WORK ANALYSIS OF PLASTERBOARD-PASTING WORKERS FOCUSED ON THE SMOOTHNESS OF ACTIVITIES USING ACCELEROMETERS

In this study, we conducted a work analysis at an active building construction site, utilizing three-axis acceleration sensors affixed to four plaster-board-pasting workers for five days. Although acceleration data is less accurate than visual or image recognition in identifying specific tasks, it can be easily captured using smartphones, even in challenging conditions such as poorly lit or obstructed construction sites. This accessibility facilitates continuous data collection over extended periods, enabling automated analysis without significant cost or time investment. In addition, this method can identify trends in worker behavior that elude conventional visual inspection. Our approach encompasses various evaluation indices, beginning with an analysis of average work time per plasterboard sheet and the differentiation of walking motions using acceleration data. Furthermore, we introduced a new evaluation index that quantifies the distribution of high- and low-intensity work based on acceleration readings. Through comparative analysis with evaluation indices from previous studies, we confirmed common trends and discussed the strengths and limitations of our proposed index. Our findings suggest a correlation between work experience and performance, as evidenced by smoother operational patterns among seasoned workers. In particular, proficient workers exhibited fewer instances of extremely intense or sporadic movements. This observation underscores the influence of experience on work dynamics.