Seasonal Factor Assignment Based on the Similarity of Hourly Traffic Patterns and Influential Variables

Abstract

Average Annual Daily Traffic is typically estimated by applying seasonal factors (SFs) to short-term counts. SFs are obtained from continuous count sites and assigned to short-term count sites. This assignment procedure is usually empirical and subjective. Some previous studies have attempted to establish relationships between SFs and influential variables to provide an objective and data-driven alternative for SF assignment. However, in rural areas, SFs are difficult to model due to low land use intensity and, sometimes, significant through traffic. This paper presents a study of relationships between monthly SFs and hourly traffic patterns, land use, and other variables, using data from 116 continuous counters in rural areas throughout Florida. It is found that hourly traffic patterns are related to traffic seasonality and can be used to improve the modeling of influential variables that affect SF. The influential variables are then used for seasonal factor assignment and estimation. The proposed method achieved an average error of four percent, with 95 percent of the estimated monthly SFs having an error of no more than ten percent

Improving Analytical Travel Time Estimation for Transportation Planning Models

“…Aashtiani and Iravani (1999) proposed a method to estimate intersection delays that took into account the significant amount of intersection delay in a large network. The link travel time, t link (x), is estimated using the modified BPR equation with consideration to cycle length and link width, as follows: (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14) where L = link distance (meters); The intersection delay is given below: In the model, street width is used instead of number of lanes; this is due to driving behavior in Tehran, where the number of lanes does not necessarily dictate the number of cars being accommodated across the width of a street (Aashtiani 1999). The significance of this model is that it provides a simple method to approximate cycle length and red time as follows: Link delay and intersection delay are separately expressed as follows: Ding (2007) and Ding et al (2009) proposed an artificial neural network (ANN) model to predict intersection delay based on traffic volumes from all movements of intersection.…”

“…4 Optimized signal plans for an intersection with subject volume of 1,400 vphpl and cross-street volume of 1,400 vphpl.································· 45 Table 3.5 Testing random seeds for the scenario that both traffic volumes from subject direction and cross-street direction are 200 vphpl. ················ 47 Table 3.6 Comparison of delays between CORSIM simulation and existing models.…”