety practices are largely age-agnostic and reactive. Wearable sensors present a transformative solution by enabling real-time, objective monitoring of psychophysiological responses. This study aimed to examine age-related influences on falling risk formation and assessment in construction activities, focusing on cognitive and physiological performance measured via wearable sensors.

Since hazard detection is the first defense against falling risks, this research first investigated the impact of age on visual cognitive processing in detecting falling-related hazards: falling from heights and same-level falling hazards, using Electroencephalogram (EEG) data. Results showed older workers exhibited stronger pre-attentive responses but weaker attentive processing, particularly underestimating same-level falling hazards while remaining highly sensitive to high-risk falling from heights. These differences, independent of work experience, suggested that cognitive aging primarily influenced falling-related hazard detection patterns.

Given that the underlying formation patterns may differ between falling from heights and same-level falling risks, this research further examined these two distinct falling risks. For falling from heights in high-altitude construction (herein referring to working at heights), age differences in vigilance, mental fatigue, attention, task engagement, and height-related anxiety were assessed through Virtual Reality (VR) simulations and EEG analysis. Older workers experienced greater mental fatigue and lower vigilance but demonstrated higher task engagement and reduced height-related anxiety. Conversely, younger workers exhibited heightened emotional height-related anxiety, with mental fatigue being the most significant risk factor for older workers.

The study extended its investigation of same-level falling risks by examining how age affected auditory attention and muscle coordination in on-foot load-carrying tasks using Electrodermal Activity (EDA) and surface Electromyography (sEMG). Results showed aging impaired cognitive-motor interactions, especially under high-load, dual-task conditions. Older workers experienced declines in auditory attention and muscle coordination, worsened by load demands. Their prioritization strategies shifted with load intensity: they balanced performance under low loads but prioritized cognition with sacrificing motor control under high loads.

Based on the exploration of age differences in falling risk formation, this research proposed an age-specific risk assessment paradigm incorporating EEG-based physiological measurements. Using VR to simulate high-risk construction scenarios and Shapley Additive Explanations (SHAP) for model interpretability, the study identified 29 significant EEG features for older workers and 30 for younger workers, highlighting the dominant role of cognitive processing and mental workload over emotional states. The support vector machine (SVM) classifier demonstrated superior performance, achieving 87.13% accuracy for older workers and 81.56% for younger workers. SHAP analysis revealed younger workers exhibited stronger interactions between beta activity and emotional arousal, whereas older workers’ falling risk assessment relied more on cognitive workload. This research advances age-related safety management in construction by deepening the understanding of age impact on cognitive and physiological performance related to falling risks. It pioneers the use of interpretable machine learning for physiological risk assessment and enables age-specific risk evaluation, fostering a more adaptive and resilient industry that embraces workforce age diversity.

Date of Award	4 Sept 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Xiaowei LUO (Supervisor)