Recently construction workers tend to work alone and are getting older. In such working atmosphere, monitoring their works at the site is necessary for keeping task performance quality and accuracy. However, the management and evaluation of workers performance can only be done by supervisors physical observation. Understanding of their works will be easier if the data that shows what kind of task is performed can be estimated from their body positions and movement, and that enable monitoring by data obtained concurrently.
Data for body positions and movements of workers was obtained and measured by using 6-DOF motion sensors, and the method to estimate the type of task by using neural network is described. Finally the authors developed the method to monitor the work at the site using the data obtained by 6-DOF motion sensors and systemized the estimation procedure.