Car Drivers Do Not Choose Their Speed in Urban Environments:

Speed Models in Tangent Streets

Yasmany García-Ramírez

, Luis Paladines

, Christian Verdesoto

and Patricio Torres

Civil Engineering Department, Universidad Técnica Particular de Loja, San Cayetano Street, Loja, Ecuador

Keywords: GPS Equipment, Roundabout Intersections, Signal-controlled Intersections, Speed Model, Stop-controlled

Intersections, Urban Tangent Streets.

Abstract: The performance-based design approach is one way to deal with speeding in the streets. Under this approach,

the geometric elements of roadways can influence on the desired operating speeds. Thus, several studies have

investigated the relationship between geometric elements and light vehicle speeds; however, no conclusive

results have been reached at this stage. In this context, this article aims to investigate the influence of several

characteristics from urban street tangents, car driver, and vehicle on their speed in free-flow conditions. Three

tangents scenarios were set out: before stop-controlled intersections, before signal-controlled intersections,

and before roundabout intersections. Speeds of light vehicles were measured at 34 streets. Speeds were

collected with in-vehicle GPS equipment. Thirty-five car drivers participated in the study with their vehicles.

Street geometric characteristics, street environment variables, driver and vehicle characteristics were also

collected. As a result, 15 regression models were calibrated and validated. Street length and objects density

were the most influential variables in those models, and not the driver and vehicle characteristics as would

suppose. This comprehensive research extends the knowledge of the most influential variables on speed in

several urban scenarios, offering useful information for urban planners and street designers.

1 INTRODUCTION

Performance-based design is one approach to deal

with speeding in the streets, where the street

geometrics and its environment elements are selected

based on their influence on the desired driving speeds,

especially in tangents. This approach presents a more

active and efficient alternative to reduce speed

vehicles (Fitzpatrick et al., 2003; Harwood et al.,

2000; Ray et al., 2014). The success of this approach

is finding the relationship between driving speeds and

street features. Several studies have developed

operating or free-flow speed models for urban streets

in order to understand this relationship. These

investigations were focused especially on tangents

before stop-controlled intersections and tangents

before roundabout intersections.

In general, five principle parameters influencing

the free flow speed: driver, vehicle, roadway,

environment, traffic operation and control (Sekhar et

https://orcid.org/0000-0002-0250-5155

https://orcid.org/0000-0002-2070-6946

https://orcid.org/0000-0002-9344-9506

https://orcid.org/0000-0001-5813-8844

al., 2016). Those variables also may influence

acceleration and deceleration choice. Speed choice is

influenced by driver characteristics, such as

personality traits (García-Ramírez, 2014; Gargoum et

al., 2016; Roidl et al., 2014), driver age (Keay et al.,

2013; Thompson et al., 2012), driver reliability

(Gstaltera and Fastenmeier, 2010), conversation and

texting tasks (Choudhary and Velaga, 2017), gender

and driving experience (Goralzik and Vollrath, 2017),

speeding intention (Dinh and Kubota, 2013b), threat-

related feelings and arousal (Schmidt-Daffy, 2013),

among others (Tarris et al., 1996).

Vehicle characteristics also impact on speed

choice, such as vehicle class (Dhamaniya & Chandra,

2013; Gargoum et al., 2016; Jevtić et al., 2015; Wang,

2006) vehicle age (Gargoum et al., 2016) or vehicle

length (Giles, 2004).

Roadway also influence on speed choice, such as

street length (Dinh and Kubota, 2013a; Wang, 2006),

number of lanes (Dinh and Kubota, 2013a; Eluru et

García-Ramírez, Y., Paladines, L., Verdesoto, C. and Torres, P.

Car Drivers Do Not Choose Their Speed in Urban Environments: Speed Models in Tangent Streets.

DOI: 10.5220/0010435904210428

In Proceedings of the 7th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2021), pages 421-428

ISBN: 978-989-758-513-5

421

al., 2013; Wang, 2006), shoulders (Eluru et al., 2013;

Gargoum and El-Basyouny, 2016), longitudinal slope

(Ding et al., 2015), roadway width (Bassani and

Sacchi, 2012; Dinh & Kubota, 2013a), lane width

(Bassani and Sacchi, 2012; Poe and Mason, 2000),

pavement markings (Ding et al., 2015; Guo et al.,

2016), pavement condition (Eluru et al., 2013; Sekhar

et al., 2016; Wang, 2006). In roundabouts, speed

choice were related to entry width (Al-Omari et al.,

2014; Gallelli et al., 2014), internal circle diameter

(Al-Omari et al., 2014; Gallelli et al., 2014), drive

curve (Al-Omari et al., 2014), entry deviation angle

(Al-Omari et al., 2014) and approach/exit speed

(Gallelli et al., 2014).

Street environment impacts on driver speed

choice too, specially sidewalk presence (Dinh &

Kubota, 2013a; Eluru et al., 2013; Wang, 2006),

access density (Wang, 2006), roadside objects density

including trees, utility poles, traffic signs, etc.

(Antonson et al., 2014; Dinh and Kubota, 2013a;

Wang, 2006), parking presence (Eluru et al., 2013;

Gargoum et al., 2016; Wang, 2006), crash barriers

(Antonson et al., 2014), bus stop (Antonson et al.,

2014), adjacent land uses (Galin, 1981; Gargoum et

al., 2016; Giles, 2004; Wang, 2006) or open landscape

and broad road (Antonson et al., 2014). On the other

hand, speeds are also significantly affected by

changes in lighting conditions, such as: sunny, cloudy,

and dark (Bassani et al., 2016; Hjelkrem & Ryeng,

2016) as well as weather conditions (Rahman &

Lownes, 2012). Likewise, the day of the week (Eluru

et al., 2013) or nighttime and daytime (Tarko et al.,

2016) also would affect driver speed choice.

Traffic operation and control could affect speed

choice, for example, speed limits (Gargoum & El-

Basyouny, 2016; Goralzik and Vollrath, 2017), point-

to-point speed enforcement system (Montella et al.,

2015), speed cameras (Schechtman et al., 2016),

speed publicity campaigns (Schagen et al., 2016),

perceived posted speed (Schechtman et al., 2016) or

photo-radar presence (Chen et al., 2000). It should be

mentioned that given the scope of the investigation,

this parameter will not be included in the analysis.

In summary, there is no doubt that driver speed

choice is affected by a range of different attributes of

driver, vehicle, roadway, environment, traffic

operation and control. However, some of those

attributes were not statically significant in all the

studies, for example, lane width (Wang, 2006) or

pavement condition (Bassani and Sacchi, 2012) did

not influence on speed choice; which means that more

research is needed.

In this context, the aim of this research is to

investigate the influence of several characteristics of

urban tangent streets, car driver, and its vehicle on the

driver speed choice in free-flow conditions. Fifteen

regression models were calibrated and validated and

three different scenarios. To show these findings, the

rest of the article is organized as follows. Section 2

gives an overview of the materials and methods. This

section describes the sample size, road test section,

measurement equipment, the selection of car drivers,

and the selection of the test vehicles. Also, this

section details the data collection, data processing,

and speed pattern analysis. Later, the results section

presents the speed model calibration process and its

validation. And the principal conclusions are

highlighted at the end of the article.

2 MATERIALS AND METHODS

2.1 Sample Size

The speed sample size was calculated based on

equation 1 (Pignataro, 1973). This equation includes

several speed percentiles, the speed standard

deviation, a level of confidence and an admissible

error. For this investigation, a standard deviation of

13 km/h was assumed (Bennett, 1994) and an error of

5 km/h.

𝑛

𝐾



∗𝜎



∗



2𝑈





2∗𝑒𝑙



(1)

Where n: simple size, K: constant related to the

confidence level, σ: standard deviation, U: normal

deviation related to the percentile of speed, el:

maximum admissible error. With a level of

confidence of 95% (K = 1.96) and value of U = 1.04

(for the worst-case scenario: 85th percentile speed), a

minimum number of 40 observations are obtained

with that equation. In this study, in just the yellow

signal, that value could not be reached (only 13

observations were collected).

2.2 Road Test Section

Streets for this study were selected based on the

following criteria: a) urban streets, b) tangents before

stop-controlled intersections, tangents before signal-

controlled intersections or tangents before

roundabout intersections, c) longitudinal slope less

than 3%, d) pavement surface in good condition, e)

speed limit of 60 km/h, and f) allow free-flow

condition. Based on those criteria, 34 streets were

selected in Loja city (Ecuador), where 13 streets were

tangents before stop-controlled intersections, 12 were

tangents before signal-controlled intersections, and 9

were tangents before roundabout intersections. In

order to optimize resources, all those streets were part

of a single circuit of 9.8 km.

VEHITS 2021 - 7th International Conference on Vehicle Technology and Intelligent Transport Systems

422

Regarding the characteristics of the tangents before

stop-controlled intersections, the length of tangents

before stop-controlled intersections was between 47-

226 m, up to 2 lanes, roadway width between 7 to 9

m, and up to traffic directions. In tangents before

signal-controlled intersections, the length of the

streets was between 94 and 121 m, up to 3 lanes,

roadway width between 7 to 10 m, and up to traffic

directions. In tangents before roundabout inter-

sections, the length of the streets was between 63 and

830 m, up to 3 lanes, the roadway width between 7

and 13 m, the traveled length within the roundabout

was between 17 and 118 m, the internal diameter was

between 11.5 and 25 m, the external diameter was

between 26 and 61 m, the width of entry and exit of

the roundabout was between 7 and 11 m, and the

roadway width inside the circle was between 9 and 14

2.3 Measurement Tool

Video VBOX Lite was selected to collect vehicle

speeds. This device was placed inside the light

vehicle. It allows recording geo-referenced digital

images, its speed, and its height, among others. The

Video VBOX Lite has an accuracy of 0.05% for

distance travelled, 0.2 km/h for speed, and ± 10 m for

height. The device, in movement, receives

information every 0.1 s from 8 satellites. The geo-

referenced digital images helped with the data

analysis. Independent variables, such as street length,

were collecting using traditional measures.

2.4 Driver and Car Selection

Car drivers who participated in this investigation

were selected from a non-probabilistic sampling.

They had to meet the following requirements: a) have

a driver's license, b) have a light vehicle, c) know the

streets of the study, and d) have driven frequently in

the last two months. Based on these restrictions, 23

men and 22 women were chosen. This distribution

was proportional to the last population census in the

country (INEC, 2010). All car drivers had an average

age of 30.5 years (min = 21, max = 60) and driving

experience of 9.3 years (min = 1, max = 40). At the

end of the trip, drivers answered two surveys: MDSI-

S (Taubman-Ben-Ari et al., 2004) and ZKPQ-50-cc

(Aluja et al., 2006) to estimate their personality traits

and driving style, respectively. This information was

used to analyse their influence on their speed choice.

A previous research found a statistical relationship

between maximum speed and certain personality

traits and driving styles on rural roads (García-

Ramírez, 2014).

Vehicles in this research were mostly Chevrolet

and Hyundai (62%). From all vehicle study, 78%

were cars and 22% pickup trucks. Average

manufacturing year was 2008 (min = 1994, max =

2015), average cylinder capacity was 1850 cm

(min

= 1000, max = 3700) and average last mechanical

check-up was 52 days (min = 15, max = 140) before

the day of data collection.

2.5 Speed Data Collection

Speeds were collected in good weather conditions,

dry pavement and during daylight. It selected the

weekends and outside of peak hours (2:00 a.m. to

6:00 p.m.) as study time, to ensure that streets were in

a free-flow condition. Video VBOX Lite was

discretely installed in each vehicle, with the

precaution of not interfering with the driving task.

The device has a GPS antenna and a camera. GPS

antenna was placed in the central part of the vehicle

roof and the camera was placed on the front

windshield, facing the street. During the device

installation, the driver was briefly explained about the

circuit and the academic use of the speed data.

2.6 Data Processing

After the collection data, position data, distance

travelled, accumulated distance and speed were

exported every second. Every speed profile, o part of

it, which was not in free-flow condition, was

eliminated. In tangents before stop-controlled

intersections, there were 21 free-flow speed profiles

for every street. In tangents before signal-controlled

intersections there were 67 free-flow speed profiles

when the traffic light was green, 45 in red light and

13 in yellow light. Free-flow speed profiles tangents

before roundabout intersections were 90, while the

free-flow speed profiles within the roundabout were

125. In each street, in its middle, the operating speed

or the 85th percentile speed, mean free-flow speed,

and free-flow speed standard deviation were

calculated, as well as considered in the previous

literature.

At the end of the trip, car drivers were asked to

answer two surveys: ZQPK-50-cc, and MDSI-S. The

ZQPK-50-cc survey has 50 questions related to the

five traits of personality: aggression - hostility,

impulsive sensation seeking, neuroticism - anxiety,

sociability, and activity. MDSI-S survey (41

questions) estimates the driving style that prevails in

the driver: risky and high-velocity style, dissociative

style, angry style, careful and patient styles, anxious

style, or distress reduction style. According to the

Car Drivers Do Not Choose Their Speed in Urban Environments: Speed Models in Tangent Streets

423

results, the majority of the drivers had the following

predominant personality traits: impulsive sensation

seeking (44.4%) and activity (42.2%). Likewise, most

of the drivers were careful and patient style (53.3%)

or risky and high-velocity style (26.7%).

2.7 Pattern Analysis

After data processing, this section analyses the

patterns of the independent variables related to the

speed, in order to detect the most influential variables.

All statistical analyses were performed using the R

program (R Core Team, 2013). In this software, a

linear regression analysis was performed at 95% level

of confidence. The variables statistically significant

from this process will use in the equation calibration

process.

The variables analysed in this process were: street

length (m), roadway width (m), lane width (m), land

use, objects density (n°/100 m), trees density (n°/100

m), access density (n°/100 m), number of lanes, and

parking and sidewalk presence. In tangents before

stop-controlled intersections the street length

influences the operating speed (v

), mean free-flow

speed (v

AVG

), free-flow speed standard deviation

). Also, the parking presence affected the v

AVG

On the other hand, in the tangents before signal-

controlled intersections, the lane width influences the

in the green light, and the objects density (n°/100

m) affects the v

in the red light. No variable was

statistically significant in the yellow light. In tangents

before roundabouts intersections, the street length

influences the operating speed (v

), and the mean

free-flow speed (v

AVG

). Variables related to the driver

and to the vehicle were not statistically significant.

3 RESULTS

Speed model calibration was performed based on the

most influential variables. In this calibration, a linear

regression analysis was performed with a 95% level

of confidence. When there were not any statistically

significant variables, fixed values were assumed.

After the calibration, models were validated with data

in another test circuit, which had similar

characteristics to the initial one. This validation was

carried out by analyzing the prediction errors.

3.1 Models Calibration

It calibrated a linear regression analysis with the

street length (see table 1).

Table 1: Proposed models for tangents before stop-

controlled intersections.

Condition Equation R

adj. #

Speed in the

middle of

the tangent

= 22.4 + 0.114 L

0.94 (2)

AVG

= 20.1 + 0.105 L

0.95 (3)

= 1.99 + 0.0146 L

0.79 (4)

𝑣



 operating speed in km/h, 𝑣



 mean free-flow speed in

km/h, 𝑣



 free-flow speed standard deviation in km/h, 𝐿

street length between 47 and 226 m, 𝑅



𝑎𝑑𝑗.  adjusted

coefficient of determination.

Vehicle speed in table 1 increases with the length

of the street when the driver perceives a long distance

to travel. Conversely, in shorter paths, drivers will

have smaller speeds because they do not have “the

physical space” to do both speed up in the tangent and

speed down before to reach the intersection.

In tangents before signal-controlled intersections

are more complex than the previous tangents because

traffic lights increase driver mental workload. The

calibrated models shown in table 2 are also consistent

with actual driving. In green light, vehicles are slower

than in yellow light, which makes sense, because

many drivers tend to speed up to pass the yellow

traffic light, especially when they are in the dilemma

zone (Bar-Gera et al., 2016).

Table 2: Proposed models for tangents before signal-

controlled intersections.

Condition Equation R

adj. #

Speed in the

middle of the

tangent in

green light

= 52.72 km/h

NA (5)

AVG

= 43.55 km/h

NA (6)

= 4.34 km/h

NA (7)

Speed in the

middle of the

tangent in

yellow light

= 56.28 km/h

NA (8)

AVG

= 43.31 km/h

NA (9)

= 9.82 km/h

NA (10)

Speed in the

middle of the

tangent in red

light

= 33.4+0.53OD

0.49 (11)

AVG

= 39.71 km/h

NA (12)

= 4.45 km/h

NA (13)

𝑣



 operating speed in km/h, 𝑣



 mean free-flow speed in

km/h, 𝑣



 free-flow speed standard deviation in km/h, 𝐿

street length between 94 and 122 m, 𝑂𝐷  object density

between 5.3 and 29.1 units per each 100 m , 𝑅



𝑎𝑑𝑗.  adjusted

coefficient of determination, NA: not available.

In this dilemma zone can also see high-speed data

dispersion because other drivers decelerate with high

rates in order to stop in yellow light. When the light

VEHITS 2021 - 7th International Conference on Vehicle Technology and Intelligent Transport Systems

424

is green, drivers have already decided that they will

pass the intersection, so the speed data dispersion is

lower than in yellow light. In the red light, speeds are

lower than the two previous conditions, because

drivers already know that they must stop before the

red light. In this condition, speed data dispersion

should be the lowest.

Models in roundabouts (see in table 3) are

applicable for tangent lengths between 63 to 312 m.

In longer tangents (> 312 m) could be assumed that

drivers have reached their desired speed. In this study,

the average desired speed was 44.80 km/h with a

standard deviation of 5.90 km/h. This average value

is similar to the result of the equation 16, using a street

length value of 312 m (45.46 km/h). This double

check strengthens confidence in the equations

developed.

The speed values from table 3 are higher than those

found in the previous scenarios (stop-controlled and

signal-controlled); because in tangents before

roundabouts, the driver has a less mental workload.

Speed data dispersion is higher than in green light and

the red light cases, as what has been seen in the yellow

light case. This is because roundabout could generate

a dilemma zone, given that the driver may doubt if

he/she continues or stops the vehicle when

approaching vehicles inside the roundabout.

Street geometric and street operation variables

were also analysed inside the roundabouts. However,

there were not any variable statistically significant, so

fixed values were adopted: an average speed of 28.60

km/h and a standard deviation of 4.66 km/h. This

average speed was similar than in previous

investigations: 30 km/h (Bassani & Sacchi, 2011), 17-

26 km/h (Gallelli et al., 2014), as well as its standard

deviation 4.13-5.21 km/h (Gallelli et al., 2014).

Table 3: Proposed models for tangents before roundabouts.

Condition Equation R

adj. #

Speed in the

middle of

the tangent

= 28.3 + 0.091 L 0.98 (14)

AVG

= 20.5 + 0.080 L 0.90 (15)

= 5.38 km/h NA (16)

𝑣



 operating speed in km/h, 𝑣



 mean free-flow speed in

km/h, 𝑣



free-flow speed standard deviation in km/h, 𝐿𝑠treet

length between 63 and 312 m, 𝑅



𝑎𝑑𝑗.  adjusted coefficient of

determination, NA: not available.

3.2 Models Validation

A validation process was performed to evaluate the

calibrated models from tables 1 to 3. For this

validation, another circuit was collected in the same

city, with similar street characteristics. Information

was collected from 8 streets with a length between 47

m and 112 m for tangents before stop-controlled

intersections, 12 streets with a length between 94 to

120 m for tangents before signal-controlled

intersections, and 6 streets with a length between 66

to 287 m for tangents before the roundabout.

Six car drivers (3 were men and 3 were women)

drove through the validation circuit. The drivers had

an average age of 26.3 years (min = 25, max = 28)

and an average driving experience of 7.2 years (min

= 5, max = 9). Chevrolet or Hyundai brands (67%)

were the vehicles that participated in the calibration

circuit. Cars were 67% and pickup trucks were 33%.

The average manufacturing year was 2007 (min =

2004, max = 2011), the average cylinder capacity was

2000 cm3 (min = 1400, max = 2700), and the average

last mechanical check-up was 53 days before the

collection day (min = 30, max = 90). Both the

measurement equipment and the time data collection

were the same in the initial circuit. Also, the same

data processing from the calibration process was used

for the calibration process.

Prediction errors were calculated in order to

validate the previous calibrated speed models. Those

errors were: mean squared error (MSE), mean

absolute error (MAE), mean absolute percentage

(MAPE), and Chi-square test (see table 4). Table 4

did not include the constant values models because is

not possible to get prediction errors, however, in

those cases, an analysis of variance (ANOVA) was

carried out, comparing values from the constant value

models with the collected values from validation.

And those values should not differ at 95% level of

confidence.

The highest values from table 4 are MSE and

MAE. Those errors are obtained by the model of

operating speed and the mean free-flow speed for the

tangents before roundabouts. When these equations

are used, caution should be taken because prediction

error was around 5 km/h.

The highest values of the MAPE is given by the

equations of the free-flow speed standard deviation,

so caution is also suggested when will be used.

Despite the values of those errors, these equations and

the other is table 4, equations are valid because the

chi-calculated did not pass the chi-critical value.

Regarding to the fixed models, the p-value from

ANOVA is more than the level of significance

(α=0.05), which means that the differences between

the means are not statistically significant, so also

those fixed models are valid.

Car Drivers Do Not Choose Their Speed in Urban Environments: Speed Models in Tangent Streets

425

Table 4: Prediction errors and chi-square values for speed,

acceleration, and deceleration equations for the

intersections in this study.

# equation

Prediction errors

p value

(ANOVA)

MSE

(km/h)²

MAE

(km/h)

MAPE

(%)

χ²

calculated

χ²

critic

(2) 3.83 1.27 4.04 0.95 14.07 -

(3) 5.85 2.13 7.69 1.68 14.07 -

(4) 0.82 0.70 22.84 2.11 14.07 -

(5) - - - - - 0.804

(6) - - - - - -

(7) - - - - - -

(8) - - - - - 0.321

(9) - - - - - -

(10) - - - - - -

(11) 8.77 2.44 5.70 1.41 12.59 -

(12) - - - - - 0.245

(13) - - - - - -

(14) 36.79 5.44 11.60 4.54 11.07 -

(15) 26.60 5.11 14.39 4.44 11.07 -

(16) - - - - - 0.560

MSE= mean squared error, MAE=mean absolute error,

MAPE=mean absolute percentage error,χ² calculated =Chi

-square calculated, χ² critic = Chi-square critic. P value:

results from ANOVA. -: Not performed o not possible to

calculate.

It is worth mentioning that a standard deviation of

13 km/h was adopted for the calculation of the sample

size, while the standard deviation in this study was

5.7 km/h. With this standard deviation, the sample

size will be 8 observations, less the minimum used in

this study. So, the sample size of this study has more

than 95% level of confidence.

4 CONCLUSIONS

This article aimed to investigate the influence of

several characteristics from urban street tangents, car

driver, and vehicle on their speed in free-flow

conditions. After analyzing the results, the following

conclusions are presented:

In tangents before stop-controlled intersections,

the street length was the most influential variable on

speed. In tangents before signal-controlled

intersections, object density was the only variable that

influenced the speed in red light. In roundabout

intersection and tangents before the roundabout, the

street length influenced the speed. The average

desired speed in tangents was 44.80 km/h, and the

average speed within the roundabouts was 28.60

km/h. The calibrated and validated models are

consistent with what happens in real driving

condition. Considering this finding, apparently car

drivers do not choose their speed; nevertheless, the

car driver is indirectly influenced by the street length.

If they perceive that the urban street is long they will

speed up, and in short streets, they will do the

opposite or the will keep their initial speed.

This study has several limitations. First, Video

VBOX Lite was used based on the assumption that

this device gives accurate speed data; which should

be studied in the future. Also, all speed models are

valid for speeds in the middle of the tangent, thus,

other points in the street speed profile should be

analyzed. Also, the calibrated speed models are valid

in a specific range, so it should be used in those

ranges. The circuits were located in an Andean city,

which could differ from other cities.

Despite these limitations, the present study helps

to extend the knowledge on urban speeds and their

relationship with street variables, offering useful

information for urban planners and street designers. It

studied three different scenarios and several variables

related to car driver characteristics, vehicle

characteristics, roadway, and street environment. It

showed that the street length is the main variable that

affects the speed in urban tangents and not the

variables related to the vehicle and driver. This

outcome suggests that the government, especially in

developing countries, should put more emphasis on

street infrastructure than on the driver or vehicle.

ACKNOWLEDGEMENTS

The authors acknowledge the support of the National

Secretariat of Higher Education, Science,

Technology and Innovation (SENESCYT) and

Universidad Técnica Particular de Loja from the

Republic of Ecuador.

REFERENCES

Al-Omari, B. H., Ghuzlan, K. A., & Al-helo, L. B. (2014).

Modeling through traffic speed at roundabouts along

urban and suburban street arterials. Journal of the

Transportation Research Forum, 53(2), 7–19. https://

trforum.org/wp-content/uploads/2017/05/2014v53n2_

01_Roundabouts.pdf.

VEHITS 2021 - 7th International Conference on Vehicle Technology and Intelligent Transport Systems

426

Aluja, A., Rossier, J., García, F., Angleitner, A., Kuhlman,

M., & Zuckerman., M. (2006). A cross cultural shortened

form of the ZKPQ (ZKPQ-50-cc) adapted to English,

French, German, and Spanish languages. Personality and

Individual Differences, 41(4), 619–628. https://

doi.org/http://dx.doi.org/10.1016/j.paid.2006.03.001.

Antonson, H., Ahlström, C., Mårdh, S., Blomqvist, G., &

Wiklund, M. (2014). Landscape heritage objects’ effect

on driving: A combined driving simulator and

questionnaire study. Accident Analysis and Prevention,

62, 168–177. https://doi.org/10.1016/j.aap.2013.09. 021.

Bar-Gera, H., Musicant, O., Schechtman, E., & Ze’evi, T.

(2016). Quantifying the yellow signal driver behavior

based on naturalistic data from digital enforcement

cameras. Accident Analysis and Prevention, 96, 371–

381. https://doi.org/10.1016/j.aap.2015.03.040.

Bassani, M., Catani, L., Cirillo, C., & Mutani, G. (2016).

Night-time and daytime operating speed distribution in

urban arterials. Transportation Research Part F:

Traffic Psychology and Behaviour, 42, 56–69.

https://doi.org/10.1016/j.trf.2016.06.020.

Bassani, M., & Sacchi, E. (2012). Calibration to local

conditions of geometry-based operating speed models

for urban arterials and collectors. Procedia - Social and

Behavioral Sciences, 53, 821–832. https://doi.

org/10.1016/j.sbspro.2012.09.931.

Bassani, M., & Sacchi, E. (2011). Experimental

investigation into speed perfomance and consistency of

urban roundabouts: an Italian case study. Poster

Session at the “3rd International Conference on

Roundabouts,” May.

Bennett, C. R. (1994). A speed prediction model for rural

two-lane highways (Phd Thesis). Auckland University.

Chen, G., Wilson, J., Meckle, W., & Cooper, P. (2000).

Evaluation of photo radar program in British Columbia.

Accident Analysis & Prevention, 32(4), 517–526.

https://doi.org/https://doi.org/10.1016/S0001-4575(99)

00071-8.

Choudhary, P., & Velaga, N. R. (2017). Mobile phone use

during driving: Effects on speed and effectiveness of

driver compensatory behaviour. Accident Analysis and

Prevention, 106(June), 370–378. https://doi.org/10.1016

/j.aap.2017.06.021.

Dhamaniya, A., & Chandra, S. (2013). Speed Prediction

Models for Urban Arterials Under Mixed Traffic

Conditions. Procedia - Social and Behavioral Sciences,

104(Hcm 2010), 342–351. https://doi.org/10.

1016/j.sbspro.2013.11.127.

Ding, H., Zhao, X., Rong, J., & Ma, J. (2015). Experimental

research on the effectiveness and adaptability of speed

reduction markings in downhill sections on urban

roads: A driving simulation study. Accident Analysis

and Prevention, 75, 119–127. https://doi.org/10.101

6/j.aap.2014.11.018.

Dinh, D. D., & Kubota, H. (2013a). Profile-speed data-

based models to estimate operating speeds for urban

residential streets with a 30km/h speed limit. IATSS

Research, 36(2), 115–122. https://doi.org/

10.1016/j.iatssr.2012.06.001.

Dinh, D. D., & Kubota, H. (2013b). Speeding behavior on

urban residential streets with a 30km/h speed limit

under the framework of the theory of planned behavior.

TransportPolicy, 29, 199–208. https://doi.org/

http://dx.doi.org/10.1016/j.tranpol.2013.06.003.

Eluru, N., Chakour, V., Chamberlain, M., & Miranda-

Moreno, L. F. (2013). Modeling vehicle operating

speed on urban roads in Montreal: A panel mixed

ordered probit fractional split model. Accident Analysis

and Prevention, 59, 125–134.

https://doi.org/10.1016/j.aap.2013.05.016.

Fitzpatrick, K., Carlson, P., Brewer, M., Wooldridge, M.,

& Miaou, S. (2003). NCHRP Report 504: Design

Speed, Operating Speed, and Posted Speed Practices.

http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_rpt_

504.pdf.

Galin, D. (1981). Speeds on two-lane rural roads-a multiple

regression analysis. Traffic Engeniering Control, 22.

Gallelli, V., Vaiana, R., & Iuele, T. (2014). Comparison

between simulated and experimental crossing speed

profiles on roundabout with different geometric

features. Procedia - Social and Behavioral Sciences,

111, 117–126. https://doi.org/10.1016/j.sbspro.2014

.01.044.

García-Ramírez, Y. (2014). Aceleraciones y desacele-

raciones de vehículos livianos en caminos de montaña

(Doctoral disertation) [Universidad Nacional de San

Juan]. http://repositorio.educacionsuperior. gob.ec/

Gargoum, S. A., & El-Basyouny, K. (2016). Exploring the

association between speed and safety: A path analysis

approach. Accident Analysis and Prevention, 93, 32–

40. https://doi.org/10.1016/j.aap.2016.04.029.

Gargoum, S. A., El-Basyouny, K., & Kim, A. (2016).

Towards setting credible speed limits: Identifying

factors that affect driver compliance on urban roads.

Accident Analysis and Prevention, 95, 138–148.

https://doi.org/10.1016/j.aap.2016.07.001.

Giles, M. J. (2004). Driver speed compliance in Western

Australia: a multivariate analysis. Transport Policy,

11(3), 227–235. https://doi.org/https://doi.org/

10.1016/j.tranpol.2003.11.002.

Goralzik, A., & Vollrath, M. (2017). The effects of road,

driver, and passenger presence on drivers’ choice of

speed: a driving simulator study. Transportation

Research Procedia, 25, 2061–2075.

https://doi.org/10.1016/j.trpro.2017.05.400.

Gstaltera, H., & Fastenmeier, W. (2010). Reliability of

drivers in urban intersections. Accident Analysis &

Prevention

, 42(1), 225–234.

Guo, Y., Liu, P., Liang, Q., & Wang, W. (2016). Effects of

parallelogram-shaped pavement markings on vehicle

speed and safety of pedestrian crosswalks on urban

roads in China. Accident Analysis and Prevention, 95,

438–447. https://doi.org/10.1016/j.aap.2015.07.001.

Harwood, D. W., Neuman, T. R., & Leisch, J. P. (2000).

Summary of design speed, operating speed, and design

consistency issues. Transportation Research Record:

Journal of Transportation Research Board, 1701, 116–

120. https://doi.org/http://dx.doi.org/10.3141/1701-15.

Hjelkrem, O. A., & Ryeng, E. O. (2016). Chosen risk level

during car-following in adverse weather conditions.

Car Drivers Do Not Choose Their Speed in Urban Environments: Speed Models in Tangent Streets

427

Accident Analysis and Prevention, 95, 227–235.

https://doi.org/10.1016/j.aap.2016.07.006.

INEC. (2010). Resultados del Censo 2010 de la población

y vivienda del Ecuador. Fascículo provincial Loja.

http://www.ecuadorencifras.gob.ec//wp-content/descar

gas/Manu-lateral/Resultados-provinciales/loja.pdf.

Jevtić, V., Vujanić, M., Lipovac, K., Jovanović, D., &

Pešić, D. (2015). The relationship between the

travelling speed and motorcycle styles in urban settings:

A case study in Belgrade. Accident Analysis and

Prevention, 75, 77–85. https://doi.org/10.1016/

j.aap.2014.11.011.

Keay, L., Munoz, B., Duncan, D. D., Hahn, D., Baldwin,

K., Turano, K. A., Munro, C. A., Bandeen-Roche, K.,

& West, S. K. (2013). Older drivers and rapid

deceleration events: Salisbury Eye Evaluation Driving

Study. Accident Analysis and Prevention, 58, 279–285.

https://doi.org/10.1016/j.aap.2012.06.002.

Montella, A., Imbriani, L. L., Marzano, V., & Mauriello, F.

(2015). Effects on speed and safety of point-to-point

speed enforcement systems: Evaluation on the urban

motorway A56 Tangenziale di Napoli. Accident

Analysis and Prevention, 75, 164–178.

https://doi.org/10.1016/j.aap.2014.11.022.

Pignataro, L. J. (1973). Traffic Engineering - Theory and

Practice. Prentice-Hall Publishing Co.

Poe, C., & Mason, J. (2000). Analyzing influence of

geometric design on operating speeds along low-speed

urban streets: mixed-model approach. Transportation

Research Record, 1737, 18–24. https://doi.org/

http://dx.doi.org/10.3141/1737-03.

R Core Team. (2013). R: A language and environment for

statistical computing. R Foundation for Statistical

Computing. http://www.r-project.org/

Rahman, A., & Lownes, N. E. (2012). Analysis of rainfall

impacts on platooned vehicle spacing and speed.

Transportation Research Part F: Traffic Psychology

and Behaviour, 15(4), 395–403. https://doi.org/

10.1016/j.trf.2012.03.004.

Ray, B., Ferguson, E., Knudsen, J., Porter, R., & Mason, J.

(2014). NCHRP Report 785: Performance-Based

Analysis of Geometric Design of Highways and

Streetsof Highways and Streets. https://doi.org/10.

17226/22285.

Roidl, E., Frehse, B., & Höger, R. (2014). Emotional states

of drivers and the impact on speed, acceleration and

traffic violations - A simulator study. Accident Analysis

and Prevention, 70, 282–292. https://doi.org/10.1016/j.

aap.2014.04.010.

Schagen, I. Van, Commandeur, J. J. F., Goldenbeld, C., &

Stipdonk, H. (2016). Monitoring speed before and

during a speed publicity campaign. Accident Analysis

and Prevention, 97, 326–334. https://doi.org/10.1016

/j.aap.2016.06.018.

Schechtman, E., Bar-Gera, H., & Musicant, O. (2016).

Driver views on speed and enforcement. Accident

Analysis and Prevention, 89, 9–21. https://doi.org/

10.1016/j.aap.2015.12.028.

Schmidt-Daffy, M. (2013). Fear and anxiety while driving:

Differential impact of task demands, speed and

motivation. Transportation Research Part F: Traffic

Psychology and Behaviour, 16, 14–28. https://

doi.org/10.1016/j.trf.2012.07.002.

Sekhar, C. R., Nataraju, J., Velmurugan, S., Kumar, P., &

Sitaramanjaneyulu, K. (2016). Free Flow Speed

Analysis of Two Lane Inter Urban Highways.

Transportation Research Procedia, 17(December

2014), 664–673. https://doi.org/10.1016/j.trpro.2016.

11.121.

Tarko, A., Hall, T., Romero, M., & Jiménez, C. G. L.

(2016). Evaluating the rollover propensity of trucks - A

roundabout example. Accident Analysis and

Prevention, 91, 127–134. https://doi.org/10.1016/j.aap.

2016.02.032.

Tarris, J., Poe, C., Mason, J., & Goulias, K. (1996).

Predicting operating speeds on low-speed urban streets:

Regression and panel analysis approaches.

Transportation Research Record, 1523, 46–54.

https://doi.org/http://dx.doi.org/10.3141/1523-06.

Taubman-Ben-Ari, O., Mikulincer, M., & Gillath, O.

(2004). The multidimensional driving style inventory—

scale construct and validation. Accident Analysis and

Prevention, 36(3), 323–332. https://doi.org/ https://doi.

org/10.1016/S0001-4575(03)00010-1.

Thompson, K. R., Johnson, A. M., Emerson, J. L., Dawson,

J. D., Boer, E. R., & Rizzo, M. (2012). Distracted

driving in elderly and middle-aged drivers. Accident

Analysis and Prevention, 45, 711–717. https://doi.org/

10.1016/j.aap.2011.09.040.

Wang, J. (2006). Operating speed models for low speed

urban environments based on in-vehicle GPS data (Phd

Thesis) (Issue May) [Georgia Institute of Technology].

https://smartech.gatech.edu/bitstream/handle/1853/105

08/wang_jun_200605_phd.pdf.

VEHITS 2021 - 7th International Conference on Vehicle Technology and Intelligent Transport Systems

428