Data Analysis of High-Capacity Vehicles by Machine Learning for Sustainable Logistics in Japan

In recent years, the Japanese logistics industry has been facing an increase in freight transportation demand and a serious shortage of truck drivers. To address the labor problems and improve efficiency for sustainable logistics, trucks with double trailers whose length is over 21 meters were introduced. They are called longer and heavier vehicles (LHV) or high capacity vehicles (HCV). In this study, the driving characteristics of the high capacity vehicles will be studied by applying k-means clustering algorithm in machine learning and geographic information system. The data used in this study were obtained from the experimental runs between October 2017 and July 2018, conducted by the Ministry of Land, Infrastructure, Transport and Tourism. Before k-means clustering algorithm is applied, the elbow method is applied to find the optimal number of clusters and the silhouette coefficient is calculated to evaluate the quality of clusters which indicates how well the data are clustered. By k-means clustering, the data are grouped into different clusters. The resultant clusters are visualized in the geographic information system. The clusters are studied and compared how the driving characteristics of the trucks differ in each cluster and how the characteristics correlate to each other. This study is focused on the heart rates and the fluctuations throughout the trip. The outliers of high heart rates and the associated characteristics are identified how they occur and in which areas the drivers can suffer stress. The emphasis is given to the comparison of the drivers’ heart rates recorded near the logistics facilities