Kanghyeok Yang

Identifying Safety Hazards Using Collective Bodily Responses of Workers

AbstractCurrent construction hazard identification mostly relies on safety managers’ ability to identify hazards using their prior knowledge about them. Consequently, numerous latent hazards remain unidentified, which poses significant risks to construction workers. To advance current hazard identification capabilities, this study examines the feasibility of harnessing and analyzing collective patterns of workers’ bodily responses (balance, gait, etc.) to identify safety hazards on a jobsite. To test the hypothesis that the abnormality of workers’ bodily responses in one location highly correlates with the likelihood of a safety hazard in that location, this project collected data on the bodily responses of 10 subjects who participated in five experiments. These test subjects wore inertial measurement unit (IMU) sensors on their body. Then the collected response data were analyzed using three metrics [average, standard deviation, and Shapiro-Wilk statistic (W)]. The data showed that the normality of worke...

Automated Detection of Near-miss Fall Incidents in Iron Workers Using Inertial Measurement Units

1 Ph.D Student, Construction Engineering and Management, Charles Durham School of Architectural Engineering and Construction, University of NebraskaLincoln, W113 Nebraska Hall, Lincoln, NE 68588; PH (402) 472-5631; email: kyang11@unl.edu 2 M.S. Student, Computer Science and Engineering Department, College of Engineering, University of Nebraska-Lincoln, W113 Nebraska Hall, Lincoln, NE 68588; PH (402) 472-5631; email: saria@cse.unl.edu 3 Assistant Professor, Construction Engineering and Management, Charles Durham School of Architectural Engineering and Construction, University of Nebraska-Lincoln, W113 Nebraska Hall, Lincoln, NE 68588; PH (402) 4727431; email: cahn2@unl.edu 4 Associate Professor, Environmental, Agricultural, Occupational Health Science, 984388 Nebraska Medical Center, Omaha, NE 68198-4388 and Construction Engineering and Management, Charles Durham School of Architectural Engineering and Construction, W113 Nebraska Hall, College of Engineering, University of Nebraska, Lincoln, NE 68588-0500; PH (402) 472-5078; email: tstentz1@unl.edu

A people-centric sensing approach to detecting sidewalk defects

A defective sidewalk inhibits the walkability of a street and may also cause safety accidents (slips, trips, and falls) for pedestrians. When a pedestrian walks along a sidewalk, his/her behaviors may vary according to the condition of the sidewalke.g., whether the surface is normal, holed, cracked, tilted, or sloped. As a result, the pedestrians stability may also change according to the built environments conditions. Accordingly, this paper examines the feasibility of using pedestrians physical behaviors to detect defects in a sidewalk. Pedestrians physical responses and paths over a sidewalk are collected using an inertial measurement unit (IMU) sensor and a global positioning system (GPS). Then, after aggregating the pedestrians bodily responses and locations, the irregularity of multiple pedestrians responses are calculated in relation to their locations. The locations that show irregularities in the pedestrian-response patterns present a high correlation with the existence of a defect. The results of this study will help improve the continuous diagnosis of defects in sidewalks, thereby enhancing these built environment systems serviceability.

Deciphering Workers’ Safety Attitudes by Sensing Gait Patterns

Workers’ unsafe behaviors are a top cause of safety accidents in construction. In practice, the industry relies on training and education at the group level to correct or prevent unsafe behaviors of workers. However, evidence shows that some individuals were identified to be showing risky behavior repeatedly and have a high rate to be involved in accidents and current safety training approach at the group level may not be effective for those workers. A worker’s evaluation of a hazard (risk perception) and tendency to take/avoid risks (risk propensity) determines how they respond to a hazard and identifying those workers with biased risk perceptions and high risk propensity can thus provide an opportunity to prevent behavior-based injuries and fatalities in the workplace. However, as risk perception and propensity are influenced not only by inherited personal traits (e.g. locus of control) but also by specific situational factors (e.g. mood and stress level), existing approaches relying on surveys are not sufficient when measuring workers’ risk perception and propensity continuously in day-to-day operations. In this context, this study examines the potential of ambulatory and continuous gait monitoring in the workplace as a means of identifying workers’ risk perception and propensity. Two experiments simulating construction work environments were conducted and subjects’ gait patterns in hazard zones were assessed with inertial measurement unit (IMU) data. The experimental results demonstrate changes in gait patterns at pre-hazard zones for most of the subjects. However, the results fail to identify the relationship between gait pattern changes at pre-hazard zones and risk propensities assessed using the Accident Locus of Control Scale.

Near-Miss Accident Detection for Ironworkers Using Inertial Measurement Unit Sensors

In the construction industry, fall accidents are the leading cause of construction-related fatalities; in particular, ironworkers have the highest risk of fatal accidents. Detecting near-miss accidents for ironworkers provides crucial information for interrupting and preventing the precursors of fall accidents while simultaneously addressing the problem of sparse accident data for ironworkers’ fallrisk assessments. However, current methods for detecting near-miss accidents are based upon workers’ self-reporting, which introduces variability to the collected data. This paper aims to present a method that uses Inertial Measurement Unit (IMU) sensor data to automatically detect near-miss accidents during ironworkers’ walking motion. Then, using a Primal Laplacian Support Vector Machine, a developed semi-supervised algorithm trains a system to predict near-miss incidents using this data. The accuracy of this semi-supervised algorithm was measured with different metrics to assess the impact of the automated near-miss incident detection in construction worksites. The experimental validation of the algorithm indicates that near-miss incidents may be estimated and classified with considerable accuracy—above 98 percent. Then the computational burden of the proposed algorithm was compared with a One-Class Support Vector Machine (OC-SVM). Based upon the proposed detection approach, highrisk actions in the construction site can be detected efficiently, and steps towards reducing or eliminating them may be taken.