Final
report of ITS Center project: Rubbernecking impact of incidents
A Research Project Report
For the
National ITS Implementation Research Center
A U.S. DOT University Transportation
Center
AN
ANALYSIS ON THE IMPACT OF RUBBERNECKING ON URBAN FREEWAY TRAFFIC
Principal Investigators:
Dr. Hualiang (Harry) Teng
An Analysis on the Impact
of Rubbernecking on Urban Freeway Traffic
By:
Jonathan P. Masinick
Dr. Hualiang (Harry)
Teng
A Research Project Report for the ITS Implementation
Center
Jonathan P. Masinick
University of Virginia
Dr. Hualiang (Harry) Teng
Department of Civil Engineering
Email: hht4n@virginia.edu
Center for Transportation Studies at the University
of Virginia produces outstanding transportation professionals, innovative research
results and provides important public service. The Center for Transportation Studies
is committed to academic excellence, multi-disciplinary research and to developing
state-of-the-art facilities. Through a partnership with the Virginia Department
of Transportation’s (VDOT) Research Council (VTRC), CTS faculty hold joint appointments,
VTRC research scientists teach specialized courses, and graduate student work
is supported through a Graduate Research Assistantship Program. CTS receives substantial
financial support from two federal University Transportation Center Grants: the
Mid-Atlantic Universities Transportation Center (MAUTC), and through the National
ITS Implementation Research Center (ITS Center). Other related research activities
of the faculty include funding through FHWA, NSF, US Department of Transportation,
VDOT, other governmental agencies and private companies.
Disclaimer: The contents of
this report reflect the views of the authors, who are responsible for the facts
and the accuracy of the information presented herein. This document is disseminated under the sponsorship of the Department
of Transportation, University Transportation Centers Program, in the interest
of information exchange. The U.S. Government
assumes no liability for the contents or use thereof.
An Analysis on the Impact of Rubbernecking
on Urban Freeway Traffic
________________________________________________________________________
A
Thesis
Presented to
The Faculty of the
School of Engineering
and Applied Science
University of Virginia
________________________________________________________________________
In Partial Fulfillment
of the Requirements
for the Degree
Master of Science
in Civil Engineering
by
Jonathan P. Masinick
Approval Sheet
This
thesis is submitted in partial fulfillment of the requirements for the degree
of
Master of Science in Civil Engineering:
__________________________________________
Jonathan
P. Masinick
This
thesis has been read and approved by the examining committeeCommittee:
__________________________________________
Dr. Brian
Brain Park (committee
memeber)
__________________________________________
Dr. Brian
L. Smith (committee memberChair)
__________________________________________
Dr. Hualiang
(Harry
HarryTeng)
Teng (advisor)
Accepted
for the School of Engineering and Applied Science:
__________________________________________
Dean, School
of Engineering and Applied Science
August 2004
An
incident influences traffic not only in the incident direction but also in the
opposite direction. There has been research
on the influence of incidents on the traffic in the incident direction. However, the research
relating toabout
the influence on the opposite direction of traffic is rare. Previous research has shown that congestion
due to incidents account for 60% of the total congestion on a freeway system. These incidents
cause the freeway system to operate inefficiently. By determining which variables contribute to the “non-recurrent”
congestion and also the impact on traffic, mitigation techniques may
be applied to minimize these effects.
In this study, the impact of incidents on the traffic
in the opposite direction was investigated with focus on rubbernecking likelihood,
occupancydelay, capacityand capacity
reductionand
delay. To achieve this study certain
objectives were met. First, Aa database
consisting of incident information, traffic and other related variables was developed.
This informational
database was used to identify whether an incident had an impact on traffic in
the opposite direction that can be reflected in occupancy changes and capacity
reduction. This project
also investigates which incidents cause such impacts and under what circumstances
such impacts would occur. In addition
to the investigation on such impacts, the delay caused by the reduction in capacity
was also be studied. The next step was
to determine whether the rubbernecking impact on the opposite direction traffic
was significant. Factors that influence
the impacts of rubbernecking likelihood were identified. Recommendations of effective countermeasures were developed
to possibly reduce rubbernecking impacts. Traffic data was investigated while congestion delay as well as
capacity reduction calculations were performed. This study is the first attempt to evaluate the rubbernecking
impact of accidents on traffic in the opposite direction based on archived traffic
and accident data.
There
are many individuals who I would like to thank for helping me along with this
journeystudy
of a project. First, I would
like to thank my thesis advisor, Dr. Hualiang (Harry) Teng. The advice and support you have given me throughout
this project study is immeasurable.
Thank you for being for guiding, encouraging, and challenging me during
this process.
I would also like to thank Dr. Brian Smith for allowing me the opportunity
to use the SMART Travel Lab to complete this studyproject. Your advice in and outside of the classroom
has been a great help.
This project would never have gotten off the
ground if it were not for the amazing help from the Smart Travel Lab. I would like to thank Tim Ngov for the support
with all aspects of the lab. I would like
to thank Simona Babiceanu and Xiaoning Lu for their effort in helping me with
the data extraction process. I would also
like to thank Jun Yao for his help through this process.
Finally, I would like to thank my family for their support not only during
this studyproject,
but also my entire life. I would not be
where I am today if it were not for them. I
would like to thank my wife, Katie, for being the most patient, yet and
persistent person during this studyproject. I would never have
finished this project
if it were not for you.
Abstract iii
Acknowledgements. iv
Table of Contents. v
List of Figures. vivivii
List of Tables. viiviiviii
Chapter 1: Introduction. 1
Chapter 2: Background
Review.. 4
2.1 Flow / Occupancy (Density) Relationships. 4
2.2 Traffic Delay. 6
2.3 Capacity Reduction. 10
2.4 Rubbernecking Effects. 11
2.5 Physical Factors. 12
2.6 Summary. 14
Chapter 3: Methodology. 15
3.1 Data Source. 16
3.3 Traffic Data. 202019
3.4 Determination of Incident
Location and Significance of 212120
Rubbernecking Impacts. 212120
3.5 Congestion Delay Calculations.
252523
3.6 Capacity Reduction. 292928
3.7 Binary Logit Model 313130
3.8 Linear Regression. 313130
Chapter 4: Results and
Analysis. 333332
4.1 Incident Data. 333332
4.2 Significant Impacts. 333332
4.3 Rubbernecking Delay
Calculations.
383836
4.3 Capacity Reduction
Modeling. 434342
Chapter 5: Conclusions. 474745
5.1 Conclusions. 474745
5.2 Future Research. 484846
References. 515148
Approval Sheet.....................................................................................................................................................
ii
Abstract.....................................................................................................................................................
iii
Acknowledgements.....................................................................................................................................................
iv
Table of Contents.....................................................................................................................................................
v
List of Figures.....................................................................................................................................................
vi
List of Tables.....................................................................................................................................................
vii
List of Tables.....................................................................................................................................................
vii
Chapter 1: Introduction.....................................................................................................................................................
1
Chapter 2: Background
Review.....................................................................................................................................................
4
2.1 Incident Detection..........................................................................................................................................................................................
4
2.2 Flow / Occupancy (Density) Relationships..........................................................................................................................................................................................
5
2.3 Traffic Delay..........................................................................................................................................................................................
7
2.5 Capacity Reduction..........................................................................................................................................................................................
11
2.6 Rubbernecking Effects..........................................................................................................................................................................................
12
2.7 Physical Factors..........................................................................................................................................................................................
13
2.8 Summary..........................................................................................................................................................................................
15
Chapter 3: Methodology.....................................................................................................................................................
16
3.1 Data Source..........................................................................................................................................................................................
17
3.3 Traffic Data..........................................................................................................................................................................................
20
3.4 Determination of Incident Location and Significance
of..........................................................................................................................................................................................
21
Rubbernecking Impacts..........................................................................................................................................................................................
21
3.5 Congestion Delay Calculations..........................................................................................................................................................................................
24
3.6 Capacity Reduction..........................................................................................................................................................................................
29
3.7 Binary Logit Model..........................................................................................................................................................................................
30
3.8 Linear Regression..........................................................................................................................................................................................
31
Chapter 4: Results and Analysis.....................................................................................................................................................
33
4.1 Incident Data..........................................................................................................................................................................................
33
4.2 Significant Impacts..........................................................................................................................................................................................
33
Table 4.xx Results of a Binary Logit Model........................................................................................................................................................................................................................................................
36
Table 4xxx........................................................................................................................................................................................................................................................
36
4.3 Delay Calculations..........................................................................................................................................................................................
37
4.3 Capacity Reduction Modeling..........................................................................................................................................................................................
43
Chapter 5: Conclusions.....................................................................................................................................................
47
5.1 Conclusions..........................................................................................................................................................................................
47
5.2 Future Research..........................................................................................................................................................................................
48
References.....................................................................................................................................................
50
Appendix A: Project SQL+
Code.....................................................................................................................................................
51
Figure (2-1): Greenshield's
Flow vs. Density Model 5
Figure (2-2): Two-Regime Flow vs. Occupancy Model 6
Figure (2-3): Cumulative
Volume Diagram - Delay due to an Incident 7
Figure (2-5): Barrier Guardrail
System on a Section of Roadway on I-64. 13
Figure (2-5): A Standard
Concrete Barrier on I-264. 13
Figure (2-6): A Double
Stacked Concrete Barrier on I-64. 14
Figure (3-1): Area Map
of Hampton Roads Freeway System.. 16
Figure (3-2): The hr.incident
Table Available from the University of Virginia STL. 191918
Figure (3-3): Example of Incident Occupancy. 222221
Figure (3-4): Incident-caused
Occupancy at Multiple Stations. 242422
Figure (3-5): Example of
Rectangular Area Approach. 262624
Figure (3-6): Example of
Cumulative Arrival and Departure Curves. 272726
Figure (3-7): Application
of Integral Estimation on Delay Calculation. 282827
Figure (3-8): Example of
Capacity Reduction Calculation. 303029
Figure (4-1): Histogram
of the Frequency of Delay. 393937
Figure (4-2): Histogram of the Frequency of the Natural Log
of Delay.
404038
Figure (4-4): Histogram
of Capacity Reduction Percentage. 444442
Figure (4-5): Histogram
of Ln(Capacity Reduction) due to Rubbernecking. 454543
Figure 1: (2-1) Greenshield's Flow vs. Density Model.. 6
Figure 2: (2-2) Two-Regime Flow vs. Occupancy Model.. 7
Figure 3: (2-3): Cumulative Volume Diagram Showing
Delay due to an Incident.. 8
Figure 4: (2-5) Barrier guardrail system on a section
of roadway on I-64.. 14
Figure 5: (2-5) A standard concrete barrier on I-264.. 14
Figure 6: (2-6) A double stacked concrete barrier
on I-64.. 15
Figure 7: (3-1) Area Map of Hampton Roads freeway
system.. 17
Figure 8: (3-2) Thehr.incident Table Available from
TheUniversity of Virginia STL.. 19
Figure 9: (3-3) Example of Incident Occupancy.. 22
Figure 10: (3-4) Incident-caused Occupancy at Multiple
Stations.. 23
Figure 11: (3-5) Example of Rectangular Area Approach.. 25
Figure 12: (3-6) Example of Cumulative Arrival and
Departure Curves.. 27
Figure 13: (3-7) Application of Integral Estimation
on Delay Calculation.. 28
Figure 14: (3-8) Example of Capacity Reduction Calculation.. 30
Figure 15: (4-1) Histogram of the Frequency of Delay.. 39
Figure 16: (4-2) Histogram of the frequency of the
natural log of delay.. 40
Figure 17: (4-4) Histogram of Capacity Reduction Percentage
with bin size of 5%.. 44
Figure 18: (4-5) Histogram of Ln(Capacity Reduction)
due to Rubbernecking.. 45
Table 4.1: Statistics of Significant Impacts of Occupancy
due to Accidents.
343433
Table 4.2 Results of a
Binary Logit Model
373735
Table 4.3 Discrete Choice
Model Results.
373736
Table 4.4: Correlation
Coefficient Matrix.
414140
Table 4.5: Congestion Delay
Model Results.
424241
Table 4.6 Capacity Reduction
Model Outcome. 464644
Freeway
incidents cause major congestion throughout the United States every year. These incidents are often vehicle-vehicle accidentscrashes, which
many times cause major backups, sometimes for miles, along freeways. These overcrowded situations are costly gridlocks,
costing the travelers time and money. Other costs incurred especially due to an incident include increased
potential for secondary accidents, additional wear and tear on vehicles, and environmental
pollution. Historical statistics show
that more than 50% of urban freeway congestion is related tocaused by
incidents (Lindley, 1989). Reducing the amount of congestion with various
Intelligent Transportation Systems (ITS) or other methods is an important area
of research. Through this
the
research and implementation, time and money and even lives can be saved.
In the past, research has been focused on determining and modeling the
impacts of incidents in the direction of traffic of the incidents.
The results from the research
incident and traffic information collected can
be used to determine system performance measures such as delay,
capacity reduction, and most importantly travel
times.
Although
the modeling of incident traffic in the same direction is important, it deals
with only half of the traffic problem. Accidents also have an impact on the opposite direction of traffic.
Even though there are no lane blockages in the opposite direction of an
accident, there are reasons to believe that an impact exists on traffic.
This impact is due to rubbernecking. According
to the Webster Dictionary “rubbernecking” means to looks about,
stare, or listen with exaggerated curiosity. Individuals driving in the opposite direction of an accident are
often distracted by the incident. It is
the curiosity of the event that leads to distraction, which and then causes
a reduction in vehicle speeds. This reduction
in vehicle speeds begins to create congestion. Although a significant part of rubbernecking
is attributed to various human factors, there are other factors such as presence
of barriers that influence the form of rubbernecking.
This
thesis investigates the impact of traffic in the opposite direction of travel
from a vehicle accident. The effects
on the opposite direction of travel can be just as significant as the effects
of the incident on the same direction of travel. The study area for this project is the Hampton Roads area in Hampton,
Virginia, in the southeast corner of the state. Although this project uses only one study area,
the results obtained can be appropriately to any urban freeway system.
The incident type and time frame examined in this project is limited to
vehicle accidents in the year 2000. In
the year 2000, there were 840 documented accidents on the Hampton Roads freeway
system. This freeway system consists of approximately
10 miles of Interstate 64 from I-564 down south to Indian River Road and also
Interstate 264 eastbound from the I-64 interchange. This incident data has been collected from
the Virginia Department of Transportation traffic center in the Hampton Roads
area. This data is accessible through
the University of Virginia’s Smart Travel Lab.
To accomplish this investigation, there are certain objectives
to accomplish.
1.
Determine whether the rubbernecking impact on the opposite direction
traffic is significant.
2.
Investigate traffic data and calculate traffic delay
and capacity reduction in the opposite direction of travel.
1.Identify the factors
that influence the impacts
in terms of rubbernecking likelihood, traffic delay, and capacity reduction.
3. Identify the factors that influence the impacts in
terms of
rubbernecking likelihood, traffic delay and
capacity reduction.Recommend effective
countermeasure on rubbernecking in the
4. Recommend effective
countermeasure on rubbernecking in the
opposite direction.
Investigate traffic
data and calculate traffic delay in the opposite
direction of travel.
To determine whether
rubbernecking impact is significant, occupancy vs. time plots are created for each
incident. Significant changes in occupancy
are observed visually and documented.
Once these significant impacts were documented tthe rubbernecking likelihood (a probability of rubbernecking occurrence),
delay
(veh*hr)
and capacity reduction (percentage of capacity loss) areare derived using various methods.
These results for delay and capacity reduction for the Hampton Roads area andare then compared with the delay
and capacity reduction in other comparative studies. To identify the influencing factors, linear and binary
regression models are developed. The variables that are statistically significant
in these models are identified as outstanding. Based on the identification of the influencing factors,
mitigation measures are recommended.
The area focused in this study
is the freeway system in the
Hampton Roads
area in Virginia. This freeway system consists of approximately 10 miles
of Interstate 64 from I-564 down south to Indian River Road and also Interstate
264 eastbound from the I-64 interchange. Incident and associated accident data has been collected by the Hampton Roads Smart
Traffic Center and archived by the University of Virginia’s Smart Travel Lab. The incident type and time frameyear examined in this study are limited to vehicle
accidents in the year 2000.
The remaining parts of this thesis include background information, previous
research on incident delays and modeling, methodology explanation, results, analysis
and findings, and finally conclusions and recommendations.
Chapter Two consists of background information and previous research done
on the various measures this projectstudy
will investigated.
Chapter Three consists of the methodology used in this study
project. A detailed list of different techniques attempted
and used in this A detailed list of different techniques attempted
and used in this study isproject
will be laid
out.
Chapter Four contains the results and evaluation of the different measures
described in the methodology.
Chapter Five will documents the analysis
of the results section, including regression models, prediction models, and summary.
Chapter Six will includes
project conclusions and recommendations.
An incident is a traffic event that has an impact on traffic conditions. Incidents come in many forms, including disabled
vehicles, abandoned vehicles, various spills and debris, environmental events
(weather), and probably the most influential, vehicle accidents. These incidents
decrease flow and add additional congestion to the already crowded urban freeways.
This causes the Level of Service (LOS) to decrease and also adds to causes
of additional incidents. Previous research
has been done on traffic impacts of incidents, incident management, incident prediction,
and other topics pertaining to these random events. Although the information gathered for these studies is typically
for traffic in the same direction as the incident, it is still valuable for this
studyproject.
2.1 Incident Detection
Inductive loop detectors are the most commonly
used traffic counting equipment in the early ages of freeway management systems.
These detectors are typically copper wire loops embedded into a roadway.
As vehicles pass over these loops, an electromagnetic field is generated
and an electrical current is created through the wire. A change in electrical current indicates that
a vehicle passed this point and thus traffic flow, speed and occupancy can be
derived. These traffic data are then transmitted
to a Traffic Management Center (TMC) for traffic operations and management decision-making.
These data are collected in intervals.
Virginia’s Hampton Roads Smart Traffic Center collects these data is collected
in 20 second intervals. Currently, these
data are archived by the Smart Travel Lab at
the
University of Virginia.
Incident detection
techniques were first introduced in Chicago in the 1960’s (Ozbay, Kachroo, 1999).
Loop detectors in the roadway collected freeway data.
When the occupancy of a section freeway was 40% or greater, it was assumed
that there was an accident present. Today,
incident detection is still a growing area of research. The quicker an accident is detected, the quicker
the response time and clean-up will be. In the United States, the average accident notification time in
urban areas in the United States was estimated to be 5.2 minutes (Evanco, 1996).
While the first detection techniques used loops detectors, today’s techniques
consist of closed circuit televisions (CCTV) and cellular phone use.
Incident notification on cellular phones takes under one minute.
The flow and density relationship has been explored since the
publication of the L-W-R theory Lighthill and Whitman in 1955 and Richards in
1956. In general, flow is defined as the number of
vehicles to pass a point during a certain time. Density is referred to as the number of vehicles per roadway length.
Many models have been developed in an attempt to determine the correlation
between flow and density. Some models such as the Greenshield model use
single regime non-linear approach (see Figure 1), while others use multi-regime
complex models.
Figure
1: (2-1): Greenshield's
Flow vs. Density Model
The
model developed by a Northwestern University research team, takes into consideration
a two-regime linear model. The first regime
considers a positive slope linear relation of flow and occupancy during non-congested
conditions.
The
other regime considers a negative slope linear relation of flow and occupancy
for congested conditions.
Note that occupancy measures the percentage of time
vehicles occupy a section of roadway where a loop detector is installed. It is often used in place of density
because it is directly proportional to density
based on a factor of average vehicle length. This two- regime linear model is diagramed below.
Figure
2: (2-2):
Two-Regime Flow vs. Occupancy Model
Congestion
delay is referred to as the difference between actual travel time and the free-flow
time on a section of freeway (Hall, 1992). It
can be determined for a wide variety of traffic situations such as freeway and
arterial systems. In freeway systems,
delay is often thought about in terms of “recurrent” and “non-recurrent” delays.
Recurrent delays are delays experienced in everyday travel based on historical
data. Non-recurrent delays are delays caused by an
event or an incident and it can be broken up into two periods, immediate delay
and residual delay. Immediate delay is
the part of the delay incurred during the duration of the incident. The residual delay is thought of as the
delay sustained after the incident has cleared. It is estimated that 60% of all freeway delay is attributed to incident
producing non-recurrent delay (Lindley, 1989).
Incident-induced
delays have been calculated using a variety of methods. Morales (1986) developed a cumulative volume
approach to calculating freeway delays. In
this approach, two A cumulative volume
curves (one for arrival and the other for departure at an incident site) are plotted are plot is a curve
of the cumulative traffic volumes of a roadway on a time axis.
. The running value represents the total number
of vehicles to pass a certain point. Although
a single cumulative plot may indicate traffic behavior at a single point, additional
location cumulative plots will serve as important tools.
Delays can be calculated by plotting cumulative curves of traffic upstream
(Arrival Curve) and downstream (Departure Curve) of a bottleneck or incident.
Assuming that Arrival Curves and Departure Curves are equal (homogeneous
conditions) during normal freeway operations, the difference in The area between these two curves
these values represents the extra delay
due to an incident. This is shown below
in the diagram belowFigure 2-3.
Figure
3: (2-3):
Cumulative Volume Diagram Showing - Delay due
to an Incident
In
addition to this delay calculation based on the cumulative curves, it has been
suggested to adjust one or both of these curves.
Daganzo (1997) proposes a ‘virtual’ arrival curve be used to determine
delays. This virtual arrival curve is
a translation of the actual arrival curve based on the “number of items that would
have been seen directly upstream of the restriction” by the beginning of the incident
duration. The actual arrival curve is
translated to the right by a value of t
Al-Deek
et al (1994) developed a new method and made improvement to Morales’ approach
by looking at delays in time slices. They
incorporated vehicle speeds in conjunction with traffic volumes to develop a delay
formula. Assumptions they made
. include:
·
Traffic speed and volume data are determined from the
loop stations on a
roadway the ssegment and
these data are homogeneous throughout the segment
·
Incident delay is calculated with respect to a reference
(or base) average speed which reflects normal conditions that may or may not be
congested. The reference speed represents
a historical speed profile which may be used to segregate (distinguish between)
incident and non-incident (recurring) congestion.
A
drawback to this approach is that it required one-minute speed averages.
Smaller interval averages (less than one minute) could lead to “noisy”
data, while larger intervals (greater than one minute) do not allow for accurate
estimation of queue boundaries. The delay
was then calculated using the formulas shown below. Different from the queuing diagram approach where
incident duration,
capacities before and after an
incident and traffic demand are used to calculate delay, the incident delay is determined
using the time-slice method
by Al-Deek (1994). The individual
slices are summed up to form the total delay shown at the bottom of the reference
below.
for 
(1)
for 
(2)
for 
(3)
where
= Delay on freeway segment
“k” during time slice “i” (vehicle-hours)
= Length of segment
k (miles)
= Length of time slice “i” (minutes)
= Flow (from loops)
on segment “k” during time cslice “i” (vehicles per hour)
= Speed (from loops) on segment “k” during time
slice “i” (miles per hour)
= Reference average speed on segment “k” during
time slice “i” (miles per hour).
The total delay on the freeway
section that is caused by the incident is given by:
(4)
In addition to the queuing diagram
and real-time traffic data based approach; computer simulation is another effective way in modeling traffic delays during incidents.
The Highway Capacity Manual (HCM) defines freeway capacity as “the maximum
hourly rate at which persons or vehicles can reasonably transverse a point or
uniform section of a lane or roadway during a given time period.” During this time period, typically 15 minutes,
the freeway must be operating under ‘ideal conditions’. When these ideal situations are not present,
typically during an incident, the capacity is reduced. The HCM provides an equation for capacity reduction
caused by basic non-ideal conditions (lane width, heavy vehicle factor, number
of lanes, etc), but not for incident situations. The HCM states that a capacity reduction of 10-20
percent is characteristic of rainy weather. A
separate study by Jones and Goolsby (1970) revealed a 14 percent loss of capacity
due to rain. This capacity loss
is based on the maximum number of vehicles able to pass a section of roadway in
a given time. Although the maximum
possible flow should not change, the actual amount of vehicles passing a section
of roadway would be reduced. A separate study
by Jones and Goolsby (1970) revealed a 14 percent loss of capacity due to rain.
The
reduced capacity used in incident modeling is called the ‘effective capacity’
and is referred to the “expected roadway capacity, over time, after accounting
for the occurrence of incidents.” (Hall, 1992)
Having Aan effective capacity formula
based on incident characteristics would be an ideal solution of calculating capacity
reduction. However, such a formula would
not hold for situations for which more than one incident is present.
It has been proposed that in the case of multiple incidents, the incident
that has the largest impact on capacity should be used in the analysis. This is to say that the capacity reduction
“is not the sum of their magnitudes, but their maximum.” The research by Goolsby in 1971 and Lindley
in 1986 developed capacity reductions for certain lane and shoulder blockages. It was concluded that, “the effective
capacity loss due to incidents is far less than the effective loss due to removing a single
lane on a four-lane roadway.”
It is a result of a human response
to the surroundings such as Ffreeway signs, scenery,
billboard ads, and many other visual “eye-candy” cause this phenomenon. From a traffic operations standpoint, rubbernecking is a serious
issue that can sometimes create traffic congestion and even traffic incidents.
Rubbernecking is a result of a
human response to the surroundings. On the other hand, Tthe attention of the driver
is focused on these surroundings and less attention is on the roadway, making
rubbernecking a safety issue as well as a traffic congestion issue.
2.3.1
Rubbernecking Effects
A 2003 study by the Virginia Commonwealth University’s Transportation Safety
Training Center (TSTC) revealed that rubbernecking was the leading cause of vehicle
crashes. These rubbernecking accidents
were not caused by landmarks or other scenery; they were caused by drivers looking
at other vehicle crashes and other roadside traffic incidents. Rubbernecking caused by vehicle crashes and
other incidents accounted for sixteen-percent of all vehicle crashes, while the
total number of outside the car distractions accounted for 35-percent. There has been research performed on calculating
rubbernecking effects of traffic in the same direction of travel as the incident,
including studies .by. These
effects are due to rubbernecking of adjacent lanes and shoulders. Although, preliminary findings
of this project suggest significant rubbernecking
effects on the opposite direction of travel, there is no documentation available
on this topic.
Drivers cannot be
distracted by events or objects they cannot see. To mitigate
rubbernecking in the opposite direction, Tthis statement should calls for barriers that block
vision to opposite direction traffic conditions. Although, this theory would hypothetically be a solution
to opposite direction rubbernecking, a full implementation of such barriers is
yet to be seen. The Hampton
Roads freeway system, consisting of I-64 and I-264, described previously, implements
a variety of different barrier techniques. Certain
segments along the freeway system have only guardrails and a grass median dividing
the freeway traffic. Certain sections
of the Hampton Roads freeway are implemented with standard 42” concrete barriers, while
other sections have double stacked concrete barriers. Below are pictures from the Hampton Roads freeway system and the
median barriers associated with it. By having the data on the availability of the different types of barriers
on roadway segments, it is possible to investigate its relationship with the rubbernecking impact on opposite traffic conditions. The derived information could help develop mitigations
to reduce rubbernecking impacts on opposite direction traffic conditions.
Figure
4: (2-5):
Barrier guardrail Guardrail sSystem on
a sSection of rRoadway on
I-64
Figure
5: (2-5):
A standard Standard Cconcrete
bBarrier on I-264
Figure
6: (2-6): A double
Double
sStacked cConcrete bBarrier on
I-64
This
chapter provided an overview of previous work done related to this studyproject. From the initial research in the 1950’s to
the complex traffic modeling of the 21st century, traffic studies attempt to provide
a safer and more efficient roadway. Previous
research in Flow vs. Density modeling, delay calculations, capacity reduction,
and rubbernecking have all contributed as background information for this studyproject. The next chapter will show the procedure taken
to complete this studyproject.
Chapter
3: Methodology
In this study, the following procedure is adopted:
1. Extract incident data from Hampton Roads freeway
system
2. Filter data by limiting type of incidents
to “accidents”
3. Determine appropriate traffic data to collect
for given incidents.
4. Plot occupancy each vehicle accident
5. Determine significant impacts on both the
same and opposite direction of accidents
6. Determine location of incidents at station
level
7. Use Binary Logit Model for determining incident impact
modeling
7. Plot cumulative volume for identified significant
impact accidents
8. Calculate delay for identified significant
impact accidents
9. Determine capacity by retrieving historical
data
10. Plot Flow rate vs. Density (Occupancy) for
incidents
11. Compare of historical capacity and incident capacity
12. Calculate capacity reduction
13. Use
Analyze
Linear
Regression Modeling for delay cand capacity
reduction results
14.
Use Binary Logit Model for determining incident impact modeling
15. Evaluation of results and recommendations
based on analysis
Data for Tthe Hampton Roads freeway
system operated by Hampton
Roads Smart Traffic Center (HRSTC) were collected for this study. is located in the southeastern
part of Virginia. This freeway
system consists of approximately 20 miles of Interstate 64 from I-564 down south
to Indian River Road and Interstate 264 eastbound from the I-64 interchange. View map
of area below. Project freeway system
is outlined and highlighted in yellow. (See Figure 73-1).
Figure
7: (3-1): Area Map
of Hampton Roads freeway Freeway sSystem
The
HRSTC uses technology to improve motorist safety and convenience, reduce area
traffic congestion and decrease motorist travel time in the Hampton Roads area. There are currently 107 closed
circuit television cameras overlooking the 20 miles of freeway traffic. This is up from just 38 cameras in 2001. These cameras assist in incident management detection as
well as any other traffic tie up. These
incidents are then documented into a database with a variety of information regarding
the incident. In addition to the 107
cameras, operated by the HRSTC, the Hampton Roads area receives loop
detectors are installed
data from
over 200 stations along approximately 20 miles of freeway system. From Tthese detectors, send real-time traffic data are collected and sent to the STC Center. every two minutes.
Thanks to VDOT, thisThese loop detector data is accessible through
the Smart Travel Lab at the University of Virginia, in Charlottesville. and aid students and faculty both, to perform needed traffic research.
3.2
Incident Data
Specifically,
the incident database was retrieved from the Smart Travel Lab’s hr.incident
table in the
Oracle 8i database. This table contains
information on each incident and includes sub-tables with additional information.
The information on incidents include incident identification number, incident
begin time (including date/time in MM/DD/YYYY HH24:MI format), incident duration
(in minutes), incident type, weather, detection source, and a brief description
of the incident. Sub-tables include information
such as the roadway of occurrence, direction, location, number of lanes and shoulders
blocked, and information about the vehicle(s) involved (make/model/color/etc).
The incident identification
number is listed as a TMS Call Number including a year
and a identification number (Example, ‘2000-00001’). The duration of the incident is defined as
the time from when the incident is detected until the clearance of the incident.
The weather during the incident is documented and
includes conditions such as rain, snow, sleet, clear, cloudy and even ice. A complete layout of the hr.incident
table can be viewed below. A brief description
of the incident is sometimes given. This brief description may include the number of personal injuries
or a more accurate location of the incident. A summary of the database can be seen below.
Figure
8: (3-2): The hr.incident
Table Available from Thethe University of Virginia STL (Smith, 2001)
Though,
useful in other studieprojects,
some of this information is not pertinent. The
relevant information used in this project
included the Iincident Iidentification Numbernumber, incident
begin time, roadway, direction, location, duration of incident, weather, number
of lanes and shoulders blocked, and description. Incidents of the year 2000 were pulled from
the hr.incident table. During this time
period, available incident types included the following:
·
Abandoned Vehicles
·
Vehicle Accidents
·
Bridge Incidents
·
Debris
·
Disabled Vehicles
·
Severe OtherWeather Conditions
·
Other
It was decided that impacts due to rubbernecking would
most likely only occur during vehicle accidents. The acquired incidents were then filtered to
only include incident whose type was ‘accident’. Information about each accident is given in
the database.
Traffic data were collected based
on the date, time, and location of each incident that are included in the database. Once the incident data was pulled
and an accessible incident database was constructed, the next step was to collected
traffic data for the date and time and location
of each incident. Note that Tthe exact sites of the incidents
cannot be readily known from
the location code of the incident in the are not documented, rather a vague
location. Tthe hr.incident database. has the location code of the incident. Each location code is a section
of roadway typically 2 miles long and havings three or four detector
stations within. Thus, Ttraffic data had to be collected
for all stations within the location code of the incident. Total volumes, average speeds, and average
occupancy were collected for an extended period, starting from one hour before the incidents
beginning time and
ending at 1 one hr after
the duration of the incident. This period
duration time
will accounts
for the time period period
where traffic is operating normally before the incident and where
traffic is recovering and once again operating normally after the duration of
the incident. See four-step incident process
above. By collecting data
for this extended period of time it is ensured that the full effects of the incident
are captured. A working SQL code was developed
to expedite this long process. A copy
of this code can be found in Appendix A.
GivenAfter the traffic and
incident information
for oneeach incident had been collected,
the incidents werep groupeduttgetting together ointo a
one single spreadsheet., The next step was to determine wwhether eachthe an incident
hads significant impact