バーチャルヒューマンモデルによる作業性、安全性の研究

第七報 ニューラルネットワークを使った荷物持上げ時の身体負荷解析

抄録

Lifting work is frequently observed in the production site of shipbuilding companies. It seems to be the work in which the body burden is bigger, and it often causes low back pain and injury of knee for aged shipbuilding workers. The low back pain concerns the compressive force on intervertebral disks, while the injury of knee concerns torque on the joint of knee. The analysis of such force and torque during the lifting job can be carried out using the biomechanical analysis, however, the calculation is usually complex and needs special know-how. Hence, the neural network method was applied in this paper to obtain the low back compression force (force on intervertebral disks) and torque on knee by developing C language program. The original teaching data were provided by the simulation using general-purpose software “Jack” applying virtual human model, and the neural network model learned these teaching data. Then, it was compared with the analytical results of “Jack” for the low back compression force and torque on knee, and both data obtained from the neural network and “Jack” were well coincided. Hence, the low back compression force and torque on knee at lifting work can be approximately gained by this neural network model without applying human software “Jack”.