Analyzing freeway travel times within a case study: Reliability of route traversal times

Abstract

With the motivation of figuring out the reliability of link and route traversal times, we decompose our research into two stages those summarize modeling of travel times as a link performance and the analysis on the distribution and reliability of reconstructed travel times. In a separate paper, we seek the performance of two models by adopting traffic flow theory based reconstruction approaches. An overall evaluation on freeway travel time reconstruction performance of both a macroscopic and a microscopic model are presented considering real data obtained within a case study conducted on a stretch of Istanbul freeway network. While the initial paper presents the simulation results on both day and flow direction basis, we further consider the results in the explicit analyses on distributions and reliabilities of model reconstructions in the present paper. In order to investigate the distributions of route traversal times by considering explicitly the day-to-day variations, we apply the fitness test of normal distribution to the overall reconstructions by 5 weekday data and then proceed with the standardization of reconstructions and derivation of day-to-day distributions. Statistics obtained as order moments of travel time random variable are considered in the indication of reliability.

Video‐based road traffic monitoring and prediction using dynamic Bayesian networks

“…There are various types of traffic network simulators referenced in the state‐of‐the‐art papers like VISSIM with MATLAB used in [14, 17, 27, 28] and so on but in our work we have used Synchro Studio traffic simulator with MATLAB. Guidelines for the calibration of the road network are very well defined in [18].…”

“…This approach is more inclined towards energy emission models to predict the type of vehicle and its trajectory. This [14] calculates mean, standard deviation and kurtosis to determine the traffic flow, dynamic state of the road by calculating mean square error and other statistical errors. Yuan et al [15] deal with traffic sign detection where as our work is more concentrated towards traffic state prediction.…”

Estimating freeway travel time and its reliability using radar sensor data

“…Link travel times can be estimated by simply extending the point speed measurements to the entire link (Van Lint and Van Der Zijpp, 2003;Soriguera and Robusté, 2011b;Bovy and Thijs, 2000). Moreover, to capture the traffic dynamics along the link, several methods have been proposed to estimate travel time based on kinematic wave theory and other traffic flow theories (e.g., Van Arem et al, 1997;Coifman, 2002;Zhang, 2006;Kesting and Treiber, 2008;Deniz et al, 2013;Aksoy and Celikoglu, 2012). Some travel time estimation methods are essentially based on flow conservation and propagation principles (e.g., Celikoglu 2007;Celikoglu 2013a,b).…”

Finding the ‘faster’ path in vehicle routing

“…The research of a dynamic network loading model for travel time estimation has been developed in [19–21]. Additionally, Deniz et al [22] conduct reliability evaluation of a macroscopic and a microscopic traffic model, respectively, analysing the distribution of reconstructed travel time by a random variable approach and comparing its statistical indicators. Celikoglu et al [23] use dynamic classification and clustering technique to categorise flow patterns, evaluating the performance of a macroscopic flow model.…”