EXPLORE PUBLICATIONS BY COUNTRIES

Seat Comfort Index, Artificial Neural Networks, Seat Comfort Modeling.

The purpose of this work is to investigate driver seat comfort and discuss some of the subjective and objective factors that impact it. Comfort describes the nature of the interaction between a human being and a specific environment and is characterized as a feeling of pleasure and satisfaction or discontent and pain. Driver seat design for comfort is complex and challenging because it is somewhat subjective in nature (e.g. mood, culture & car brand). However, there are certain aspects of a seat comfort, which are objective (e.g. anthropometrics, pressure distribution on seat) and can be modeled mathematically. This paper discuses some of the objective and subjective measures which influence seat comfort. In addition, it provides a mathematical model of seat comfort index based on neural networks in terms of some of the objective measures, which influence it. The results of this work show that the objective measures included lead to a correlation of 0.794 to the overall comfort index identified by twelve drivers testing five different types of seats. This implies that our selected objective measures/inputs can capture about 80% of the comfort index identified by test drivers. The remaining 20% variation in comfort index not captured by the model utilized in this work can be attributed to subjective measures and/or additional objective measures which can be added as inputs to the neural network.

A comparison of this study to a previously published work [1] which also utilizes neural networks to model seat comfort index reveals important facts. This previous study uses a very large neural network with a significant number of adjustable parameters to model seat comfort. As a result, their neural network is very prone to memorizing the data associated with seat comfort index without capturing the underlying mathematical behavior. The neural network proposed in this work, however, has an optimal architecture, which captures the mathematical model describing comfort index accurately, and is not prone to memorizing the seat comfort data.