Analysis of Driving Parameters using Pearson Correlation and K-means Cluster: A case study of Sarawak, Malaysia

Abstract

- Driving Cycle is widely adapted by automotive industry in evaluating vehicle fuel consumption and emission towards manufacturing of a sustainable and efficient vehicle. It has become a debatable issue in automotive industry due to tailorable features in line with driving and road condition of a specific place. This paper intends to clarify on the selection of representative driving parameters from the pool of microtrips based on the data gathered throughout Sarawak continent. Joint method of car chasing and circulation techniques were employed during data collection. Routes were selected by assessing the traffic volume extracted from Road Traffic Volume Malaysia (RTVM) and three traffic conditions were incorporated. Data collection had extended to collect three different periods of the week namely peak hour, off-peak hour and weekend to account potential traffic condition. Pearson product-moment correlation was applied to express the correlation between parameters and K-means cluster method in order to classify the microtrips. Through analysis, it was convinced that percentage of idle is most suited to be used to evaluate representativeness of microtrips. These findings could significantly contribute to developing a comprehensive and representative region-specific driving cycle.