建築工事における作業内容の推定方法に関する研究 : 6自由度データに基づく構成動作の推定方法

抄録

This paper discussed on the estimation of building works with 6-DOF motion sensors. The 6-DOF (Degree-of-freedom) sensor captures 3-axis acceleration and 3-axis angle. The actual data on the sensors is computed to the feature parameters in each time-unit for the recognition of the work motions. The neural network is applied for the method of the recognition. The neural network model is designed in the three layers Back-propagation network with the standard connections. The neural network outputs show guidelines to select the work motions. The authors collected the 6-DOF motion data in partition works, and estimated work motions based on those data. The differences between the actual and the estimated motions were evaluated, and it was clarified that the method is applicable to estimate work motions in building works.