다변량 통계기법을 활용한 표준 운전자세의 도출

Abstract

The package layout of a vehicle is very important to the driver's postural comfort. It considers human body characteristics of drivers and passengers, and uses the anthropometric data on driver's posture. Especially, since the human bodies were scanned in Size Korea project, three-dimensional human shape data have been considered the useful data in vehicle design stage. The proper driver's postural data are necessary in order to apply three-dimensional human data or ergonomic human models to package layout. Traditional postural data were provided to each joint angle individually, such as elbow angle, shoulder angle, hip angle, and knee angle. However, it is not appropriate to account for the human posture because the limbs move in coupled with each other. In the study, some whole-body posture were derived by multivariate statistical analysis to develop driver's standard posture, which may be used for reference data in application of human model software or human shape data in CAD environment. From the clustering analysis five whole-body postures were classified for the driver's representative posture. This study demonstrates that the combination of the most frequent angles of each joint can not explain the driver's actual posture in car, and that the joint angles must be considered together as a whole-body posture. It is expected that this study is helpful for the package layout designer to understand the driver's posture and to apply the human shape data to vehicle design.