Deep Learning-Based Prediction of Comfortable Driving Postures for Elderly Chinese Drivers

Abstract

With the intensification of aging in China, the number of elderly drivers is continuously increasing. Uncomfortable driving postures can cause serious physical damage to elderly drivers, yet there is still a lack of ergonomic design for comfortable driving postures specifically tailored to the elderly driver population in China. This paper first identifies the main factors and their weights affecting the comfort of driving postures through analysis and experimental methods. Subsequently, the anthropometric data of the elderly from GB10000-2023 are adjusted with a forward-looking predictive correction to align with current and future design requirements for driving postures. Finally, a predictive model for comfortable driving postures is constructed using deep learning, enabling personalized predictions of comfortable driving postures for the elderly. Practical applications demonstrate that this method can effectively improve the comfort of driving postures for elderly drivers, reducing the risk of physical injury and offering significant practical value.