Development of Statistical Geometric Models for a Driver’s Hip and Eye Locations

Abstract

Drivers’ hip locations (HLs) and eye locations (ELs) have been used as reference data to design an ergonomic automobile interior. Although many prediction models of a driver’s HL and EL have been developed, the developed models have limitations in terms of prediction accuracy and stability. The objectives of the present study are (1) development of statistical geometric models (SGMs) of a driver’s HL and EL, and (2) evaluation of the accuracy of the SGMs. Forty drivers’ preferred driving postures were measured by a motion capture system in 3 different vehicle conditions (coupe, sedan, and SUV). The SGMs were developed by incorporating the geometric relationships between HL, EL, anthropometric dimensions, and driving postures and the statistical relationships between body link lengths and surface landmark lengths. The SGMs were evaluated quantitatively by comparing the Reed et al. (2002)’s models in terms of prediction accuracy. As a result, the average adj. R² of SGMs is 1.1 ~ 3.7 times higher than Reed et al.’s models and root mean squared error (RMSE) of the SGMs is 1.7 ~ 1.8 times smaller than the Reed et al.’s models. Moreover, RMSE of the SGMs in three vehicle conditions are 1.7 ~ 4.3 times smaller than the Reed et al.’s models. These results indicate that the accuracy of the SGMs are more accurate and stable than the Reed et al.’s models in all three vehicle conditions. The developed SGMs have high applicability to the ergonomic design of automobile interiors such as the seat adjustment range and windshield height.