ArticlePublisher preview available

Predicting operating speed: comparison of linear regression and structural equation models

Authors:
Behzad Bamdad at Transport & Mobility Leuven
Behzad Bamdad
Babak Mirbaha at Imam Khomeini International University
Babak Mirbaha
Luca Sgambi at Catholic University of Louvain
Luca Sgambi
Ali Abdi Kordani at Imam Khomeini International University
Ali Abdi Kordani
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

The aim of this research was to predict operating speed by considering geometric and roadside factors. Although most previous studies have employed linear regression modelling (LRM) to predict operating speed, this study recommends structural equation modelling (SEM) for the prediction of operating speed on rural multi-lane highways. In addition to geometric variables, LRM takes roadside variables into account. When employing SEM in this work, two latent variables were defined, namely ‘roadside effects’ and ‘geometric effects’. The first latent variable was the combination of land-use type, land-use density and number of accesses per segment, while the second was extracted from the segment length, highway grade, curvature, the presence of a guardrail and flat roadside slope, and the posted speed limits. The residual analysis and R² values for LRM and SEM suggest that SEM demonstrates superior modelling performance compared with LRM. These results show the significant role of latent variables in predicting speed, which cannot be achieved through ordinary LRM.
Figures - available from: Proceedings of the Institution of Civil Engineers - Transport
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Emerald.
Content available from Proceedings of the Institution of Civil Engineers - Transport
This content is subject to copyright. Terms and conditions apply.
Predicting operating speed:
comparison of linear regression
and structural equation models
Behzad Bamdad Mehrabani
PhD candidate, Faculty of Architecture, Architectural Engineering and
Urban Planning, Université catholique de Louvain, Tournai, Belgium
(Orcid:0000-0001-8585-7879) (corresponding author:
behzad.bamdad@uclouvain.be)
Babak Mirbaha
Associate Professor, Civil Engineering Department, Imam Khomeini
International University, Qazvin, Iran (Orcid:0000-0003-1004-7673)
Luca Sgambi
Assistant Professor, Faculty of Architecture, Architectural Engineering and
Urban Planning, Université catholique de Louvain, Tournai, Belgium
(Orcid:0000-0002-4132-2087)
Ali Abdi Kordani
Associate Professor, Civil Engineering Department, Imam Khomeini
International University, Qazvin, Iran (Orcid:0000-0003-3175-2566)
The aim of this research was to predict operating speed by considering geometric and roadside factors. Although
most previous studies have employed linear regression modelling (LRM) to predict operating speed, this study
recommends structural equation modelling (SEM) for the prediction of operating speed on rural multi-lane highways.
In addition to geometric variables, LRM takes roadside variables into account. When employing SEM in this work, two
latent variables were defined, namely ‘roadside effects’and ‘geometric effects’. The first latent variable was the
combination of land-use type, land-use density and number of accesses per segment, while the second was extracted
from the segment length, highway grade, curvature, the presence of a guardrail and flat roadside slope, and the
posted speed limits. The residual analysis and R
2
values for LRM and SEM suggest that SEM demonstrates superior
modelling performance compared with LRM. These results show the significant role of latent variables in predicting
speed, which cannot be achieved through ordinary LRM.
Notation
Xiindependent variables
Yidependent variable
β0,β1,…,βnmodel coefficients
εerror term
1. Introduction
Operating speed is a significant traffic parameter that explains
the effects of the geometric design of roads and roadside
factors on driver behaviour –it is the speed at which drivers
are observed to operate their vehicles during free-flow con-
ditions. The 85th percentile of the distribution of observed
speeds is the measure of the operating speed. Operating speed
has various applications in road traffic and safety studies.
Predicting the safety conditions of a particular highway
depends mainly on operating or average speed (Gargoum and
El-Basyouny, 2016).
Numerous studies have attempted to predict the operating speed
of rural highways. The majority of these studies have aimed to
predict operating speed on rural two-lane highways (Abbas
et al., 2011; Bassani et al., 2015; Bella, 2013; Boroujerdian
et al., 2016; Eboli et al., 2017; Esposito et al., 2011; Garcia
et al., 2013; Hashim et al., 2016; Shallam and Ahmed, 2016;
Yagar and Van Aerde, 1983) rather than on multi-lane highways
(Gargoum and El-Basyouny, 2016; Gong and Stamatiadis,
2008; Himes and Donnell, 2010; Semeida, 2013). Previous
studies have demonstrated that operating speed on rural roads is
determined by various factors, including geometric, traffic and
roadside parameters (Bella, 2013; Boroujerdian et al., 2016;
Garcia et al., 2013; Park and Saccomanno, 2006). Most such
studies have addressed geometric factors (Boroujerdian et al.,
2016; Eboli et al., 2017; Gong and Stamatiadis, 2008; Park and
Saccomanno, 2006; Sil et al., 2020) and, moreover, several
studies have concentrated on curves while neglecting tangents
(Abbas et al., 2011; Echaveguren et al., 2015; Gong and
Stamatiadis, 2008; Hassan and Sarhan, 2011; Park and
Saccomanno, 2006; Shallam and Ahmed, 2016).
In addition, the majority of previous studies have only
considered observed variables and overlooked the effects of
latent variables on operating speed and endogeneity issues.
According to Schumacker and Lomax (2004), latent variables
can only be observed or measured indirectly, and hence are
regarded as constructs inferred from observed variables.
Structural equation modelling (SEM) is a method for
estimating the impacts of latent variables (specified as a linear
combination of observed variables). SEM can handle both
endogenous and exogenous variables and help to investigate
the direct and indirect impacts of exogenous variables on
endogenous variables in causal systems.
In the research described in this paper, two distinct approaches
to modelling to predict operating speed on multi-lane highways
were explored. The first approach attempted to provide a
prediction by employing linear regression modelling (LRM),
which only considers observed variables. In the second
224
Cite this article
Bamdad Mehrabani B, Mirbaha B, Sgambi L and Kordani AA (2023)
Predicting operating speed: comparison of linear regression and structural equation models.
Proceedings of the Institution of Civil Engineers –Transport 176(4): 224–236,
https://doi.org/10.1680/jtran.20.00065
Transport
Research Article
Paper 2000065
Received 21/05/2020;
Accepted 17/09/2020;
First published online 08/10/2020
Emerald Publishing Limited: All rights reserved
Keywords: roads & highways/traffic
engineering/transport planning
Thinking Styles, Conspiracist Belief, and the Mediating Role of the Dunning-Kruger Effect in Modeling Belief Change
Thesis
Full-text available
  • Dec 2023
This study develops a structural model of belief change using the latent constructs of thinking styles and conspiracist belief while accounting for the Dunning-Kruger effect (i.e., overconfidence in one’s knowledge the less one knows about a topic) as a mediator. A combined two-hundred and twenty-six participants from both Amazon’s Mechanical Turk and the introductory psychology SONA subject pool were given two knowledge measures on topics of genetic modification and vaccination before and after reading refutational texts containing current evidence-based information. Belief change was measured as the difference between overall pre- and post- knowledge assessment scores, while The Dunning-Kruger effect was measured by cross-tabulating high-low median split knowledge measure scores with self-reported confidence ratings. A series of questionnaires functioned as measured indicators of the latent Thinking Styles and Conspiracist Belief constructs. Adequate model fit was achieved with the sample data and all paths of the initial structural model were significant. Thinking Styles predicted Conspiracist Belief, which was then predictive of Belief Change via mediation by the Dunning-Kruger effect. Though large path coefficients were not obtained, the significance of this structural model demonstrates important underlying relationships between individual differences and a person’s ability to change incorrect beliefs – something that has become increasingly important in the current era of misinformation and elusive truth.
View
Safety and Economic Effects of Median Guardrails in No-Passing Zone Sections of a Two-Lane Two-Way Rural Highway
Article
  • Jan 2024
In this study, the effect of median guardrails in no-passing zone sections of a two-lane two-way highway was investigated. In this regard, the Firuzkuh-Qaemshahr highway in Mazandaran province of Iran which has 5.5 km median guardrails was selected. All types of crashes studied in two periods, before-and after-the implementation of median guardrails, were considered, and using crash predictive models for the two-way two-lane highway, the effect of implementation was explored. To evaluate the economic effects of this treatment, various crash cost factors were converted to monetary equivalents to compare the cost of crashes before-and after-applying the treatment. This study revealed that using median guardrails caused the monthly average of fatal and injury crashes to decrease by 84% and 29%, respectively. Moreover, the proportion of fatal and injury crashes to total crashes had a 55% drop, indicating that the severity of crashes was reduced, leading to fewer deaths or injuries. Also, this treatment had an impressive reduction in the average monthly cost of crashes, so that the largest reduction among all cost factors belonged to the cost of fatalities in traffic crashes with an 85% reduction, followed by the cost of lost time due to the traffic jams caused by fatal crashes with an 84% decrease. Furthermore, the economic analysis of the implementation of median guardrails revealed that this treatment reduced the average monthly traffic crash cost by 65%. Therefore, the median guardrails in no-passing zone sections of the two-way two-lane Firuzkuh-Qaemshahr highway had a remarkable economic benefit.
View
View
Two-Stream Multi-Channel Convolutional Neural Network for Multi-Lane Traffic Speed Prediction Considering Traffic Volume Impact
Article
Full-text available
  • Mar 2020
Traffic speed prediction is a critically important component of intelligent transportation systems. Recently, with the rapid development of deep learning and transportation data science, a growing body of new traffic speed prediction models have been designed that achieved high accuracy and large-scale prediction. However, existing studies have two major limitations. First, they predict aggregated traffic speed rather than lane-level traffic speed; second, most studies ignore the impact of other traffic flow parameters in speed prediction. To address these issues, the authors propose a two-stream multi-channel convolutional neural network (TM-CNN) model for multi-lane traffic speed prediction considering traffic volume impact. In this model, the authors first introduce a new data conversion method that converts raw traffic speed data and volume data into spatial–temporal multi-channel matrices. Then the authors carefully design a two-stream deep neural network to effectively learn the features and correlations between individual lanes, in the spatial–temporal dimensions, and between speed and volume. Accordingly, a new loss function that considers the volume impact in speed prediction is developed. A case study using 1-year data validates the TM-CNN model and demonstrates its superiority. This paper contributes to two research areas: (1) traffic speed prediction, and (2) multi-lane traffic flow study.
View
A multilevel generalized ordered probit fractional split model for analyzing vehicle speed
Article
Full-text available
  • Dec 2018
Vehicle operating speed plays a significant role in many fields of transportation engineering including safety, operation, design and management. The current research effort contributes to literature on examining vehicle speed on arterial roads methodologically and empirically. Specifically, we propose and estimate a panel mixed generalized ordered probit fractional split (PMGOPFS) model to examine critical factors contributing to vehicle operating speed on roadways. The proposed modeling framework allows for the exogenous variable impacts to vary across the alternatives. Further, the model is formulated to allow for the impact of common unobserved factors across multiple levels (roadway, segment, direction, day and time period). To the best of the authors’ knowledge, this is the first time such an econometric model is proposed and estimated in any literature (not just in transportation). The proposed model is estimated employing a maximum simulated quasi-likelihood based objective function. Vehicular speed data obtained from 8 arterial roads in Orlando for the year 2016 is used for estimating the model. The data is obtained for weekday morning and evening peak and off-peak hours for one randomly chosen week for each roadway throughout the year. The exogenous variables that are considered in the current empirical study include geometry, roadway, traffic, land use and environmental attributes. The model estimation results are further augmented by conducting elasticity analysis to highlight the important factors affecting the vehicular speed profile.
View
Modeling the Operating Speed in Tangents and Curves of Four-lane Highways Based on Geometric and Roadside Factors
Article
Full-text available
  • Jan 2019
Operating speed is an index that represents drivers' speeding behaviors on different highways and shows the comfort and safety levels they experience. Many models had been proposed in previous studies to predict operating speed and most of these works had used geometric, with few of them conducted using roadside variables to predict operating speed. Also, the operating speed study in multilane rural highways had been gained less attention by researchers. In this study, two four-lane rural highways (Kole jub-Borujerd and Borujerd-Khorramabad) had been surveyed for analyzing the operating speed. More than 13,800 spot speed data was gathered in 108 tangent and 30 curve segments. Two linear regression models were developed to predict operating speed in the tangent and curve segments using geometric and roadside factors simultaneously with the acceptable R-squared statistic (0.730 and 0.854 respectively). The results showed that segment length, guardrail median, and flat roadside configuration have a positive effect while slope, accesses density, curvature, and adjacent land use length have a negative effect on operating speed. Moreover, the sensitivity analysis demonstrates that the effect of slope on operating speed in curves is twice comparing to its effect in tangents; while, operating speed in tangents is approximately 2.5 times more sensitive to access density than operating speed in curves. Thus, it can be concluded that not only geometric features affect operating speed but also roadside features affect it. The outcomes of this study can be useful in design and safety planning studies of rural multilane highways.
View
Speed Factors on Two-Lane Rural Highways in Free-Flow Conditions
Article
Full-text available
  • Jan 2005
The mean free-flow speed and its variability across drivers are considered important safety factors. Despite a large body of research on operating speeds, there is still much to learn about the factors of free-flow speeds, especially on tangent segments of two-lane rural highways. The roadway factors of speed dispersion across drivers are largely unknown. Also, the use of the entire free-flow speed distribution suggested by other authors has not yet been addressed. Consequently, the existing models are not aimed to evaluate the speed variability at a site. This paper presents free-flow speed models that identify factors of mean speed and speed dispersion on tangent segments and horizontal curves of two-lane rural highways. Ten highway variables, six of them functioning as both mean speed and speed dispersion factors, were identified as speed factors on tangent segments. Four highway and curve variables, two of them functioning as both mean speed and speed dispersion factors, were identified as speed factors on horizontal curves. The developed free-flow speed models have the same prediction capabilities as traditional ordinary-least-squares models developed for specific percentile speeds. The advantages of the developed models include predicting any user-specified percentile, involving more highway characteristics as speed factors than traditional regression models, and separating the impacts on mean speed from the impacts on speed dispersion.
View
FIXED AND RANDOM EFFECT MODELS TO PREDICT OPERATING SPEED PERCENTILES ON RURAL TWO-LANE HIGHWAYS
Conference Paper
Full-text available
  • Jan 2015
This paper presents the results obtained from the estimation of free-flow speed on two-lane rural highways. The data used for the analysis were collected in the Northwest of Italy using video cameras and a laser speed gun. The model structure adopted separates the estimate of the central tendency of speeds from the typical deviations of individual speeds. Hence, in the model the same set of variables can be used to determine the mean value and the standard deviation of the speed distribution; the desired speed percentile is then calculated by considering the associated standard normal random variable (Z). Random effects (RE) were included in the model to account for the variability in time and space of the data that contains repeated measurements for the same road/section/direction and to remove the dependency between any estimation errors from individual observations. Fixed effect (FE) models are also calibrated for comparison purposes and the BIC criterion is used for variable selection and applied to both the FE and RE models. Estimates reveal many geometric variables in the FE model to be significant, while the RE model only selects the section curvature and the pedestrian density for the estimation of the mean, and nearly all the geometric variables in the estimation of the standard deviation. The random effects relative to the road are not significant, which means that, once the effects related to the random selection of cross sections and direction have been observed, the speed measurements are relatively stable.
View
Operating Speed Models on Horizontal Curves for Two-lane Highways
Article
Full-text available
  • Dec 2016
A variety of physical factors may influence speeds on two-lane highways. In the present study, the highway alignment of Shillong bypass is critically reviewed at several curved section to understand their speed characteristics. The purpose of this paper is to develop curve-speed prediction equations using data collected at 10 horizontal curves of the highway. It was found that the radius of curvature, sight distance and deviation angle had significant relation to the operating speed. Three equations were developed and the radius of curvature had the maximum impact on the operating speed at the circular portion of the horizontal curve, whereas the sight distance and the deviation angle had maximum impact on the operating speed at the entry of the curve.
View
Analysis of Geometric Design Impacts on Vehicle Operating Speed on Two-Lane Rural Roads
Article
Full-text available
  • Dec 2016
The rural road level of service quality has highly related to the vehicles operating speed. The vehicle operating speed has been generally influenced by geometric design characteristics. Previous studies have constructed various analytical models to measure the effect of miscellaneous geometric design variables on vehicle operating speed. This study investigates the main effect of road slopes and its interaction with other variables in Two-Lane rural roads. The vehicles operating speed in 42 sections of Two-Lane rural roads in north of Iran, a recreation area, has been recorded by video. The obtained films are analysed by image processing and operating speeds are measured by an appropriate software with acceptable precision. The distribution of vehicle speed along tangent sections was analysed and it was found speed had normal distribution. So it had two parameters: mean and variance. The relation between these two parameters and geometric characteristics of tangent sections was studied. The results show that the interaction between vehicle classification (passenger car or heavy vehicle) with slope has been significant parameter. Moreover, the initial speed of vehicle at start of tangent section is a significant parameter. This result demonstrates the effect of previous geometric section on vehicle speed. The proposed models are validated by the standard error of estimate and observation versus estimation graph. The models will help road designer to estimate the average operating speed of vehicles along tangent sections and useful for policy makers to enact policies to control vehicles operating speed. The estimated speed can be helpful for analysis the safety of road concepts such as design consistency.
View
Modelling Operating Speeds for Multilane Divided Highways
Chapter
  • Jan 2020
Different geometric parameters (e.g., radius, curve length and preceding tangent length) affects vehicle speed in horizontal curve. Inconsistency in driver’s speed selection may lead to unsafe situation at horizontal curve. The differences between operating speed of successive elements and deviation with design speed of single element have been used as a measure to evaluate geometric design consistency and safety. Researchers have studied speed behaviour considering strong lane discipline to predict preferred vehicle operating speed in two lane highways. However, studies considering weak lane discipline to predict operating speed for multi-lane divided highways are limited. Therefore, in this study, passenger car and heavy vehicle speed data at the starting and centre of ten horizontal curves in a four-lane divided highway are collected. The 85th percentile speeds at eight sites are analysed to develop a linear speed prediction model. Curve length is found to be the only explanatory variable in the model developed for location at the starting of a curve. Whereas, radius is found as the explanatory variable to predict speed at centre of the curve. The developed models are validated at two different sites. Statistical analysis shows that I-value is lesser than 0.2, which confirms the acceptability of the proposed model.
View
Modeling the endogeneity of lane-mean speeds and lane-speed deviations using a Bayesian structural equations approach with spatial correlation
Article
  • Jun 2018
  • TRANSPORT RES A-POL
This study focused on the development of speed prediction models for a multilane highway which incorporate the potential endogenous relationship between adjacent lane speeds and speed deviations while considering geometric design, traffic flow, and other variables in the model specification and accounting for the correlation structures due to multilevel nature of data. The Full Bayesian framework was employed to build the hierarchical models which accounted for three correlation structures at multiple levels: the correlation between speeds of adjoining lanes due to multivariate nature; spatially structured correlations between the adjacent segments, and spatially unstructured correlations among segments. The model estimates which influence the lane-mean speed indicated the directional variation of exogenous factors. For the westbound traffic, the afternoon and night hours were observed to be influential while eastbound traffic was more sensitive to the morning period. The segment length revealed a lane-varying correlation where longer segments influenced a speed reduction for the three lanes closer to the median while the speed in outermost lane exhibited a reverse trend as it increased with longer segments. Both models, with mean speed and speed deviations, demonstrated the significant presence of endogeneity due to mean speeds and speed deviations of adjacent lanes, respectively. This study also assessed the accuracy of predicted mean speed and speed deviations by calculating the measures of discrepancy between the observed and model predicted speeds. The Bayesian residuals, which incorporated the normal, multivariate, and spatial correlation structures, exhibited significant superiority at the prediction accuracy than the Normal ones. This discrepancy in prediction performance reflected the impact of consideration or exclusion of random effects.
View
Structural equations modelling of drivers' speed selection using environmental, driver, and risk factors
Article
  • Sep 2017
  • ACCIDENT ANAL PREV
Drivers' speed has significant implications on road users' safety in general, and particularly so if a crash occurs. This paper explores the influence of environmental and road characteristics, situational factors, and individual characteristics on drivers' observed speed selection in a simulator experiment. The paper presents a theoretical framework for drivers' speed selection, and applies structural equation modeling for the various factors examined. The simulator experiments collected data of 111 drivers driving in 4 different scenarios composed of 22 segments for each scenario. The dataset was analyzed in several resolutions: Driver level, Trip level, and Segment level. The three models revealed that gender, age, and driving frequency are all significant in determining drivers' perceptions and attitudes, which in turn influence speed selection. Situational factors such as traffic speed, enforcement, and time-saving-benefits are also related to speed selection, as well as infrastructure characteristics. These findings demonstrate that structural equations provide a flexible modeling tool able to concurrently analyze the variety of factors that relate to speed selection. As a result, Structural Equations Modeling provides more accurate and refined explanations for the combined effects of various factors on drivers' speed selection than previous research so far. These tools can be useful in developing speed management strategies to improve road safety.
View