REGRESSION ANALYSIS OF OXYGEN SATURATION LEVEL FOR CRITICAL DRIVING FATIGUE FACTORS USING BOX-BEHNKEN DESIGN

Abstract

Driving fatigue research is becoming more popular as the number of fatigue-related road incidents rises. However, there are limited studies examining the role of physiological factors and the significant variables influencing them to determine driving fatigue. Hence, the purpose of this study was to perform a regression analysis to evaluate whether variables such as driving duration, driving speed, body mass index (BMI), gender and types of roads are relevant in influencing oxygen saturation levels and how these variables interact to suggest driving fatigue. A regression analysis was carried out utilizing the Box-Behnken design. The results showed that the values of Prob > F for all variables were less than 0.01%, showing that all variables significantly impacted the oxygen saturation levels. The oxygen saturation levels declined as driving duration, speed, and BMI rose. The same pattern emerged between female and male drivers. Furthermore, the oxygen saturation levels were lower on straight roads than on rather difficult uphill/downhill roads. The regression model was evaluated to determine its accuracy by comparing output data from software prediction and actual driving experiments. First, the output data prediction interval obtained by both approaches was within a 95% confidence interval, satisfying the minimum quantitative limitation of a 90% predictive interval. Residual errors were also less than 10%.  The findings may be helpful to researchers and decision-makers involved in the field of road safety in an effort to lower the number of traffic accidents brought on by driver fatigue.