Using Spatial-Temporal Data from Digital Tachographs to Build a Model for Predicting Travel Times – The Example of the Intercity Bus

碩士 === 清雲科技大學 === 經營管理研究所 === 95 === As the amount of traffic rapidly increases, traffic problems are becoming increasingly complex and this has made improving the level of road service an important issue for the current government. In this study, a digital tachograph was used for predicting travel...

Full description

Bibliographic Details
Main Authors: Chang Jia-Ming, 張家銘
Other Authors: Chu Song-Wei
Format: Others
Language:zh-TW
Published: 2007
Online Access:http://ndltd.ncl.edu.tw/handle/19825171142774053282
Description
Summary:碩士 === 清雲科技大學 === 經營管理研究所 === 95 === As the amount of traffic rapidly increases, traffic problems are becoming increasingly complex and this has made improving the level of road service an important issue for the current government. In this study, a digital tachograph was used for predicting travel times on highways because tacographs have lower maintenance costs; can record more detailed data and have wider applications than traditional methods. 60,387,756 pieces of historical data were collected during this study. After filtering this data, 4,744,510 pieces of data were placed in the matrix, which is a large amount of data. A suitable data processing procedure was developed in this study to process this massive amount of historical data. In addition, the processed historical data was sent through regression tree models created in a SQL Server 2005 for each section of the road. Finally, validated data was used to test the model. From this study, after the regression tree predict models derived from regression tree models were tested with validated data and then applied with mean absolute percentage error and root mean square error measurement, it was discovered that the regression tree model for each section of the road had reasonable predictability. It was also found that for some sections, predictability was highly accurate. This model can be provided to commercial vehicle managers in order to help them understand the operation status of their cars and for scheduling backup vehicles in case of problems. This model can also be used to provide travel information to commuters.