A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Classification of construction hazard-related perceptions using: Wearable electroencephalogram and virtual reality
Abstract Improving construction workers' safety is one of the most critical issues in the construction industry. Methods have been developed to better identify construction hazards on a jobsite by analyzing workers' physical and physiological responses collected from the wearable devices. Among them, electroencephalogram (EEG) holds unique potential since it shows immediate abnormal responses when a hazard is perceived. However, there remain limitations in the current knowledge base to attain the ultimate goal of ubiquitous hazard identification. In this context, this study investigates the feasibility of identifying construction hazards by developing an EEG classifier based on the experiments conducted in an immersive virtual reality (VR) environment. Results show that the CatBoost classifier achieved the highest performance with 95.1% accuracy. In addition, three important channel locations (AF3, F3, and F4) and two frequency bands (beta and gamma) were found to be closely associated with hazard perception.
Highlights Fifteen ML classifiers were developed to classify hazard-related EEG signals. CatBoost classifier achieved the highest performance (95.1% accuracy). Nine critical EEG features contributing to classification performance were found. Classifier retrained using significant features achieved acceptable performance. Three EEG channels and two frequency bands are closely related to hazard perception.
Classification of construction hazard-related perceptions using: Wearable electroencephalogram and virtual reality
Abstract Improving construction workers' safety is one of the most critical issues in the construction industry. Methods have been developed to better identify construction hazards on a jobsite by analyzing workers' physical and physiological responses collected from the wearable devices. Among them, electroencephalogram (EEG) holds unique potential since it shows immediate abnormal responses when a hazard is perceived. However, there remain limitations in the current knowledge base to attain the ultimate goal of ubiquitous hazard identification. In this context, this study investigates the feasibility of identifying construction hazards by developing an EEG classifier based on the experiments conducted in an immersive virtual reality (VR) environment. Results show that the CatBoost classifier achieved the highest performance with 95.1% accuracy. In addition, three important channel locations (AF3, F3, and F4) and two frequency bands (beta and gamma) were found to be closely associated with hazard perception.
Highlights Fifteen ML classifiers were developed to classify hazard-related EEG signals. CatBoost classifier achieved the highest performance (95.1% accuracy). Nine critical EEG features contributing to classification performance were found. Classifier retrained using significant features achieved acceptable performance. Three EEG channels and two frequency bands are closely related to hazard perception.
Classification of construction hazard-related perceptions using: Wearable electroencephalogram and virtual reality
Jeon, JungHo (author) / Cai, Hubo (author)
2021-09-21
Article (Journal)
Electronic Resource
English
Image-based construction hazard avoidance system using augmented reality in wearable device
| British Library Online Contents | 2017
Virtual reality improves hazard awarene
| IuD Bahn | 2003
| Emerald Group Publishing | 2021