19(2): S2: 468-478, 2024 www.thebioscan.

com

Modeling Time-to-Collision for Road Work Zones under Varying Geometric and
Traffic Conditions
1Mr.S.B.Pawar, 2Mr.O.S.Bidkar, 3Dr.Y.P.Pawar, 4Dr.S.S. Kadam, 5Dr.S.D.Jagdale, 6 Mr.S.R.Patil
1,2,3,4,5,6
Assistant Professor,
E-mails: 1 [email protected], 2 [email protected], [email protected], 4
[email protected], 5 [email protected], 6 [email protected]
Mobile No’s: +91-8329006248, +91-9561460476, +91-91455557631, +91-8600502360, +91-9730873511, +91-909083934;
1,2,3,4,5,6
Civil Engineering Department, SKNSCOE, Pandharpur, Korti

DOI: https://doi.org/10.63001/tbs.2024.v19.i02.S2.pp468-478

KEYWORDS ABSTRACT
Trajectory data,
Highway transportation is an essential mode of transportation from both accessibility and mobility perspectives. There is
Work zone,
more traffic demand for road transportation than for other modes. In addition, there is haphazard development of road
Time-to-collision,
construction in all states of India. Road construction and maintenance activities are more common due to catering to more
Modelling.
supplies of traffic demands. In particular, the work zone (WZ) is an area designed for road construction activities in which
a specific road section is blocked. Therefore, traffic safety issues are much more for various road conditions. Therefore,
Received on:
studying traffic safety at the microscopic level for varying roadway, geometric, and traffic conditions is necessary. For this,
01-08-2024 the present study adopts a surrogate safety measure, namely time-to-collision (TTC), to study traffic safety at a microscopic
level. The analysis involves the average value of TTC, deceleration rate to avoid a collision, and trajectories for geometric
Accepted on:
parameters of WZ. Modeling of TTC is performed for various geometric parameters of the WZ. The study outcome will
20-11-2024 benefit the construction authority in planning safety measures to minimize the probability of conflicts and correspondingly
achieve safety reasonably.

INTRODUCTION creating SSM. Ozbay et al. 2008 (3) developed and verified a
novel SSM based on simulation. They suggested a modified
Road Transportation is a crucial mode of transportation, having
simulation-based SSM, which better captures the chance of
more demands from overall states of Indian country. Road
crashes and the severity of these potential crashes than the
construction and maintenance activities are common in India.
previous one. These surrogate safety indices should be limited to
Road construction and maintenance activities create various
the analysis of linear conflicts, as they are first advised for link-
types of construction work zone (WZ). Traffic safety is the more
based research.
concerning issue in the construction WZ area. Therefore, it is
Morando et al. 2018 (4) examined autonomous vehicle’s (AV)
imperative to study traffic safety at a microscopic level. The
safety impact using simulation-based substitute safety measures.
researcher derives various surrogate safety measures, but their
The safety study was performed for two case studies-a signalized
actual behavior concerning change in the geometric parameters
intersection and a roundabout with various AV penetration rates.
of WZ is not considered. Hence, the present study modeled time-
The findings show that AV significantly improves safety with high
to-collision (TTC) for varying road geometric and traffic
penetration rates, especially at short headways, to boost road
conditions under WZ conditions. Gettman et al. 2003 (1) studied
capacity and reduce delay. With AV penetration rates between
various surrogate safety measures (SSM) from traffic simulation
50% and 100% for the signalized intersection, conflicts are
models such as TTC, post-encroachment time, deceleration rate,
reduced by 20% to 65%, statistically significant at a 95%
maximum speed, and speed differential. Measures of conflict
confidence range. With 100% AV penetration, the roundabout’s
intensity include the TTC, post-encroachment time, and
conflict rate drops by 29% to 64% (statistically significant at a
deceleration rate. To gauge the severity of prospective crashes,
95% confidence interval). SSM for a simulation-based traffic
use maximum speed and the speed differential. A post-
study was researched by Wang et al. in 2013 (5). To determine
processing tool is used to compute the statistics for the various
the crash risk of simulated conflicts, the distributions of
measurements and compare design alternatives after the
response time and maximum braking rates are added to a
simulation model has been run through several iterations.
probabilistic crash propensity model. This model can deliver the
Wang et al. 2019 (2) reviewed SSM and their applications in
ACPM for the three crash types crossing, rear end, and lane
connected and automated vehicle (CAV) safety modeling. Their
change. As part of an experimental validation effort, 12
main objectives included providing a thorough and organized
crossings were simulated using the simulation program VISSIM. To
assessment of critical SSM studies, highlighting challenges and
estimate the ACPM, valuable conflict data are retrieved utilizing
openings for future SSM and CAV research, and helping
the surrogate safety assessment approach. The ACPM can assess
researchers and practitioners select the best SSM for safety
the relative safety of various traffic facilities and treatments
investigations. Several significant challenges are also identified,
using Spearman rank tests. Table 1 summarizes the available
including SSM for CAV trajectory optimization, SSM for individual
simulation-based literature and their implication on safety.
vehicles and vehicle platoons, and CAV as a new data source for

468
 TABLE 1 Summary of simulation-based literature

Safety
Category Reference Research Features Research Outcomes Measures

Study on the effect of work-zone lane Integration software and safety implications are Deceleration
Zhu et al. (6) closures on traffic safety calculated regarding deceleration and speed
variance.
Development of a cellular automata (CA) CA model is the best-fitted model. A model shows Travel time
Meng et al. (7) model to study the WZ traffic the close agreement of travel time and traffic
delay for CA model data and field data.
Development of an improved cellular Results showed that the improved model could Deceleration
automaton model for work zone traffic accelerate, decelerate, keep their velocity, and
A Hou et al. (8)
simulation change lanes more realistically while passing the
work zone.
Investigating the effectiveness of VSL algorithms can yield a substantial increase in Delays, TTC
Lin et al. (9) variable speed limit controls on highway both WZ throughputs and a reduction in vehicle
WZ operations delays.
Development of a document that guides Traffic volume involves various case study TTC
the analyst, researcher or manager in examples, discussions, and analyses designed to
Hardy et al. (10)
performing a specific work zone analysis provide helpful information to the WZ analysis.
project
Study of rural interstate work zone Simulation for WZ Traffic Management Plan gives Acceleration
Schrock et al.
traffic management plans in iowa using a better alternative: implementing a nonstop
(11)
simulation work schedule until project completion.
Validation of driving simulator for WZ It is found that there is a statistical difference DRAC
Bella et al. (12) design. between field and simulated data.
Calibration and validation of freeway VISSIM Microsimulation results showed that the Acceleration
Park et al. (13) work zone using VISSIM microsimulation procedure effectively calibrated and validated a
freeway work zone network.
Multivariate regression for estimating The estimated statistical models can generate a Acceleration
Edara et al. (14) driving behaviour parameters in work range of parameter values that produce a wide
B zone simulation range of capacities used by state DOTs in the US.
Assessment of the impacts of highway WZ The results of this evaluation will be of interest Acceleration
Moriarty et al. strategies using simulation models to State and local transportation engineers
(15) responsible for planning and designing work zone
strategies.
Study on the various simulation models All models are calibrated and validated with field Acceleration
Maze et al. (16) data and simulation data.
Testing of WZ traffic control devices There are significant changes in state and Acceleration
Nelson et al. (17) using driving simulations national standards.
Developing an improved cellular The ICA model is used to estimate capacity in the Deceleration
Meng et al. (18) automata model for simulating WZ.
heterogeneous traffic in work zone
Estimation of arterial WZ capacity using Simulation of arterial WZ showed that the Capacity
Heaslip et al. (19) simulation distance of WZ to the downstream intersection
affects the capacity of WZ.
C
Replication of WZ capacity values in the The study provides an appropriate method of Capacity
Chatterjee et al.
simulation model choosing the lane-changing and car-following
(20)
parameters.
Estimation of freeway WZ capacity by Capacity calculated from analytical models of Acceleration
Heaslip et al. (21) using simulation and field data CORSIM (version 5.1) is within 1% of the capacity
of HCM 2000 and field data.
Note: A = Simulation using various modeling techniques, B = Hence, it is necessary to study the descriptive statistics of SSM
Simulation using various simulation software, and C = Capacity along with various road parameters which affect the modeling.
estimation using simulation technique. One of the research gaps is no literature available in the study
where the modeling of the SSM has taken place. SSM modeling is
The previous literature shows that traffic safety affects important because it gives insights into various parameters that
macroscopic and microscopic traffic flow parameters. affect SSM. Hence, this motivation study uses different study
Microscopic parameters, namely various derived SSM, are used to parameters to model the SSM.
capture the behavior of crashes and conflicts on various roads.

469
 SSM Methodology
FIGURE 1 Stepwise methodology of the study

Figure 1 shows various tasks of the present study in the form of 4. Data were imported into SSAM software for each simulation
research methodology and involved the following seven steps. . policy. Once data was extracted, MATLAB coding was
1. A detailed literature review was performed related to the applied to calculate leader-follower pairs. After the
SSM and modeling. calculation of leader-follower pairs, the TTC value is
2. The required data were collected in Western Expressway of calculated for each pair.
India for WZ and Without (WWZ) sections using traffic data 5. Analysis of Surrogate Safety Measures involves the
extractor software. MATLAB coding was used to calculate Histogram of SSMs, Heat map of SSMs and Descriptive
leader-follower pairs, which were calculated by considering Statistics of SSMs.
three conditions: (1) the leader should be ahead of their 6. The generalized Linear Model is used to model TTC for
follower, (2) the lateral distance should be less than the various roadways and traffic conditions. The model is
lateral threshold, and (3) the distance between the leader developed for multiple lane drops, tapers and lengths of
and follower should be minimum. Work-Zones.
3. Once the leader-follower pairs were drawn, TTC was 7. Finally, various results and conclusions are made.
calculated for each pair. Various simulation runs are 1 ANALYSIS OF SURROGATE SAFETY MEASURES
developed in VISSIM software with varying lane drops, 1.1 SSM Histogram
tapers and lengths of WZ. Calibration and validation were Histograms of SSM are plotted for lane drops, WZ length, and
done using the Widemann 74 Models. Macroscopic and taper rate for various traffic volumes like low, medium and high
microscopic validations were then performed. Macroscopic traffic volume, respectively, as shown in Figure 2. It is shown
validation involved speed versus flow plots for field data that the variation of TTC decreases as an increase in lane drops.
and normal data. Plots were made, and it was found that Similarly, there is more variation of TTC with an increase in the
simulated data was matched with field data. MAPE values WZ length and taper rate, respectively. Further, this variation
are also less than 15%. Microscopic validation also involved also varies for the varying conditions for the same traffic
relative velocity and spacing histograms. The results volume. This shows that the WZ conditions significantly influence
showed that simulated data significantly matched field data traffic safety and operations. Additionally, for all three
with 95% confidence interval. Various lane drops, tapers conditions, the TTC is normally distributed.
and WZ lengths were varied, considering the guidelines of
IRC SP:55 2014.
Histo gram o f TTC fo r vario us lane-dro ps o f Lo w Vo lume Histo gram o f TTC fo r vario us lane-dro ps o f Medium Vo lume Histo gram o f TTC fo r vario us lane-dro ps o f High Vo lume
Variable 0.6 Variable
0.25 Variable 0.4 3-LDM 3-LDH
3-LDL 2-LDM
0.5 2-LDH
2-LDL 1-LDM 1-LDH
0.20 1-LDL
0.3 0.4
Density
De ns ity

0.15
De ns ity

0.3
0.2
0.10
0.2
0.1
0.05 0.1

0.00 0.0 0.0

0.0 1.5 3.0 4.5 6.0 7.5 9.0 0 1 2 3 4 5 6 -0.0 0.6 1.2 1.8 2.4 3.0 3.6
Data Data Data

(a) 1-Lane-drop (b) 2-Lane-drop (c) 3-Lane-drop

470
 His togram of TTC with varaition of le ngth of Work-Zone for Low Volume His togram of TTC with varaition of le ngth of Work-Zone for M e dium Volume His togram of TTC for varition of Le ngth of Work-Zone for High Volume
0.5 0.5 Variable
Variable
100 M 0.5
100 L Variable
150 L 150 M 100 H
0.4 0.4
200 L 200 M 0.4 150 H
200 H
0.3 0.3

De ns ity
De ns ity

De ns ity
0.3

0.2 0.2
0.2

0.1 0.1 0.1

0.0 0.0 0.0

0 1 2 3 4 5 6 -0.0 0.8 1.6 2.4 3.2 4.0 4.8 5.6 0.00 0.75 1.50 2.25 3.00 3.75 4.50
TTC (Se c) TTC (s e c) TTC (Se c)

(d) WZ length = 100 m

His togram of TTC with variation of tape r for Low Volume
(e) WZ length = 150 m (f) WZ length = 200 m
His togram of TTC with variation of tape r of Work-Zone for High Volume
His togram of TTC with varaition of tape r M e dium Volume
0.4 0.9 0.9
Variable Variable
200 H 1:03 M Variable
0.8 0.8
200 M 1:12 M 1:03 H
200 L 0.7 1:15 M 0.7 1:12 H
0.3
1:15 H
0.6 0.6
De ns ity

De ns ity

De ns ity
0.5 0.5
0.2
0.4 0.4
0.3 0.3
0.1 0.2 0.2
0.1 0.1
0.0 0.0 0.0
0 1 2 3 4 5 6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 -0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8
TTC (s e c) TTC (Se c) TTC (Se c)

(g) Taper rate = 1:03 (h) taper rate = 1:12 (i) taper rate = 1:15

Low volume Medium volume High volume

FIGURE 2 TTC distribution for various WZ conditions

1.2 Heat Map of TTC-Based Conflicts criteria for identification of conflicts. Heat maps are plotted for
The number of conflicts is calculated based upon the TTC as various lane drops, tapers and WZ lengths, as shown in Figure 3.

(a) Lane drop

471
 (b) WZ Length

( c) Rate of Taper
FIGURE 3 Spatial distribution of TTC-based conflicts in the WZ area

It is found that the number of conflicts increases with an length. It is also found that the number of conflicts decreases
increase in lane drops and decreases with an increase in the WZ with an increase in the taper rate of WZ.

472
2.3 Trajectory Analysis

(a) Lane-dropwise Trajectories

(b) WZ Length wise trajectories

(c) Taper-wise trajectories

FIGURE 4 Vehicular trajectories for various WZ conditions

473
The analysis section also involves trajectory data at certain road vehicle type, volume type, longitudinal velocity, lateral velocity
sections for various lane drops, WZ lengths and taper rates. At a and length of WZ, tapering length, etc. Table 5 shows the
microscopic level, traffic behavior is studied. Lane drop, WZ description and type of variables used for modeling.
length, and taper rate-wise vehicular trajectories are plotted at 2.4 Box and Whisker Plots of Acceleration
a microscopic level, as shown in Figure 4. Similarly, for modeling Box and whisker plots of acceleration are plotted concerning
TTC concerning geometric parameters such as lane drops, WZ lane drops, tapers and length of work zones. Figure 5 shows the
length, and taper rate, various types of variables were Box and whisker plots for acceleration for various lane drops,
considered, such as the transverse angle between vehicles, tapers and lengths of work-zone.

(H=High Volume, M=Medium Volume, L=Low Volume)

FIGURE 5 Box and Whisker plots of Acceleration

474
 It is found from the figure that the acceleration increases with (FT). The corresponding mean value for various traffic flow
the increase in the length and the rate of tapers of the work conditions for the TTC, DRAC and FT with their standard
zone. deviation in the parenthesis is shown in Table 2. The results
TTC MODELLING UNDER VARYING GEOMETRIC show that all three SSM values vary significantly for all three
CONDITIONS traffic states with varying WZ conditions.
The analysis section involves the average value of TTC,
Deceleration-Rate to Avoid Collision (DRAC), and Following Time

TABLE 2 The average value of TTC, DRAC, and FT concerning geometric parameters

Section Type
SSM Traffic State
1-Lane drop 2-Lane drop 3-Lane drop 100 m 150 m 200 m 01:03 01:12 01:15
High (sec) 3.12 (1.12) 2.12 (0.96) 1.06 (1.24) 1.21 (1.03) 2.03 (2.01) 3.04 (1.85) 1.15 (1.04) 2.16 (0.96) 3.01 (2.01)
TTC Medium (sec) 4.26 (1.08) 3.24 (2.04) 2.03 (1.64) 1.45 (1.21) 2.15 (1.54) 3.24 (2.01) 1.34 (0.96) 2.64 (1.21) 3.64 (1.13)
Low (sec) 5.42 (2.04) 3.89 (1.64) 2.85 (1.96) 1.64 (1.45) 2.48 (2.04) 3.75 (1.64) 1.74 (1.04) 3.02 (1.23) 4.04 (1.07)
High (sec) 3.22 (1.11) 4.23 (2.04) 5.23 (0.96) 3.44 (0.86) 4.64 (1.07) 6.24 (0.89) 2.85 (1.28) 4.08 (1.39) 5.22 (1.24)
DRAC Medium (sec) 2.64 (1.08) 3.46 (1.64) 4.16 (1.25) 2.85 (1.16) 4.01 (1.16) 5.87 (1.37) 2.61 (2.14) 4.21 (1.06) 4.28 (1.56)
Low (sec) 2.01 (1.34) 2.65 (1.32) 3.27 (1.34) 2.07 (1.28) 2.84 (1.63) 5.21 (0.93) 2.22 (1.67) 4.84 (0.93) 5.08 (0.97)
High (sec) 4.6 (1.02) 3.9 (1.05) 3.4 (0.92) 2.2 (0.79) 3.2 (1.01) 4.6 (0.81) 1.8 (1.01) 2.2 (1.08) 3.1 (0.86)
FT Medium (sec) 5.4 (0.86) 4.6 (0.76) 4.1 (1.03) 2.9 (0.87) 3.8 (0.95) 5.1 (0.79) 2.1 (0.89) 2.7 (0.76) 3.8 (1.04)
Low (sec) 6.2 (0.91) 5.2 (0.89) 4.7 (0.87) 3.5 (1.06) 4.5 (0.61) 5.8 (1.05) 2.4 (0.95) 3.2 (0.92) 4.2 (0.84)

TABLE 3 Description and type of variables

Sr. No Variables Description Variable Type Measurement (Min-Max)

1 Transverse angle Continuous 5 degrees to 45 degree
2 Vehicle type (VehT) Discrete 1-2W, 2-3W, 3-CAR, 4–LCV, 5-HCV,6-Bus
1-Low Traffic Volume, 2-Medium Traffic Volume, 3-High
3 Volume type (VolT) Discrete
Traffic Volume
4 Longitudinal velocity (LongV) Continuous 0.8 m/sec -15 m/sec
5 Lateral velocity (LatV) Continuous 0.2 m/sec – 1.4 m/sec
6 WZ length (Length) Continuous 100 m-200 m
7 Tapering length (TL) Continuous 30 m- 75 m
8 Lane drops (LD) Continuous 1-Lane drop, 2-Lane drop, 3-Lane drop

2.1 Generalized Linear Model functions f (or probability mass function, in the case of a
A generalized linear model (GLM) is a versatile generalization of discrete distribution), can be expressed by equation 1, are
traditional statistical linear regression. By allowing the linear included in the over-dispersed exponential family of
model to be connected to the response variable via a link distributions.
function and by allowing the size of each measurement's
variance to be a function of its predicted value, the GLM
generalizes linear regression. John Nelder and Robert
Wedderburn developed generalized linear models to combine
many different statistical models, such as linear regression,
logistic regression, and Poisson regression. They suggested an ……………….(1)
iteratively reweighted least squares method for maximum where θ and τ are the parameters for the exponential
likelihood estimation (MLE) of the model parameters. MLE is still distribution model and y is an independent variable.
widely used and is often used as the default method in statistical
computer programs. Other methods, such as least squares fitting The quantity incorporating knowledge of the independent
to variance-stabilized answers and Bayesian regression, have variables into the model is called the linear predictor. A linear
been developed. The families of probability distributions predictor is represented by the symbol (Greek "eta"). The link
parameterized by the bold symbols theta and tau, whose density function connects it to the expected value of the data. The
functions f (or probability mass function, in the case of a expression for is given as linear combinations of unknown
discrete distribution), can be expressed in the form, are parameters (thus, "linear"). The matrix of independent variables
included in the over-dispersed exponential family of X is used to represent the linear combination coefficients.
distributions. The families of probability distributions Consequently, it can be written as
parameterized by the bold symbols theta and tau, whose density

…………………(2)
where β is a constant and X is a matrix of the independent variables.

475
 There are five essential assumptions for which the generalized 2.2 Modeling of TTC for Varying Lane-drop
linear model will work: linear relationship, multivariate The generalized linear modeling technique is used for modeling
normality, no or little multicollinearity, no auto-correlation, and various lane drops based on traffic conflicts. The model summary
homoscedasticity. and significance of different traffic state parameters with their
statistical value are shown in Table 4.

TABLE 4. Model summary for various lane drop conditions

Model Variables
Model Types Model Statistics
Intercept Transverse angle VehT VolT LongV LatV Length TL

Coefficient 2.329 0.026 0.270 -0.517 -0.006 0.005 0.004 0.121

Standard Error 0.234 0.002 0.029 0.00 0.002 0.008 0.001 0.048
1-Lane drop Wald Chi-Square 98.795 223.009 85.420 164.620 7.214 0.472 9.958 6.307
p-value 0.000 0.000 0.000 0.000 0.007 0.492 0.002 0.012
Goodness-of-fit Measures AIC = 1044.28, BIC = 1085.92, LL = -456.25, MAPE = 14.2
Coefficient -0.933 0.037 0.980 -1.338 -0.054 0.201 0.010 0.196
Standard Error 5.971 0.017 0.00 0.029 0.041 0.028 0.020 0.040
2-Lane drop Wald Chi-Square 0.024 4.920 52.187 20.985 20.985 0.779 4.905 0.070
p-value 0.876 0.027 0.000 0.000 0.000 0.378 0.027 0.791
Goodness-of-fit Measures AIC = 387.56, BIC = 452.25, LL = -1893.50, MAPE = 15.3
Coefficient 6.163 0.054 1.021 -1.891 -0.294 0.402 0.021 0.301
Standard Error 1.004 0.005 0.017 0.029 0.050 0.008 0.004 0.018
3-Lane drop Wald Chi-Square 37.683 102.329 38.979 34.139 34.322 34.322 4.229 3.589
p-value 0.000 0.000 0.000 0.000 0.000 0.000 0.040 0.058
Goodness-of-fit Measures AIC = 1678.25, BIC = 1713.24, LL = -830.45, MAPE = 15.3

2.3 Modeling of TTC for Varying WZ Lengths as a function of various traffic state variables. Table 5 shows the
WZ length effect on TTC has been performed. For that, 100 m, model summary of TTC for varying WZ lengths.
150 m and 200 m WZ lengths are considered, and TTC is modeled

TABLE 5 Model summary for various WZ length conditions

Model Variables
Model Types Model Statistics
Intercept Transverse angle VehT VolT LongV LatV LD Scale

Coefficient 6.821 0.028 0.487 -0.526 -0.296 0.443 -0.896 13.926a

Standard Error 1.043 0.006 0.210 0.375 0.076 0.077 0.426 0.880
100 m WZ
Wald Chi-Square 42.742 15.958 5.362 16.501 14.782 32.722 5.927 -
length
p-value 0.000 0.000 0.021 0.000 0.000 0.000 0.015 -
Goodness-of-fit Measures AIC = 1044.28, BIC = 1085.92, LL = -456.25, MAPE = 14.2
Coefficient 3.813 0.036 0.775 -0.775 -0.425 0.627 -1.140 5.309a
Standard Error 0.629 0.004 0.089 0.124 0.039 0.044 0.134 0.335
150 m WZ
Wald Chi-Square 36.685 7.914 19.116 4.921 7.493 8.288 115.584 -
length
p-value 0.000 0.005 0.000 0.027 0.006 0.004 0.000 -
Goodness-of-fit Measures AIC = 1044.28, BIC = 1085.92, LL = -456.25, MAPE = 14.2
Coefficient 3.813 0.048 0.845 -0.945 -0.815 0.827 -2.015 5.309a
Standard Error 0.829 0.006 0.089 0.424 0.059 0.044 0.134 0.335
200 m WZ
Wald Chi-Square 8.685 9.914 6.116 5.021 8.493 8.288 115.584 -
length
p-value 0.000 0.005 0.000 0.037 0.008 0.004 0.000 -
Goodness-of-fit Measures AIC = 1044.28, BIC = 1085.92, LL = -456.25, MAPE = 14.2

476
 2.4 Modeling of TTC for Varying Taper Rates Taper rate-wise TTC was modeled for various taper rates such as
1:03, 1:12 and 1:15. Table 6 shows the model summary and their
significance.

TABLE 6 Model summary for various taper rate conditions

Model Variables
Model Types Model Statistics
Intercept Transverse angle VehT VolT LongV LatV LD Scale

Coefficient 8.215 1.034 1.201 -0.526 -0.496 0.528 -0.985 13.926a

Standard Error 1.043 0.006 0.210 0.375 0.076 0.077 0.426 0.880
1:03 Taper
Wald Chi-Square 42.742 15.958 5.362 16.501 14.782 32.722 5.927 -
rate
p-value 0.000 0.000 0.021 0.000 0.000 0.000 0.015 -
Goodness-of-fit Measures AIC = 1044.28, BIC = 1085.92, LL = -456.25, MAPE = 14.2
Coefficient 3.813 0.052 1.852 -0.826 -0.625 0.726 -1.314 5.309a
Standard Error 0.629 0.004 0.089 0.124 0.039 0.044 0.134 0.335
1:12 Taper
Wald Chi-Square 36.685 7.914 19.116 4.921 7.493 8.288 115.584 -
rate
p-value 0.000 0.005 0.000 0.027 0.006 0.004 0.000 -
Goodness-of-fit Measures AIC = 387.56, BIC = 452.25, LL = -1893.50, MAPE = 15.3
Coefficient 3.813 0.085 2.155 -1.124 -0.864 0.863 -2.615 5.309a
Standard Error 0.829 0.005 0.093 0.524 0.049 0.064 0.246 0.335
1:15 Taper
Wald Chi-Square 10.685 9.914 10.116 5.921 9.493 9.288 10.584
rate
p-value 0.000 0.004 0.000 0.057 0.008 0.006 0.000
Goodness-of-fit Measures AIC = 1678.25, BIC = 1713.24, LL = -830.45, MAPE = 15.3

2.5 Results and Discussion CONCLUSION

It is found that TTC of the WZ increases with an increase in the
transverse angle of the WZ, increases with an increase in the Traffic safety is an essential issue on most Indian roads. Many
size of follower vehicles, decreases with an increase in the traffic crashes and conflicts occur on Indian roads. Traffic safety
volume type, decreases with an increase in the longitudinal will affect more at the microscopic traffic levels for varying
velocity, increases with increase in the lateral velocity, roadway and traffic conditions. SSM is the best platform for
increases with increase in the WZ length and increases with microscopic analysis of traffic safety. With this motivation,
increase in the tapering length. It is also observed that various analysis modeling of SSM is done in this study. Based on this
independent variables are more sensitive to TTC with a decrease study, the following conclusions are drawn:
in the lane drops of the WZ. The main reason for increasing the 1. Surrogate safety measures such as TTC have a more
sensitivity of independent variables is the more maneuverability significant effect on the variation in roadway geometry
of vehicles while moving through the WZ as the movement of of the work zone.
vehicles increases with an increase in the lane drops. Hence TTC 2. TTC can be modeled as a function of various
of the vehicles becomes sensitive to various parameters such as microscopic traffic variables of the WZ. The results
transverse angle, vehicle type, volume type, the longitudinal show that TTC becomes more sensitive with the
velocity of follower type, lateral velocity, WZ length and increase in lane drops, tapers, and lengths of the work
tapering length. It is observed that TTC increases with an zones.
increase in the angle between the vehicles, increases with an 3. TTC becomes more sensitive with a decrease in the
increase in vehicle size, decreases with an increase in traffic number of lane drops, an increase in the length of the
volume, decreases with an increase in longitudinal velocity, work zone and an increase in the rates of tapers.
increases with an increase in lateral velocity, decreases with an 4. The results of this study will help control TTC in the
increase in lane drop. The WZ lengths increase, then TTC field by varying various independent parameters of the
becomes more sensitive to the various independent parameters model to avoid conflicts and crashes. The future scope
of WZ. As the WZ lengths increase, vehicles are more of the study involves understanding the parameters
maneuverable concerning space and time. Hence, TTC will affecting the TTC.
become more sensitive to various independent parameters of 5. Future studies should involve the factors critical for
WZ. As the WZ lengths decrease, TTC will become less sensitive TTC reduction and ways to improve them to achieve
to the different independent parameters of the WZ. The model higher TTC values between leader and follower pairs.
results show that TTC increases with an increase in the angle, Once a higher value of TTC is achieved between the
increases with vehicle size, decreases with an increase in traffic leader and follower pairs, there will be fewer chances
volume, decreases with an increase in longitudinal velocity, of conflicts and crashes.
increases with an increase in lateral velocity, and decreases with REFERENCES
an increase in lane drop. It is also found that as the taper rates • Gettman, D. and Head, L., 2003. Surrogate safety measures
increase, TTC becomes more sensitive to various independent
parameters. When the taper rates increase, vehicles have more from traffic simulation models. Transportation Research
maneuverability concerning time and space. Hence, TTC Record, 1840(1), pp.104-115.
becomes more sensitive to various independent parameters of • Wang, C., Xie, Y., Huang, H. and Liu, P., 2021. A review of
WZ. As the taper rates decrease, TTC becomes less sensitive to surrogate safety measures and their applications in connected
various independent parameters of WZ. and automated vehicles safety Modelling. Accident Analysis &
Prevention, 157, p.106157.

477
• Ozbay, K., Yang, H., Bartin, B. and Mudigonda, S., 2008.
Derivation and validation of new simulation-based surrogate
safety measure. Transportation research record, 2083(1),
pp.105-113.
• Morando, M.M., Tian, Q., Truong, L.T. and Vu, H.L., 2018.
Studying the safety impact of autonomous vehicles using
simulation-based surrogate safety measures. Journal of
advanced transportation, 2018.
• Wang, C. and Stamatiadis, N., 2013. Surrogate safety measure
for simulation-based conflict study. Transportation research
record, 2386(1), pp.72-80.
• Zhu, J. and Saccomanno, F.F., 2004. Safety implications of
freeway work zone lane closures. Transportation research
record, 1877(1), pp.53-61.
• Meng, Q. and Weng, J., 2010. Cellular automata model for
work zone traffic. Transportation Research Record, 2188(1),
pp.131-139.
• Hou, G. and Chen, S., 2019. An improved cellular automaton
model for work zone traffic simulation considering realistic
driving behavior. Journal of the Physical Society of
Japan, 88(8), p.084001.
• Lin, P.W., Kang, K.P. and Chang, G.L., 2004, July. Exploring
the effectiveness of variable speed limit controls on highway
work zone operations. In Intelligent transportation
systems (Vol. 8, No. 3, pp. 155-168). Taylor & Francis Group.
• Hardy, M. and Wunderlich, K., 2009. Traffic analysis tools
volume IX: work zone modeling and simulation: a guide for
analysts.
• Schrock, S.D. and Maze, T.H., 2000. Evaluation of rural
interstate work zone traffic management plans in Iowa using
simulation. In Mid-Continent Transportation Symposium 2000
Proceedings.
• Bella, F., 2005. Validation of a driving simulator for work zone
design. Transportation Research Record, 1937(1), pp.136-144.
• Park, B. and Qi, H., 2006, September. Microscopic simulation
model calibration and validation for freeway work zone
network-a case study of VISSIM. In 2006 IEEE Intelligent
Transportation Systems Conference (pp. 1471-1476). IEEE.
• Edara, P. and Chatterjee, I., 2010. Multivariate regression for
estimating driving behavior parameters in work zone
simulation to replicate field capacities. Transportation
Letters, 2(3), pp.175-186.
• Moriarty, K.D., Collura, J., Knodler, M., Ni, D., Heaslip, K. and
Hall, M., 2008. Using simulation models to assess the impacts
of highway work zone strategies; case studies along interstate
highways in Massachusetts and Rhode Island. In TRB 2008
Annual Meeting.
• Maze, T. and Kamyab, A., 1999. Work zone simulation model.
• Nelson, A.A., Chrysler, S.T., Finley, M.D. and Ullman, B.R.,
2011. Using driving simulation to test work zone traffic control
devices (No. 11-1515).
• Meng, Q. and Weng, J., 2011. An improved cellular automata
model for heterogeneous work zone traffic. Transportation
research part C: emerging technologies, 19(6), pp.1263-1275.
• Heaslip, K., Jain, M. and Elefteriadou, L., 2011. Estimation of
arterial work zone capacity using simulation. Transportation
Letters, 3(2), pp.123-134.
• Chatterjee, I., Edara, P., Menneni, S. and Sun, C., 2009.
Replication of work zone capacity values in a simulation
model. Transportation research record, 2130(1), pp.138-148.
• Heaslip, K., Kondyli, A., Arguea, D., Elefteriadou, L. and
Sullivan, F., 2009. Estimation of freeway work zone capacity
through simulation and field data. Transportation research
record, 2130(1), pp.16-24.

478