Day-wise Travel Time Pattern Analysis Under Heterogeneous Traffic Conditions

Abstract

Traffic parameters in general follow some patterns and identifying them is important since it will lead to more efficient end applications. With the introduction of various automated sensors, huge amounts of data are being collected, which can be used for identifying traffic patterns. In general, traffic patterns can be classified as yearly, monthly, weekly, daily and hourly. Travel time has been recognized as one of the most important parameters to fully facilitate many of the Intelligent Transportation Systems (ITS) applications. However, to predict the travel time, its weekly, daily and hourly patterns need to be analyzed. The present study analyzes the pattern followed by Global Positioning System (GPS) based bus travel time data. The travel time pattern may be different for different days of the week and hence the analysis was carried out separately for each day of the week. A systematic statistical analysis was carried out to rank these patterns in the order of significance for each day of the week separately. The statistical test namely, the Z-test for the mean of a population of differences for „paired‟ samples data, was used for hypothesis testing at 5% level of significance using one month‟s data (with a total sample size of 5700). It was observed that all days have a similar pattern except Sunday.