Travel Time Variability Analysis for Bluetooth Sensor Data in Highways

EasyChair Preprint no. 4377

Abstract

Travel time variability and travel time reliability are the two sides of one coin. Increasing travel time variability will lead to a decrease in the predictability of travel time, therefore, travel time reliability will decline too. Distribution of travel time data is characterized to mainly assess the variability of travel time. Variation in day-to-day travel time deteriorates the transportation system reliability by trespassing the user’s expectations. Variability of travel time is time-dependent and differently affects the departure time choice behavior of travelers at different times of the day. The most current researches on pricing strategies are based on the premise that the preferences of travelers depend on the time periods, such as the morning peak hour. This research is aimed at providing a comprehensive model based on the division of time data with equal time intervals. This model is called the conditional model function with the random effect function. This model is built around the multiplication of the probability density function of each time interval with a random-effects. The causality of using random owing to the interface of each time interval random interaction is seen for another time interval. Current state-of-the-art travel time variability researches assume that travel times follow a unimodal distribution during the off-peak hour and bimodal distribution such as lognormal and gamma mixture during peak-hour. Particularly, a Burr, Weibull, or lognormal distributions are unimodal distributions employed to characterize the system reliability. Dividing a day time period into constant time frames to pursue travel time variability is the main technique employed by the previous studies. However, this ignores the interconnected nature of the travel time among the different time periods of a day. To bridge this gap, the current study presents a conditional likelihood model that relates to the different time windows using a random-effects model.

Keyphrases: Bluetooth sensor, Random effects model, Travel time distribution, travel time reliability, travel time variability

@Booklet{EasyChair:4377,
  author = {Mohammad Ghavidel and Navid Khademi and Ehsan Bahrami Samani and Sadaf Milani},
  title = {Travel Time Variability Analysis for Bluetooth Sensor Data in Highways},
  howpublished = {EasyChair Preprint no. 4377},

  year = {EasyChair, 2020}}