USF Libraries
USF Digital Collections

Operational evaluation of right turns followed by U-turns at signalized intersections as an alternative to direct left turns

MISSING IMAGE

Material Information

Title:
Operational evaluation of right turns followed by U-turns at signalized intersections as an alternative to direct left turns
Physical Description:
Book
Language:
English
Creator:
Fan, Jingjing
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
DLT
delay
travel time
RTUT
Dissertations, Academic -- Civil Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Direct left turn (DLT) movements have a lot safety and operation problems. They increase conflicts, delay and reduce capacity of the roadway. Increasing, many states and transportation agencies are considering using right-turns followed by u-turns as an alternative to direct left turns. Recently, Florida state are planning to close median openings in order to reduce the safety and operational problems caused by DLT. The left-turn egress movement would then to be made by turning right to the arterial road and then making a u-turn at an available downstream median opening or a signalized intersection. FDOT sponsored a project in 2001 (Methodology to quantify the Effects of Access Management Treatments on Roadway Operations and Safety) which has provided useful information on operational effect of right-turn followed by u-turn at median opening as an alternative to the direct left turn from driveways. However, the impact of the treatment---right-turn followed by u-turn at signalized intersection as an alternative to direct left turn is still not clear. A operational comparison of this treatment would be valuable. Field data collection was conducted to quantify the operational effects of right turn followed by u-turn at signalized intersection versus direct left turns. Totally eight sites around the Tampa Bay area were selected and more than 300 hours field data including delay, travel time, traffic volume, traffic control, headway, signal timing and geometric data were collected. Data reduction was conducted after that and several statistic models were developed evaluating total travel time, travel delay of these two movements and the ratio of drivers choice of RTUT versus DLT. As the conclusion, the study found that RTUT followed by U-turn at signalized intersection has longer delay and longer travel time as compared to DLT and the existence of long cycle downstream signal will discourage driver's selection of RTUT. There conclusions are very important in that they provided decision makers with useful information regarding the selection of the suitable U-turn location.
Thesis:
Thesis (M.S.C.E.)--University of South Florida, 2004.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Jingjing Fan.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 113 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001478787
oclc - 56563592
notis - AJS2477
usfldc doi - E14-SFE0000431
usfldc handle - e14.431
System ID:
SFS0025123:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Operational Evaluation of Right Turns Followed by U-turns at Signalized Intersections as an Alternative to Direct Left Turns by Jingjing Fan A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Civil Engineering Department of Civil and Environmental Engineering College of Engineering University of South Florida Major Professor: Jian John Lu, Ph.D. Ram Pendyala, Ph.D. Manjriker Gunaratne, Ph.D. Date of Approval: April 30, 2004 Keywords: RTUT, DLT, Operations, Travel Time, Delay, Model Copyright 2004 Jingjing Fan

PAGE 2

ACKNOWLEDGEMENTS It is of great pride that I had a chance to work with the brilliant minds affiliated to the Department of Civil and Environmental Engineering, University of South Florida. First, I would like to thank Dr. Jian John Lu for his continued support and able guidance in my research efforts. I would also like to thank Dr. Ram Pendya la and Dr. Manjriker Gunaratne for serving on my committee. A nd I would like to acknowledge the Florida Department of Transportation for providing f unding for the research. Assistance of the Graduate Research Assistants at the Departme nt of Civil and Envir onmental Engineering, who participated in the difficult and time c onsuming data collection and dada reduction process, is also highly acknowledged.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT vii CHAPTER 1. INTRODUCTION 1 1.1 Background 1 1.2 Research Statement 4 1.3 Research Purposes and Objectives 6 1.4 Outline of the Report 7 CHAPTER 2. LITERATURE REVIEW 8 2.1 General 8 2.2 Current Rules and Re gulations in Florida 8 2.3 Delay and Travel Time Models 10 2.3.1 Delay Models at Signalized Intersection 11 2.3.2 Delay Models at Unsignalized Intersection 15 2.4 U-turn as an Alternative to Direct Left Turn 17 2.5 U-turn at Signalized Intersections 22 2.6 Summary 26 CHAPTER 3. METHODOLOGY 28 3.1 Delay and Travel Time Models 28

PAGE 4

ii 3.1.1 Operations Analysis of Direct Left Turns 29 3.1.2 Operations Analysis of Right Turn Plus U-turns 31 3.2 Driver Selection of RTUT 34 3.2.1 Driver Selection of RTUT at Signalized Intersection 35 3.2.2 Fitting a Binary Logistic Regression Model 37 3.3 U-turn Effects on Signalized Intersection 38 3.3.1 Procedures for Estimating Saturation Flow Rate 38 3.3.2 U-turn Effects on Left-turn Saturation Flow Rate 41 3.3.3 Adjustment Factor for U-turn Movement 43 CHAPTER 4. DATA COLLECTI ON AND REDUCTION 47 4.1 Site Selection 48 4.2 Data Collection 55 4.3 Data Reduction 58 CHAPTER 5. OPERATIONAL EFFECTS 60 5.1 General 60 5.2 Average Delay 61 5.2.1 Delay Model for Direct Left Turn 61 5.2.2 Delay Model for Right Turn Followed by U-turn 66 5.2.3 Delay Comparison of DLT and RTUT 72 5.2.4 Delay Comparison of Two U-turn Approaches 74 5.3 Average Total Travel Time 75 5.3.1 Travel Time Model for DLT 75 5.3.2 Travel Time Model for RTUT 78

PAGE 5

iii 5.3.3 Travel Time Compar ison of DLT and RTUT 82 5.3.4 Travel Time Comparison of Two U-turn Approaches 83 5.4 Amount of RTUT under Both Choices 84 5.5 U-turn Effects on Signalized Intersection Capacity 89 5.6 Summary 94 CHAPTER 6. Summary and Conclusions 95 6.1 Summary 95 6.2 Conclusions 97 6.3 Recommendations 98 REFERENCES 101

PAGE 6

iv LIST OF TABLES Table 2.1 Left-turn Delay Under Different Volume Conditions 20 Table 2.2 Adjustment Factor for U-turns 28 Table 3.1 Regression Results for Ratio of RTUT at Median Opening 35 Table 4.1 Description of Selected Sites 49 Table 5.1 Descriptive Statisti cs of the Collected Data 63 Table 5.2 Regression Results for Delay Models of DLT 63 Table 5.3 Descriptive Statisti cs of the Collected Data 69 Table 5.4 Regression Results for Delay Model of RTUT 69 Table 5.5 Descriptive Statisti cs of the Collected Data 76 Table 5.6 Regression Results for Delay Model of RTUT 76 Table 5.7 Descriptive Statisti cs of the Collected Data 79 Table 5.8 Descriptive Statisti cs of the Collected Data 79 Table 5.9 Regression Results for Delay Model of RTUT 86 Table 5.10 The Regression Results for the Average Queue Discharge Time 91 Table 5.11 Adjustment Factor for U-turn Movements 93

PAGE 7

v LIST OF FIGURES Figure 2.1 Vehicular Conflict Points at a Typical Four-way Intersection Versus a Directional Median Opening 18 Figure 2.2 Unsignalized Direct ional Median Openings 19 Figure 2.3 Michigan U 22 Figure 3.1 DLT Egress Movements 29 Figure 3.2 RTUT Movements 31 Figure 3.3 The Distance from Driveway to Signalized Intersection 34 Figure 3.4 Comparison of Queue Discharge Patterns for Different Percentage of U-turning Vehicles in Left-turn Lane 41 Figure 3.5 The Relationship between Average Discharge Time & The Percentage of U-turning Vehicles 45 Figure 4.1 Site 1 Fowler Avenue and 56 th Street 50 Figure 4.2 Site 2 Fowler Avenue and 22 nd Street 51 Figure 4.3 Site 3 Hillsborough Avenue and Webb Street 51 Figure 4.4 Site 4 Dale Mabry Hi ghway and North Dale Street 52 Figure 4.5 Site 5 Bruce B. Downs Boulevard and Fletcher Avenue 53 Figure 4.6 Site 6 34 th Street and 54 th Street 53 Figure 4.7 Site 7 54 th Street and 22 nd N. Street 54

PAGE 8

vi Figure 4.8 Site 8 Dale Mabry Hi ghway and Maple Dale Street 55 Figure 4.9 Equipments Setup in the Field 56 Figure 4.10 Equipments Setup in the Field 56 Figure 4.11 Equipments Setup in the Field 57 Figure 4.12 The Basic Camera Locations in the Field 57 Figure 5.1 Traffic Flows Affecting the Delay of DLT 62 Figure 5.2 Curves for the Average Total Delay for DLT 66 Figure 5.3 Two Different Weaving Patterns 67 Figure 5.4 Curves for the Average Total Delay for RTUT 72 Figure 5.5 Comparison of Average Delay of Two Movements 73 Figure 5.6 Comparison of Average Delay of Two U-turn Approaches 74 Figure 5.7 Curves for the Average Total Travel Time for DLT 78 Figure 5.8 Curves for the Average Total Travel Time for DLT 82 Figure 5.9 Curves for the Average Total Travel Time for RTUT 83 Figure 5.10 Comparison of Average Travel Time of Two Movements 84 Figure 5.11 Travel Time Comparis on of Two U-turn Approaches 88 Figure 5.12 Average Queue Discharge Time versus the Percentage of U-turning Vehicles in Queue 91

PAGE 9

vii OPERATIONAL EVALUATION OF RIGHT TURNS FOLLOWED BY UTURNS AT SIGNALIZED INTERSEC TIONS AS AN ALTERNATIVE TO DIRECT LEFT TURNS Jingjing Fan ABSTRACT Direct left turn (DLT) movements have a lot safety and opera tion problems. They increase conflicts, delay and reduce capacity of the roadway. Increasing, many states and transportation agencies are considering us ing right-turns followed by u-turns as an alternative to direct left turns. Recently Florida state are planning to close median openings in order to reduce the safety and operational problems caused by DLT. The leftturn egress movement would then to be made by turning right to the arterial road and then making a u-turn at an available downstream me dian opening or a signalized intersection. FDOT sponsored a project in 2001(Methodo logy to quantify the Effects of Access Management Treatments on Roadway Operatio ns and Safety) which has provided useful information on operational effect of right-turn followed by u-turn at median opening as an alternative to the direct left turn fr om driveways. However, the impact of the treatment---right-turn followed by u-turn at si gnalized intersection as an alternative to direct left turn is still not clear. A opera tional comparison of this treatment would be valuable.

PAGE 10

viii Field data collection was conducted to quantify the operational effects of rightturn followed by u-turn at signalized intersectio n versus direct left turns. Totally eight sites around the Tampa Bay area were select ed and more than 300 hours field data including delay, travel time, traffic volume, traffic control, headway, signal timing and geometric data were collected. Data reduc tion was conducted after that and several statistic models were developed evaluating to tal travel time, travel delay of these two movements and the ratio of drivers choice of RTUT versus DLT. As the conclusion, the study found that RTUT followed by U-turn at signalized intersection has longer delay and longer tr avel time as compared to DLT and the existence of long cycle downstream signal will discourage drivers selection of RTUT. There conclusions are very important in that they provided decision makers with useful information regarding the selection of the suitable U-turn location.

PAGE 11

1 CHAPTER 1 INTRODUCTION 1.1 Background In the past few decades, more and more states came to realize the importance of access management to the state roadway syst em. In 1979, the nations first systemwide comprehensive access management program was adopted in Colorado. In 1988, the Florida Legislature adopted the State Highway System Access Management Act, Statutes 335.18, which was an important legal foundation of Florida statewide access management program. In 2003, the Transporta tion Research Board published the first national access management manual. Since 199 3, there have been five national access management conferences (USDOT/FHS 1993, 1996, 1998, 2000 and 2002) held in this nation. Recently, several NCHRP projects we re established to conduct comprehensive research in this area. Over 100 access ma nagement techniques were identified and divided into four broad categ ories: traffic operations, traffic safety, environment, and economic (including transporta tion service and land use). Access management was defined as the system atic control of the location, spacing, design, and operation of driveways, medi an openings, interchanges, and street connections to a roadway. Access Manageme nt helps achieve the necessary balance between traffic movement and property access by careful control of the location, type, and design of driveways and street intersec tions. This is accomplished by classifying

PAGE 12

2 highways with respect to the level of access an d mobility they are expected to provide, and then, identifying and applying the most effe ctive techniques to pres erve that function. The benefits of access management include improved safety, improved traffic flow and fuel economy, increased capacity and re duced delay and vehicle emissions. As indicated in Access Management Manua l, one of the major principles of access management is to use nontraversable me dians to manage left turn movements. Left turn movements have been considered as one of the major resources of traffic operations and safety problems in the roadway system. A lot of previous studies have indicated that left turn move ments increase delays, conflic ts, and accidents, and they reduce capacity and mobility in the major tr affic. For example, a research conducted in 1998 revealed that 74% of tota l access-related crashes invo lve left turning vehicles. Many states and transportation agencies have started using non-traversable medians in state highway system. Since 1993, the FDOT mandated that all new or reconstructed multi-lane arterial roads with design speeds over 40 mph be designed with restrictive medians. In additi on, Florida uses directional median openings to manage left turns and crossing maneuvers. By closing existing median openings in some major arterial roads or replacing them with di rectional median openings, Florida prohibits left-turn exits onto major arterials. Left turn egress movements would be made by turning right onto the arterial road and then maki ng U-turns at downstream median opening or signalized intersection. Replacing full median openings with dire ctional median openings has been found to substantially reduce crash rate. In prac tice, however, to close the existing median opening is very sensitive and is very diffic ult to be handled. Business owners generally

PAGE 13

3 believe that the median closure will have some adverse impacts on their business. In addition, some people often oppose being for ced to make a righ t-turn followed by a U-turn due to the perception that it results in a much longer travel time than a direct left turn or a believe that U-turns are unsafe. Several studies have been conducted in Florida and nationally concerning the economic impacts of non-traversable medians. These results generally indicated that median projects have little overall adverse impact on business activity. However, there are fewer studies regarding th e operational effects of pr oviding right turns followed by U-turns as an alternative to direct left turns before 2000. The project sponsored by the FDOT in 2000 (Methodology to Quantify the Effects of Access Management Treatments on Roadway Operations and Safe ty) provided very useful information on the access management treatment on Right-Turn followed by U-turn at Median Opening as an Alternative to the Direct Left Turn from Driveways and Side Streets. The study took three basic approaches in evaluating this issue and involved huge amount of field data co llection where extensive data were gathered at several appropriate locations using video cameras followed by lengthy data reduction process in the lab. This project proved that under high through traffic volume condition, direct left-turns resulted in higher traffic conflicts, stop delay and travel time as compared with right turns followed by U-turns. The 2000 research was focused on U-turn at median opening. In the real world, however, there are many other conditions in which U-turn is accommodated at signalized intersections. Very few researches have been conducted concerning th is specific condition and the operational effects of right turns fo llowed by U-turns at signalized intersection

PAGE 14

4 are still not clear. In add ition, there are many controversie s surrounding the selection of different U-turn approaches, including pr oviding U-turn in advance of signalized intersection, U-turn at signalized intersection, and U-turn after signalized intersection. Currently there is no widely accepted pro cedure to estimate the operations of these different U-turn approaches. With the increasing installation of restrictive medians on multilane arterial roads, it is becoming more and more urgent to conduct an extensive study to address these issues. 1.2 Research Statement To make left turns from a driveway, dr ivers usually have two alternatives: (1) making direct left turns from the driveway into the arterial, (2) making right turns followed by U-turns at downstream median opening or signalized intersection. The past FDOT project (Methodology to Quantify the Ef fects of Access Management Treatments on Roadway Operations and Safety) prove d that under high through traffic volume condition, direct left-tur ns resulted in higher traffic conf licts, stop delay and travel time as compared with right turns followed by U-turn s. In the real world, however, there are many other conditions in which U-turn is provided at downstream signalized intersections. As compared with U-turns at median opening, there are three new issues need to be considered concerning U-turn at signalized intersection. First, in addition to the characteristics such as major street volume, driveway volume, left turn in volume, distribution of traffic, new factors need to be considered incorporating the effects of signalized in tersection on operations of RTUT movement such as signal timing, signal control type, right on red from side street, and the distance from the driveway to signalized intersection, etc.

PAGE 15

5 Second, U-turn movements may have some adverse impacts on signalized intersection capacity. For example, U-turn movement may have longer discharge headway and start up lost time as compared with left turn movement. Two previous studies have substantiated th is assumption. However, similar research hasnt been conducted in Florida. Third, drivers selection of RT UT is another issue need to be considered. The previous project indicated that more drivers make right turns followed by U-turns at downstream median opening instead of direct left turns from driveway when ther e was relatively high left-turn-in flow rate and major-road throughtraffic flow rate. When considering U-turn at signalized intersections, however, drivers choice behavior may be different from the condition where U-turn is provided at median opening. Clarifying the drivers preference of U-turn at median opening or signalized intersection will help decision maker select suitable U-turn approaches. Few researches have been conducted concer ning operational effects of right turns followed by U-turns at downstream signalized intersection as an alternative to direct left turns from driveway. Transportation engine ers generally rely on broad or subjective methods to analyze the performance of this sp ecific treatment. Very few field data are available to substantiate the reasonablen ess of this specific access management technique. This is also one of the reasons why this treatment is still very controversial in the real world. In order to address public conc erns and to analyze the potential cost and benefits of the treatment, it is necessary to conduct a systematic study based on field observations to quantify the operational perf ormance of RTUT at downstream signalized intersection versus di rect left turns.

PAGE 16

6 In this study, a field experiment was c onducted to collect data at 8 sites in the Tampa Bay area. Delay, travel time and pe rcentage of drivers choosing right turn followed by U-turn at downstream signalized in tersection rather than a direct left turn were used to quantify the operational eff ects of this specific access management treatment. U-turn effects on signalized intersection capacity will be estimated by determination of the U-turn adjustment f actor on left-turn satu ration flow rate. The analysis results will be compared with the results of 2000 research, so that the operations of two U-turn treatments can be compared The research results can be directly applied to evaluate the operational effects of median trea tments such as installing restrictive median, closing existing median openings, and replaci ng a full median opening with a directional median opening. 1.3 Research Purposes and Objectives The primary purpose of this project was to conduct a detailed evaluation and investigation on a widely used access mana gement technique: right-turns followed by U-turns at signalized intersection as an alte rnative to direct left turns from driveway This research took two main approaches to evaluate this specific access management technique including operational analysis and conf lict analysis. Operati onal effects of right turn followed by U-turn at signalized intersect ion will be quantified in this report through field studies and data collection. Empirical models concerning delay, travel time and drivers selection of RTUT will also be developed using collected field data. More specifically, the objective consists of the following parts: 1) To determine under what volume conditions (major-road, left-turn-in, and driveway) would DLT have more delay or travel time as compared to RTUT.

PAGE 17

7 2) To estimate delay for RTUT at signalized intersection as a function of conflicting major and minor-road flow rate s and signalization conditions. 3) To estimate delay for DLT as a function of conflicting major and minor-road flow rates. 4) To determine under what roadway traffic and geometric conditions would drivers using RTUT instead of DLT. 5) To evaluate U-turn effects on signalized intersection capacity. 6) To compare two widely used U-turn approaches: U-turn in advance of signalized intersection and U-turn at signalized intersection, from operations point of view. 1.4 Outline of the Report This report consists of five chapters Chapter 1 provides an introduction of the whole project. Chapter 2 describes a summary of the past studies in this area. Chapter 3 explains the methodology employed in achievi ng the previously mentioned objectives, which consists of three subsections: delay and travel time models, drivers selection of RTUT or DLT and U-turn effects on signali zed intersection. Chapter 4 focuses on data collection and data reduction procedure. An alysis results and research findings are presented in Chapter 5. Chapter 6 provides summary, conclusions and recommendations of this research.

PAGE 18

8 CHAPTER 2 LITERATURE REVIEW 2.1 General In order to develop a methodology to quan tify the operations of a widely used access management treatment: providing right turns followed by U-turns at downstream signalized intersection as an a lternative to direct left turn s, extensive work was conducted to search current rules and regulations, de sign standards and policies in Florida and nationally. In addition, past studies and reports related to this topic will also be searched and reviewed. Generally, the references can be categorized into three parts: current rules and regulations, delay and travel time mode ls, and operational effects of U-turns. 2.2 Current Rules and Regulations in Florida Access management is a relatively new response to traffic congestion and accidents. In the past few decades, more and more states realized the importance of managing access from private property to th e state highway system. In 1988, the Florida Legislature enacted the State Highway System Access Management Act, Statutes 335.18. The act indicated that Regula tion of access to the State Highw ay System is necessary in order to protect the public health, safety, and welfare, to preserve the functional integrity of the State Highway System, and to promote the safe and efficient movement of people

PAGE 19

9 and goods within the state. It was also indica ted by this act that w ithout such a program, the health, safety, and welfare of the residents of this state may be placed at risk, due to the fact that unregulated access to the State Highway System is one of the contributing factors to the congestion and f unctional deterioration of the system. As mandated by this act, the FDOT adopted administrative rules: Chapter 14-96 and Chapter 14-97. Based on these, Florida established the statewide comprehensive access management program. The purpose of rule Chapter 14-96 is to implement the State Highway System Access Management Act for the regulation and control of vehicular access and connection points of ingress t o, and egress from, the State Highway System, and other transportation facilitie s under the Departments jurisdic tion except for limited access facilities. Rule 14-96 is known as the Permits Rule in that it de als with the access permitting procedures, permit requests, and permit modifications or conditions. Rule Chapter 14-97 adopts an access classification system and standards to implement the State Highway System Access Management Act of 1988 for the regula tion and control of vehicular ingress to, and egress from, the St ate Highway System. The implementation of the classification system and standards is in tended to protect public safety and general welfare, provide for the mobility of pe ople and goods, and preserve the functional integrity of the State Highway System. As it in dicated in this rule, All segments of the State Highway System shall be assigned an access classification and standard. The standards shall be the basis for connection permitting and the planning and development of Department construction projects. Rule 14-97 is also known as the Standards Rule since it governs access classification of hi ghways and provides spacing standards for driveways, median openings, and signals.

PAGE 20

10 Florida is heavily encouraging restrictiv e medians on its higher designed at-grade arterial roadway. The 1993 Multi-lane Facilities Median Policy required that all new or reconstructed multilane highways with a desi gn speed over 40 mph be designed with a restrictive median. It also directed designers to find ways to use restrictive medians in all multi-lane projects, even those with de sign speed below 40 mph. One of the major purposes of installing restrictive medians is to eliminate left turn movements. By closing existing median openings in some major arterial roads or replacing th em with directional median openings, Florida prohibits left-turn exits onto major arte rials. Thus, the left turn egress movements would be made by turning right onto the arterial road and then making U-turns at a downstream median ope ning or signalized intersection. Another related policy is the Median Opening Decision Process (October 1996). The purpose of this procedure is to provide di rection for engineering decisions to support deviation from median opening standards stated in Administrative Rule 14-97. This policy established a process for handling desi gn decisions related to deviations from median opening standards, including techni cal and policy guidelines and an access management review committee to consider requ est for deviation. As it indicated in this procedure, there are three essential principl es that should be used when considering deviations from median openi ng and signal spacing standards: traffic safety, traffic efficiency, and functional integrity. 2.3 Delay and Travel Time Models Delay and travel time are two important measures of effectiveness (MOEs) of traffic operations. In the real world, people often opposed ma king a right turn followed by a U-turn because many of them generally be lieve that right turn followed by U-turn

PAGE 21

11 has much longer delay and travel time as comp ared with DLT. However, very few studies have been conducted concerni ng the delay and travel time of right turns followed by U-turns as a whole procedure, especially for the specific condition where U-turn is provided at signali zed intersection. 2.3.1 Delay Models at Signalized Intersection Delay is an important parameter that is used in the estimation of the level of service at signalized intersecti ons. In addition, delay is a measure that most directly relates the drivers experience, in that it describes the amount of time consumed in traversing the intersection. There are many diff erent ways to define delay. As illustrated in Traffic Engineering (Second Edition), the most frequently used forms of delay are defined below: 1) Stopped Time Delay: Stopped time delay is defined as the time a vehicle is stopped while waiting to pass through the intersection. 2) Approach Delay: Approach delay includes st opped time, but also includes the time lost when a vehicle decelerates from its ambient speed to a stop, as well as while accelerating from the stop back to its ambient speed. Sometimes it is very difficult to measure decelerate delay in the field w ithout sophisticated tracking equipment. 3) Travel Time Delay: Travel time delay is defined as the difference between the drivers desired total time to traverse the intersection and the actual time required to traverse it. 4) Time-in-Queue Delay: Time-in-Queue delay is the total time from a vehicle joining an intersection queue to its discharge across the stop-line or curb-line.

PAGE 22

A lot of researches have been conducted to estimate delay at signalized intersections. Among them, the most often quoted model is perhaps the Webster model. In this model, Webster estimated delay at isolated traffic signals as a sum of uniform delay (d u ) and random delay (d r ). Uniform delay is the delay at signalized intersection assuming uniform arrival rate. As indicated in HCM 2000, the uniform delay can be expressed as: d 1 = CgXCgC,1min115.02 (2-1) where, d 1 = uniform delay assuming uniform arrivals (s/veh); C = Cycle length (s); cycle length used in pretimed signal control, or average cycle length for actuated control; g = effective green time for lane group (s); green time used in pretimed signal control, or average lane group effective green time for actuated control X = v/c ratio or degree of saturation for lane group. The random delay can be expressed as: XcXdr12 (2-2) where c is the capacity of a lane group. Webster also estimated an adjustment term by simulation and concluded that control delay can be approximated as d = 0.9 (d u + d r ). Webster model is a very classical delay estimation model and it was widely accepted as an accurate depiction of delay for the idealized case of uniform arrivals, stable flow and no initial queue. Following Websters work, a number of stochastic 12

PAGE 23

13 models have been developed, including those by Newell, Miller. McMeil, and Heidemann. These models generally assume that arrivals are Poisson distributed, with an underlying average rate of ve hicles/unit time, and the system remains under-saturated over the analysis period. Therefore these models can not be directly used when traffic demand exceeds intersection capacity for a significant period of time. The HCM 2000 use control delay as the cr iteria for LOS of both signalized and unsignalized intersections. In this manual, th e total delay was defined as the difference between the travel time actually experienced and the reference travel time that would result during base conditions, in the absence of incident, cont rol, traffic, or geometric delay. Control delay was defined as the propo rtion of total delay attributed to control measures. Control delay includes initial dece leration delay, queue move-up time, stopped delay, and final acceleration delay. With resp ect to field measurements, control delay is defined as the total elapsed time from the tim e a vehicle stops at th e end of the queue to the time the vehicle departs from the stop line. The HCM 2000 developed a procedure to estimate average control delay for a given lane group. The average control delay wa s divided into three components. The first component represents delay assuming the uniform arrival of vehicles. The second component adds an incremental delay to acco unt for stochastic arrivals and occasional oversaturation. The third component adds dela y as the result of an initial queue at the beginning of the analysis period. The average control delay per vehicle for a given lane group is given by the following equation: d = d 1 (PF) + d 2 + d 3 (2-3)

PAGE 24

where, d = control delay per vehicle (s/veh); d 1 = uniform control delay assuming uniform delays (s/veh); PF = uniform delay progression adjustment factor, which accounts for effects of signal progression; d 2 = incremental delay to account for effect of random arrivals and oversaturation queues, d 3 = residual demand delay to account for initial queues. In this model, d 1 has the same form as the uniform delay in Webster model (2-1). The incremental delay d 2 can be estimated by the following equation: d 2 = 900T cTklXXX8112 (2-4) where T is the length of the analysis period (hrs), k is the incremental delay factor that is dependent on controller settings, and I is the upstream filtering/metering adjustment factor. The model is adjusted for traffic-actuated control with factor k depending on unit extension and degree of saturation. For isolated pretimed signals k= 0.5 and I=1.0. A couple of researches have been conducted also to test and compare existing delay models. Luttinen compared the HCM2000, Danish DanKap, and Swedish Capcal 2 models with simulation data and indicated that HCM 2000 underestimate capacity and overestimate delay at high degrees of saturation (X> 0.75). For traffic-actuated control HCM 2000 estimated somewhat too low delays at low degrees of saturation. Another problem with HCM 2000 model is that it does not consider the extra delay due to the blocking effect of short turning lanes. This effect is emphasized especially in the already problematic situation with high degrees of saturation and a large number of left-turning vehicles. 14

PAGE 25

15 Qureshi (2003) suggested using simulation soft ware to estimate delay for intersections with actuated control. He also illustrated that using current analytical procedures to estimate delay at actuated co ntrolled signalized intersection has the following limitations: 1) The variability of traffic demand within a given control period cannot be fully considered. Analyses are typically usi ng the average demand within a period. 2) Unusual arrival and service patterns that do not follow traditional statistical distributions cannot be modeled. 3) The models cannot be used to analyze real -time traffic operations, as such operations are typically concerned with instantaneous and cyclic flows rather than average flows. 2.3.2 Delay Models at Unsignalized Intersection There have been many studies on deve loping capacity and delay models to evaluate traffic operations at unsignalized intersections. Radwan and Kumares developed a delay-flow rate relationship for u ndivided and divided 4-lane highways. In this study, delay was defined as seconds per vehicle for ma jor and minor roads. The flow rate is the combination of major-minor flow rate. A linear fitting was tried between delay per vehicle in seconds and flow rates on major highways. It was found that the slope of the fitted line for the undivided highway case was much higher than that for the divided highway case. This result was as expected because the highway median permits drivers to perform their crossing maneuver in two steps and consequently, they experience less delay. Moreover, delay for the undivided highway was found to be less than the delay for

PAGE 26

16 divided highways as long as the major fl ow rates were less than 290 and 315 vph for minor rates of 100 and 50 vph, respectively. The Highway Capacity Manual has set up a procedure to estimate the delay, capacity, and level of service of unsignalized intersections. A study by Tian, Kyte and Colyar indicated that using the HCM procedure could overestimate delay and underestimate capacity when a minor street left-turn vehicle would cross the nearest approach and stop in the median position while waiting to join the major street traffic, resulting in a two-stage gap acceptance process. The two-stage priority situation as it exists at many un-signalized intersections within multilane major streets provides larger capacities and smaller delay co mpared to intersections wi thout central storage areas. A study by Robinson presented theoretical models to adjust the basic capacity or delay equations to account for some common occurre nces at TWSC intersections: two-stage gap acceptance, flared minor-street approaches, effects of upstream signals, and effects of pedestrians. However, these theoretical models have not been calibrated against empirical data. The HCM 2000 provided updated models to calculate the capacity and delay of unsignalized intersections, including two-way stop-cont rolled (TWSC) and all-way stop-controlled (AWSC). The procedures for TWSC intersections also account for certain conditions such as effects of upstream signals and of median storag e where minor street vehicles can proceed through the intersection in a two-stop process, namely a two-stage gap acceptance process. However, as st ipulated in the HCM 2000 methodology, each major-street approach can have up to two through lanes and one exclusive right and/or left-turn lane. Each minor-street approach can have up to three lanes, a maximum of one

PAGE 27

lane for each movement. This is a limitation of the research on which the procedures are based. The HCM 2000 uses the following model to estimate control delay at TWSC intersections: d = TcvccvcvTcxmxxmxmxxmxxm4503600119003600,,2,,, + 5 (2-5) where, d = control delay (s/veh); v x = flow rate for movement x (veh/hr); C m, x = capacity of movement (veh/hr); and T = analysis time period (hr) (T=0.25 for a 15-min period) As discussed in the research scope, only major arterials with 6 to 8 through lanes (3 or 4 each direction) were investigated for delay and travel time comparison in this study. Therefore, the HCM procedure for unsignalized intersections could not be directly applied to estimate the delay or travel time of right-turns and left-turns at driveways. 2.4 U-turn as an Alternative to Direct Left Turn In the roads designed with restrictive medians, left turn egress movements are only permitted at full median openings. As indicated before, however, left turn movements posed a lot of problems to roadway safety and operations. Many states have taken very strict restriction on median opening spacing to reduce the density of full median opening. The Access Management Manual illustrated that when providing a full median opening on the fringe of an urban area, it is important to consider the potential for future signalization. A full median opening that is located where signalized intersection will 17

PAGE 28

interfere with efficient traffic progression may need to closed or reconstructed as a directional median opening. The directional median opening means an opening in a restrictive median which provides for U-turn only, and/or left-turn in movements. Replacing full median opening with a directional median opening will reduce conflict points, simplify driving tasks, and was found to significantly reduce crash rates (Figure 2.1). Figure 2.1 Vehicular Conflict Points at a Typical Four-way Intersection Versus a Directional Median Opening Florida makes extensive use of directional median openings in the State Highway System. By closing existing median openings in some major arterial roads or replacing them with directional median openings, Florida prohibits left-turn exits onto major arterials. Left turn egress movements would be made by turning right onto the arterial road and then making U-turns at downstream median opening or signalized intersection (Figure 2.2). 18

PAGE 29

Figure 2.2 Unsignalized Directional Median Openings (a) Downstream from the Signalized Intersections and (b) Upstream from the Signalized Intersections Several studies have been conducted to evaluate the operational effects of providing U-turns at median openings as an alternative to direct left turns from a driveway. An analytical model was developed and calibrated in NCHRP 420 to estimate the travel time savings when unsignalized left turns are diverted for various distances. It can apply to both suburban and rural environments where there are no nearby traffic signals. The key findings are as follows: First, a right turn followed by a U-turn will require up to one minute of travel time, assuming a diversion distance of about 1,320 ft. Second, a single-stage left-turn exit (where medians are too narrow to safely store two or more vehicles) will involve the following delays (not including acceleration times): 19

PAGE 30

20 Table 2.1 Left-turn Delay Under Different Volume Conditions Volumes(vph) Artery Left-Turn Delay per Vehicle (Two directions) Exit (Seconds) 1,000 50 20 1,000 100 25 2,000 50 200 2,000 100 530 These values suggest that when arterial traffic exceeds 375 to 500 vphpl on a four-lane facility the computed delays would exceed those a ssociated with the right turn/U-turn movement. Higher volumes (700-900 vphpl) that are common along many suburban arterials would produce ev en higher left-turn egress delays in theory. In practice, motorists become impatient when gaps exceed 1 to 2 min and are apt to avoid the direct left turn egress. Third, the two-stage left turn process, where medians can safely store waiting vehicles, reduces delays to left-turning traffic. Nevertheless, this process still results in long delays to left-turning vehi cles when the volumes on the major street are relatively high (i.e., more than 2,000 vph), and the left turns exceeds 50 per hour. In these cases, even with substantial circuity (1,320 ft or 402m from the access drive to the U-turn median opening, or a 0.5 mi of additional travel) the right turn followed by a U-turn involves less time than calculated left-turn egress movements under moderate to high volumes. The project sponsored by FDOT in 2001 (Met hodology to Quantify the Effects of Access Management Treatments on Roadway Operations and Safety) provided very useful information on the access management treatment on Right-Turn followed by U-turn at Median Opening as an Alternative to the Direct Left Turn from Driveways and

PAGE 31

21 Side Streets The study took three basic approaches in evaluating the issue including operational evaluation, conflict data analysis, and crash data analysis. This project involved huge amount of field da ta collection where the extens ive data were gathered at several appropriate locations using video cameras followe d by lengthy data reduction process in the lab. Delay and travel time mode ls were developed using collected data to quantify the relationship between delay and trav el time to explanatory variables. Basic conclusions got from this project including: 1) The curves based on delay and travel time models indi cated that under high major road and driveway volume conditions, vehicles making a direct left turn experienced longer delay and travel times than those th at made a right turn followed by a U-turn. 2) Directional median openings may provide more efficient traffic flow than full median openings when the major-road through-traffic flow rate is more than 4,000 vph in both directions and the left-t urn-in flow rate from the major-road is over 150 vph ; 3) There are no significant impacts on thr ough traffic speed by either movement because these two movements have no impact on the platoon speed, they only affect the speed of random arri vals between platoons; 4) The percentage of RTUT movements increases with major-road through-traffic flow rate and left-turn-in flow rate from major-road; 5) The average running time of a vehicle maki ng a RTUT from a dr iveway has a linear relationship with the length of weaving se gment or the running time increases as the weaving distance gets longer; 6) The average weaving speed of RTUT linearly increases with the increase of weaving distance; and

PAGE 32

7) The before and after study indicated that there was about 15-22% less delay for the drivers turning left from a driveway after the median opening was replaced with a directional median opening, forcing them to make a RTUT at a median opening 420 feet downstream, in place of a DLT. 2.5 U-turn at Signalized Intersections As mentioned previously, the 2001 USF research was focused on providing right turn followed by U-turn at median opening. In the real world, however, there are many other conditions where U-turn is provided at downstream signalized intersections. Considering these two different U-turn treatments, each state has different policies. For example, in Wisconsin, U-turns are not legal at signalized intersections. U-turn movements are provided at pre-U-turn openings near signalized intersections. Michigan uses U-turn channels on highways with wide medians and prohibits all turning turns at signalized intersections. U-turn lanes can be provided downstream of signalized intersection. It is also called Michigan U. Figure 2.3 Michigan U 22 In Florida Median Handbook, Sokolow mentioned that there are three different U-turn approaches including U-turn at signalized intersection, U-turn in advance of a signal and U-turn after signal. He also indicated that a U-turn in advance of a signalized intersection will result in two successive left-turn lanes and unless there is a substantial

PAGE 33

23 length of full median width, drivers may mist akenly enter the U-turn lane. It was still recommended by the handbook that where medians are of sufficient width to accommodate dual left-turn lanes, U-turn can be provided from the inside left turn lane at signalized intersections. For this specific condition, Florida Median Handbook mentioned about three issues need to be considered: 1) Consider righton-red restrictions for side streets. 2) Remember to look at signal operation. 3) Dont let the signalization intersection work against U-turns. NCHRP 420 analyzed three different U-turn approaches including providing U-turn lanes in advance of, at, or beyond signalized intersec tions. As indicated in this report: 1) Left-turn lanes can be provide d for U-turning vehicles in advance (i.e., upstream) of signalized intersections. This avoids c oncentrating development-related turning traffic at signalized junctions of major crossroads. 2) Dual left-turn lanes can be provided at signalized intersec tions with the inner lane dedicated to U-turns. Many states now provide these la nes; however, they still require multiphase traffic signal controls. 3) Leftand U-turn lanes can be provided downstream of signalized intersection, thereby allowing two-phase traffic signal controls. One of the major concerns of U-turn at signalized intersections is that U-turn movements may reduce signalized intersection capacity. A study conducted by Webster and Cobbe in 1966 recommended the follo wing relationship between radius and saturation flow rate for the exclusive left-turn movements:

PAGE 34

R s92.412080 (2-6) Where: s = saturation flow rate for exclusive left-turn movement (vphpl); and R = radius of curvature (ft). The equation shows that the saturation flow rate increases with increasing turning radius. Since U-turns usually have smaller turning radius than left turn movements, it is anticipated that U-turn may have lower turning speed than that of left-turn movement. Therefore U-turn may have some adverse impacts on left turn lane capacity. Two previous researches have substantiated this assumption. A study conducted at North Carolina State University in 1993 evaluated the U-turns effects on left-turn saturation flow rates. The study team selected four intersections with exclusive left-turn lanes and protected signal phasing and recorded saturation flow rates and U-turn percentages for 198 queues during weekday midday peaks. The data analysis showed that a saturation flow reduction factor appears necessary for left-turn lanes that have large percentages of U-turns.Saturation flow rates were significantly lower when queues had more than 65% U-turns. However, the analyses also showed no correlation between saturation flow and the percentage of U-turns for queues with 50% or fewer U-turns. The results of this study suggest tentative saturation flow reduction factors of 1.0 for U-turn percentages below 65, 0.90 for U-turn percentages between 65 and 85, and 0.80 for U-turn percentages exceeding 85. A follow-up investigation should focus on intersections that have high percentages of U-turns, restrictive geometry, or high percentages of U-turning heavy vehicles. 24

PAGE 35

25 Tsao and Chu (1995) recorded 600 headways of left-turning passenger cars and 160 headways of U-turning passenger cars in Taiwan.Their study revealed that the average headways of U-turning passenger cars are significantly larger than those of left turning passenger cars. The effects of Uturning vehicles depend upon the percent of U-turning vehicles in the left-turn lane, as well as the order of formation in the traffic stream. When preceded by a left-turning vehi cle, the average headway of U-turning passenger car is 1.27 times that of left-t urning passenger cars. When preceded by a U-turning vehicle, however, the average head way of U-turning passenger cars is 2.17 times that of left-turning passenger cars. C onsidering U-turn adjustment factors for varying percents of U-turning vehicles in left-turn lanes, this study got a little bit different conclusion from the North Carolina research, which is listed in the following table: Table 2.2 Adjustment Factor for U-turns Percent of U-turn 0 2 4 6 10 15 20 25 30 Average Value 1 0.98 0.97 0.95 0.92 0.89 0.86 0.84 0.81 Conclusions from these two studies can not be directly used in this project. The reasons are: 1) Both of these two studies assume that the di scharge flow rate of the vehicle reaches saturation state after fourth or fifth di scharged vehicle. The field measurement, however, shows that when there are U-turn ing vehicles in the left-turn lane, the discharge flow rate do not display an easily identifiable steady maximum rate. 2) Because of the variations in drivers beha vior in different areas, conclusions from these two projects may not be hold in Florida.

PAGE 36

26 Based on our literature review, generally there is no widely accepted procedure for estimating U-turn effects on signalized inte rsection. In HCM, Uturns are treated as left-turns when estimating saturation flow rate In practice, however, when a full median opening is replaced by a directiona l median opening, the direct left turns must be diverted to make a right turn followed by a U-turn at downstream signali zed intersection. The increased U-turning movements at signal may further degrade the signalized intersection and maybe the effects should not be ignored. Sometimes when there is a large combined left-turn plus U-turn volume at signalized in tersection, the installation of dual left-turn lane could be considered and U-turn can be provided from the in side left-turn lane. Similar study hasnt been conducted in Florid a before and U-turns ef fects on signalized intersection capacity ar e still not clear. 2.6 Summary After an extensive literature search, whic h included current ru les and regulations, design standards and policies in Florida and nationally, and a computer search of the Transportation Research Information Service (T RIS) database, conclusions can be made that little documentation is available on operational effects of providing right turns followed by U-turns at signalized intersections as an alternative to direct left turns. It is necessary to conduct an extensive study on this specific access management treatment since it has been widely used in Florida. Other findings including: 1) There is no widely accepted procedure for estimating U-turning vehicles effects on signalized intersection capacity. Related st udy hasnt been conducted in Florida before.

PAGE 37

27 2) Vehicles delay at signalized intersect ion is largely affected by signalization conditions, including g/c ratio, cycl e length, and demand flow rate. 3) Providing U-turn movements in advance of or at downstream signalized intersection is a controversial topic. However, little documentation is avai lable concerning the comparison of these two U-turn approaches.

PAGE 38

28 CHAPTER 3 METHODOLOGY Methodologies that were used in studyi ng the operational eff ects of right-turn followed by U-turns at signalized intersection as an alternative to direct left turns are explained in this chapter. This chapter consists of three sections. The fist section explains the methodology used in specification of dela y and travel time m odels. The second part deals with drivers selection of RTUT or DL T on the basis of accessibility considerations. The third part of this chapter discusses U-turn effects on signalized intersection capacity. 3.1 Delay and Travel Time Models Vehicle delay and travel time are very im portant parameters used by transportation professionals to evaluate the performance of intersections. This importance of vehicle delay and travel time is reflected in the us e of these parameters in both design and evaluation practices. In additi on, in the real world, one of the major public concern is people often believe that U-turn has much longe r delay and travel time as compared with direct left turn. This is why some people dont like to make a U-turn. One of the major objectives of this projec t is to develop delay and travel time models of RTUT and DLT. In order to addre ss public concerns, it is necessary to compare delay and travel time of RTUT versus DL T under specific volume and roadway geometric conditions. In addition, from the decision maker s point of view, it is also necessary to quantify the relationship of delay and travel time to possible explanatory variables,

PAGE 39

including conflicting traffic flow rates, signalized intersection characteristics, and roadway geometric characteristics. 3.1.1 Operations Analysis of Direct Left Turns The Highway Capacity Manual 2000 identified the priority of right-of-way given to each traffic stream at unsignalized intersection. Based on the definition, DLT egress from a driveway or minor street has the lowest priority. Theoretically, DLT egress must therefore yield to all other movements at unsignalized intersections. Thus, it is the most likely movement to be delayed. However, in the real world, when left-turn drivers wait for longer periods, they become more aggressive and enter the median opening without yielding to other maneuvers, such as left-turn-in vehicles from the major road. On the arterials with wide medians, which can allow one or two vehicles to stop, a DLT maneuver may require four steps, as shown in Figure 3.1 and as explained as follows. Figure 3.1 DLT Egress Movements Step 1 Stopping and waiting at the driveways, Step 2 Selecting a suitable gap, accelerating across major-road through-traffic lanes and coming to a stop in the median. Sometimes, drivers can cross the median without stopping at the median openings if there is a suitable gap in both directions, 29

PAGE 40

30 Step 3 Stopping at the median, and waiting for a suitable gap from right-side through-traffic. Some drivers only need to se lect a suitable gap for the inside lane, accelerate and merge into through traffic, wh ereas some others need at least two clear lanes. Sometimes when several left-t urn vehicles stop para llel at the median opening, the vehicles stopped at the right side may block visibility for other drivers. This may result in crashes between left-turning vehicles and through traffic; and Step 4 Accelerating to operating speed on the major roadway. This may force through traffic to decelerate or make a lane cha nge when the left-turning drivers select a small gap. Based on the operations analysis of a DLT movement, the aver age delay and total travel time of DLT can be defined by the following equations: TT L = t L1 + t L2 + t L3 (3-1) TD L = t L1 + t L2 (3-2) where, TT L = average total travel time of DLT movements, TD L = average total waiting delay of DLT movements, t L1 = average waiting delay of DLT vehicles at the driveway, t L2 = average waiting delay of DLT vehi cles at the median opening, and t L3 = average running time for vehicles leaving th e driveway till completing the left turn movement (not including t L1 and t L2 ). From the above equations, the average tota l delay of DLT is the sum of average waiting delay of left turns at a driveway and the average waiting delay at a median opening.

PAGE 41

The average total travel time of DLT is equal to the average total delay plus the average running time for vehicles from the time they leave the driveway to when they stop at the median opening (t L3 ). 3.1.2 Operations Analysis of Right Turn plus U-turns In order to eliminate problems associated with DLT movements, many states and transportation agencies have started installing restrictive medians and directional median openings in the state highway system. Left turn egress movements would be replaced by turning right onto the arterial road and then making U-turns at downstream median opening or signalized intersection. As shown in Figure 3.2, a vehicle making a RTUT at downstream signalized intersection also requires four steps. Figure 3.2 RTUT Movements Step 1 Stopping at the driveway, and making a right turn when there is a suitable gap from left-side through-traffic. This is much easier than left-turn egress because it does not need to yield to other movements at the unsignalized intersection at the same time. So, usually when the upstream signal for the major-road through-traffic turns red, there is a large gap created for right turns. There is a potential conflict between a right turn from a driveway and a U-turn at the median opening. Drivers can easily 31

PAGE 42

32 overlook this conflict, which can result in an accident when their attention is focused on the major-road through traffic; Step 2 Accelerate, weave to the inside lane, and decelerate to a stop at the exclusive left turn lane of downstream signalized in tersection. This movement will cause conflicts such as decelerati on and lane change of through traffic. There may also be speed reduction of through traffic in the weaving section. Sometimes when left turn lane is not long enough, U-turn vehicles may be blocked by through traffic already queued at the traffic signal. It will cause extr a delay to RTUT vehicle. Step 3 Waiting until the signal turns green to ma ke a U-turn. Delay in this step is decided by signalization conditions and demand flow rate. Sometimes it takes long delay in this step especially when the signal has long cycle length and large left turn demand. As mentioned previously, U-turn movement may have longer discharge headway and start up lost time as compared with left turn movement. Therefore U-turn movements may have some adverse impacts of left turn capacity. In this step, if U-turn is made during protecte d signal phase, drivers dont need to take care of the right-side through-traffic. Therefore there is no conflict between U-turn vehicles and through-traffic. For the condition when U-turn is made during permitted signal phase, the drivers need to find a suitable gap from right side and then make a U-turn. This condition is very similar to where U-turn is provided at median opening. Basically this condition will not be considered in this project. Attention should be paid that there is a potential conflict he re between U-turn vehicles and right-on-red vehicles in the other approach of the road. Both U-turn

PAGE 43

33 drivers and right-turn drivers can easily overlook this conflict, which will result in accident. Step 4 Accelerate to the operating speed of through traffic. As compared with DLT movement, this step will not result in speed reduction in through traffic if U-turn is made during protected signal phase. Accordingly, to estimate total travel tim e for vehicles making RTUT movements, the following equations can be used: TT RU = t RU1 + t RU2 + t RU3 + t RU4 (3-3) TD RU = t RU1 + t RU2 (3-4) t RU4 = 0.68 (l/v T ) (3-5) where, TT RU = average total travel time of RTUT movements (seconds), TD RU = average total waiting delay of RTUT movements (seconds), t RU1 = average waiting delay of right turn vehicles at the driveway (seconds), t RU2 = average waiting delay of U-turn vehicles at the exclusive left turn lane of downstream signalized intersection (seconds), t RU3 = average running time from leaving the driv eway to stopping at the exclusive left turn lane (not including t R1 and t R2 ) (seconds), t RU4 = average running time of vehicles crossing the whole roadway sect ion at the posted speed of through-traffic (seconds), l 1 = weaving distance from the studied drivew ay to the median U-turn opening (ft.), l = the distance form the studied driveway to the U-turn bay, includ ing weaving distance and the left turn storage bay (ft), l=l 1 +l 2

PAGE 44

Weaving Distance (l 1 ) Left Turn Storage Bay (l 2 ) Figure 3.3 The Distance from Driveway to Signalized Intersection The average total waiting delay of RTUT vehicles includes the delay of right turns at the subject driveway (t RU1 ) and the delay of U-turns at signalized intersection (t RU2 ). The average total travel time of a RTUT movement is the sum of average total waiting delay, the average running time in the weaving section, and the average running time needed for a vehicle traversing the length of the whole roadway segment (weaving section plus exclusive left turn lane) at the operating speed of through-traffic. The average total delay and travel time were used to quantify operational effects of RTUT vs. DLT. 3.2 Driver Selection of RTUT Drivers selection of RTUT or DLT may be affected by different roadway traffic and geometric characteristics. Conducting a study regarding drivers selection of RTUT can provide decision makers with useful information to select a suitable median treatment, including median opening closure and replacing a full median opening with a directional one, etc. 34

PAGE 45

3.2.1 Driver Selection of RTUT at Signalized Intersection The project conducted by University of South Florida in year 2000 developed a linear regression model to describe the relationship between the percentage of drivers choosing RTUT at median opening and the combination of left-turn-in flow rate and major-road through-traffic flow rate. The regression results are given in the following table: Table 3.1 Regression Results for Ratio of RTUT at Median Opening N R-Square Intercept TV LTIN SPLIT Coefficients -1.48 0.0002 0.004 -2.19 105 0.36 tstatistics -2.95 3.89 4.83 -2.94 SPLITTVLTINeRatio1.20002.0004.023.0 (3-6) where, Ratio = percentage of RTUT at fifteen-minute intervals, LTIN = left-turn-in flow rate from the major-road (vph), TV = flow rate of major-road through-traffic (vph), and SPLIT= percentage of upstream through-traffic flow rate. The model shows that the percentage of RTUT movements increases with major-road through-traffic flow rate and left-turn-in flow rate from major road. In practice, when there is a suitable U-turn median opening downstream, some drivers prefer to make a RTUT rather than a DLT when the median storage space has been occupied by conflicting vehicles. This decision is encouraged when there are a large number of left-turn-in vehicles from the major road. 35

PAGE 46

36 When U-turn movements are accomm odated at downstream signalized intersection, however, driver s choice behavior may be largely different from the condition where U-turn is provided at median opening. As indicated earlier, a driver making a U-turn at signalized intersection doesnt need to wait and find a suitable gap form right-side through-traffic since U-turn is made at protected signal phase. Some drivers may prefer this option wi th the perception that it is safer. However, there are also some drivers dont like making a U-turn at signa lized intersection since they think it takes long time to wait for the signal to turn green. Another factor that may affect drivers choice behavior is the distance from driveway to downstream signalized intersec tion. In the field, peopl e generally dont like to make a right turn followed by a U-turn at downstream signalized intersection when the intersection is located out of drivers sight distance. When the distance is too short, drivers need to select a suitable simultaneous gap in al l through lanes and then make a direct entry into the inside lane, wait until the signal turns green to make a U-turn. Sometimes it is very difficult to finish this procedure especially when through volume is very heavy and the driveway is blocked by through traffic already queued at the traffic signal. In these conditions, some drivers may make a right turn on to the major arterial road, cross the intersection, and then make a U-turn at a median opening downstream of the signalized intersection. To decide a suit able distance from driveway to downstream intersection or median opening is a challenging topic since th ere are a lot of different factors need to be considered. More details considering this topic will be included in another project.

PAGE 47

3.2.2 Fitting a Binary Logistic Regression Model Logistic regression is a technique for analyzing problems in which there are one or more independent variables which determine an outcome that is measured with a dichotomous variable in which there are only two possible outcomes. In the case of binary logistic regression models, the relationship between a binary response variable and one or more explanatory variables are modeled. For a binary response variable y, the linear logistic regression model has the form, XpppLogitiii)1log()( (3-7) where, ip = Prob. (y i = y 1 /X i ) is the response probability to be modeled, and y 1 is the first ordered level of y, = the intercept parameter, = the vector of slope parameters, and Xi = the vector of explanatory variables. This logistic regression equation models the logit transformation of the i th individuals event probability, p i as a linear function of the explanatory variables. Logistic regression was widely used to estimate bounded fractional dependent variables. As compared with OLS regression, the predicted value from a logistic regression can be guaranteed to lie in the unit interval ( 10 ip ). It is natural to model its population regression as a linear function since )1/(logiipp can take on any real values as varies between 0 and 1. In this study, a binary logistic regression model was developed to estimate the percentage of drivers selecting RTUT at downstream signalized intersection ip 37

PAGE 48

38 as an alternative to direct left turns from driveway under certain roadway traffic and geometric conditions. 3.3 U-turn Effects on Signalized Intersection In some cases, when a full median opening is replaced by a directional median opening, the direct left turns must be diverted to make a right turn followed by a U-turn at downstream signalized intersection. As indicated before, the increased U-turn movements at signalized intersection may reduce the inte rsection capacity. This problem may become cute when there is large percentage of Uturning vehicles at le ft-turn lane and the medians are not wide enough to accommodate dual left-turn lanes. In this condition, a U-turn prohibition could be considered at signalized intersection and U-turn movements could be relocated in adva nce of or after signal. Saturation flow rate is one of the most critical factors in estimating capacity of a lane or lane group at signalized intersections In this study, U-turn effects on signalized intersection capacity are estimated by developing a U-turn ad justment factor on left-turn saturation flow rate. 3.3.1 Procedures for Estimating Saturation Flow Rate As indicated in HCM 2000, saturation flow ra te is the equivalent hourly rate at which previously queued vehicles can traver se an intersection approach under prevailing conditions, assuming that the green signal is available at all times and no lost time are experienced. Based on this definition, the Saturation flow rate is the maximum flow rate that can pass through a given lane group un der prevailing conditions. In estimating saturation flow rate, different adjustment f actors are applied to address the impacts of prevailing conditions, including lane width and la teral clearance, number of lanes, heavy

PAGE 49

vehicles and grades, turning movements, interchange density, lane distribution, and environmental factors. A saturation flow rate for each lane group can be estimated according to the following equation: RpbLpbRTLTLUabbpgNVwffffffffffNfss0 (3-8) where, s = saturation flow rate for subject lane group, expressed as a total for all lanes in lane group (veh/h); 0s = base saturation flow rate per lane (pc/h/ln); N = number of lanes in lane group; wf = adjustment factor for lane width; HVf = adjustment factor for heavy vehicles in traffic stream; gf = adjustment factor for approach grade; pf = adjustment factor for existence of a parking lane and parking activity adjacent to lane group; bbf = adjustment factor for blocking effect of local buses stop within intersection area; af = adjustment factor for area type; LUf = adjustment factor for lane utilization; LTf = adjustment factor for left turns in lane group; RTf = adjustment factor for right turns in lane group; Lpbf = pedestrian adjustment factor for left-turn movements; and Rpbf = pedestrian-bicycle adjustment factor for right-turn movements. 39

PAGE 50

As an alternative to the estimation of saturation flow rate using Equation 3-10, saturation flow rate for each lane group can also be estimated by field measurement. As indicated in HCM, the measured values of prevailing saturation flow rate in the field will produce more accurate results. Discharge headway research is widely used in the field measurement of saturation flow rate at a signalized intersection. In practice, when the green signal is initiated, headways between departing vehicles will be observed as vehicles cross the stop line. The first headway will be the time between the initiation of the green signal and the crossing of the first vehicle over the stop line. The second headway is the time between the first and second vehicles crossing the stop line. Any reference points can be used when recording headways, as long as the identical point is maintained through measurement. Common practice is to measure the headways as the rear wheels of the reference vehicle cross the curb line. The study uses the rear wheel as the reference point in field measurement. Most of the previous studies have indicated that the discharge headway converges to a constant headway, which is usually achieved after the fourth to sixth discharged passenger car crossing the stop line after the beginning of green. The constant headway is defined as the saturation headway. The relationship between saturation flow rate and saturation headway is shown in the following equation: hs3600 (3-9) where, s = saturation flow rate (vphpl); h = saturation headway (sec); 3600 = seconds/hour. 40

PAGE 51

3.3.2 U-turn Effects on Left-turn Saturation Flow Rate In order to develop a procedure for estimating U-turn effects on signalized intersection, a pilot survey was conducted at the early stage of this project. An intersection with exclusive left-turn lane and protected signal phasing was selected for this survey. The intersection is located on Fowler Avenue in Tampa, which is a six-lane principle arterial road. The signal is actuated controlled with an average cycle length of 149 sec. The study team recorded discharge headways for 138 left-turning vehicles and 54 U-turning vehicles in 27 discharging queues during weekday peak hour. To focus on the characteristics of passenger-car flows, the data related to heavy vehicles and all vehicles behind a heavy vehicle are excluded from analysis. The queue discharge patterns for queues with different percentage of U-turning vehicles are shown in the following figure: Figure 3.4 Comparison of Queue Discharge Patterns for Different Percentage of U-turning Vehicles in Left-turn Lane 41

PAGE 52

42 Past studies generally indicated that when a vehicle queue is released by a traffic signal turning green, the discha rge flow rate of the vehicl es quickly reaches a steady state. The collected data matches this c onclusion. As shown in figure 3-2, the average discharge headways for left-tur ning vehicles converge to a relatively constant state from forth or fifth discharged vehicle after green onset. For the situations in which there are U-turning plus left-turning vehicles in the di scharging flow, however, the queue discharge patterns do not display an easily identifia ble steady maximum rate Field measurement found that it is very difficult to get satu ration headway in the field measurement when there are U-turning vehicles in the discharging flow. In addition, the figure shows that the average discharge headway increases with the percentage of U-turning vehicles in discharging queue. This is because of the di fferent turning characteristics of these two movements. As mentioned earlier, U-turn move ment has shorter turning radius than left turn movement. Consequently it has lower turning speed as compared with left turn movement. Field observation found that there is a conflict point between U-turning vehicle and the preceding left-turning vehicles In the real world, when a vehicle is making a U-turn at signalized intersection, sometimes the preceding left turn vehicles have to make a break in or der to avoid a rear end coll ision. In this condition, the saturation state is broken by the U-turning vehicle. Preceding vehicles may need to accelerate again before reach ing a saturation state. In HCM, U-turns are treated as left-turns in the current procedure for estimating saturation flow rates at signalized intersec tions while the operati ng characteristics of these two movements are different. Two pa st studies and the pilot survey have substantiated the assumption that the average discharge headway of U-turning vehicles is

PAGE 53

43 larger than that of left-turning vehicles. Moreover, U-tu rning vehicles cause greater effects to their succeeding ve hicles than left-turning vehicles. Therefore conclusion can be made that U-turn movements reduce left turn saturation flow rate especially when there is a large proportion of U-turning vehicl es in left-turn lane. It was recommended in this project to apply an adjustment factor for U-turning movements when estimating left turn saturation flow rate. 3.3.3 Adjustment Factor for U-turn Movement Similar to other turning movement adju stment factors, such as right-turn adjustment factor and left-turn adjustment factor, the U-turn adjustment factor also depends on a number of variables, including: 1) Whether U-turns are made from exclusiv e left turn lanes or shared lanes; 2) Type of phasing (protected, permitted, or protected-plus-permitted); 3) Proportion of U-turning vehicles in the left turn lane. In this project, only the condition in which U-turn movements being accommodated at exclusive left turn lane w ith protected signal phasing was considered. The study conducted by Tsao and Chu in 1996 use discharge headway research to estimate U-turns effects on the traffic flow in left-turn lanes. The study recorded discharge headway for 600 left-turning vehicl es and 160 U-turning vehicles in left-turn lanes. This research assumed that discharge flow rate of the vehicle reaches saturation state after fourth or fifth discharged vehi cle, and only the head ways after the fifth discharged vehicle were recorded. In this st udy, U-turn adjustment factor was estimated by the following equations:

PAGE 54

)(21minmaxhhhhfLLLLUT (3-10) UULLhahaah)100(1001)(max (3-11) ULLULLhahahaah10010010021)(min (3-12) where, f UT = the adjustment factor for U-turns; h min (a) = the lower limit of average headway with a% of U-turning vehicles; h max (a) = the upper limit of average headway with a% of U-turning vehicles; h LL = the average headway between two successive left-turning vehicles (sec); h UU = the average headway between two successive U-turning vehicles (sec); and a = percentage of U-turning vehicles. As mentioned earlier, this equation can not be directly used in our project because the past study assumes that the discharge flow rate of the vehicle reaches saturation state after fourth or fifth discharged vehicle. The field measurement, however, indicates that when there are U-turning vehicles in left-turn lane, the discharge flow rate do not display an easily identifiable steady maximum rate. Field observation found that the number of U-turning vehicles in the first four discharged vehicles will affect the discharge headways of preceding vehicles. In the field, when there are U-turning vehicles in the first four discharged vehicles, preceding left turn vehicles can not fully speed up to reach the maximum saturation state. Past study didnt consider this condition since only the headways after the fifth discharged vehicle were recorded. This previous study was conducted in Taiwan. Considering the variations in drivers behavior in different areas, the conclusion may not be hold in Florida. 44

PAGE 55

In this project, the determination of U-turn adjustment factor is achieved by analyzing the relationship between percentage of U-turning vehicles in the left turn lane and the average queuing discharge time of the whole discharging flow, including the first four discharged vehicles. As indicated before, U-turn movement has larger average discharge headway than left-turn movement. In addition, a U-turn movement will force preceding left-turning vehicles slow down to avoid a rear end collision. Therefore, when there are U-turning vehicles in exclusive left turn lane, the discharging queue will consume more green time than the queue will only left-turning vehicles. Theoretically, the difference increases with the percentage of U-turning vehicles in the discharging flow. To analyze the relationship between the percentage of U-turn vehicles in the left turn lane and the average queuing discharge time, a linear regression model was specified using field data from pilot survey. The relationship is illustrated in the following equation: Figure 3.5 The Relationship between Average Discharge Time & The Percentage of U-turning Vehicles 45

PAGE 56

UTPehh0017.00 (3-13) where, h = average queuing discharge time for U-turn and left-turn mix flow (sec); h 0 = base average queuing discharge time for left-turn only flow (sec); and P UT = percentage of U-turn vehicles from inside left-turn lane (%). The pilot survey shows that, when there is no U-turning vehicle in left-turn lane, the average queuing discharge time for each left turning vehicle is 2.11 sec. The average queuing discharge time increase with the percentage of U-turning vehicles. The relationship is illustrated in equation 3-14. Based on the definition of the adjustment factors for turning movements, the U-turn adjustment factor for the left-turn saturation flow rate can be estimated by the following equation: UTUTPPUTeehhhhf0017.00017.00011109.21109.236003600 (3-14) where, f UT = adjustment factor for U-turn movement; h = average queuing discharge time for U-turn and left-turn mix flow; h 0 = base average queuing discharge time for left-turn only flow (sec); and P UT = percentage of U-turn vehicles from inside left-turn lane (%). The equation 3-14 can not be directly used in estimating adjustment factor for U-turns because the sample size is limited and the R-square value is low. However, the procedure for estimating U-turn adjustment factors for different percentage of U-turning vehicles in left-turn lane is used in further study. 46

PAGE 57

47 CHAPTER 4 DATA COLLECTION AND REDUCTION This study consists of huge amount of field data collection work. During June 2002 to July 2003, field measurement was conducted on eight urban or suburban arterial street segments in Tampa Bay area, where extensive data were collected using video cameras. A total of more than 300 hours of field data was collected. This chapter discusses the detailed efforts of data colle ction and data reduction work. The major objective of this project is to quantify the operationa l effects of right turn followed by U-turn at downstream signalized intersection as an alternative to direct left turn. The data needed to achieve this objective are listed as follows: 1) Traffic volume: major-road through-traf fic volume, left-turn-in volume from major-road, left-turn-out volume from driv eway and side street and right turn followed by U-turn volume; 2) Traffic delay: delay of left turns and right turns at the s ubject driveway, delay of left turns at median openings, and delay of U-turns at signali zed intersection; 3) Traffic running time: average running time of RTUT crossing the weaving segment, and average running time of DL T crossing the through lanes;

PAGE 58

48 4) Signal parameters: green arrow time, cy cle length, queue discharge time, queue discharge headways for left-turning and U-turning vehicles, and leftturn volume from inside left turn lane; 5) Geometric data: cross section, lane assignments, weaving di stance, length of left-turn storage bay, and median type; and 6) Traffic control features: speed limit, traf fic control signs and traffic signals. 4.1 Site Selection Site selection work was conducted during June 2002 to November 2002. The major purpose of site selection is to find compatible site with high RTUT and DLT volumes. More specifically, the geometric crit eria of selecting specific sites include: 1) The arterial should have a raised-curb median with either a full median opening or a directional median opening that ca n safely store waiting vehicles; 2) The arterial should have 6 or 8 through traf fic lanes (3 or 4 lanes each direction). Passenger cars can normally make U-turn s along a divided six-lane arterial; 3) Speed limit on the arterial shou ld be 40 mph or higher. The FDOT mandates that all new multi-lane projects with design speeds of 40 mph or greater be designed with a restrictive median; 4) The studied driveway should ha ve either two lanes (one for right-turn and another for the left-turn) or one wide lane with a flared curb so that the two movements do not interfere with each other; 5) The driveway volumes should be high so th at there were a considerable number of RTUT and/or DLT vehicles; 6) The median width should be wide enough to store the left-turning vehicles; and

PAGE 59

49 7) The downstream signal should have exclusive le ft turn lane and protected left turn phasing in the studied approach. The condi tion in which U-turn movements being accommodated at permitted left turn phase is not considered in this study. Based on these criteria, eight sites located in Tampa Bay area were selected for field measurement. The selected sites are listed in table 4.1. Table 4.1 Description of Selected Sites Note: N1: # of through lanes; N2: # of ex clusive left turn lanes at signalized intersection; D: directional me dian opening; F: Full median ope ning; l: the distance from driveway to downstream signalized intersection, including weaving distance and left-turn storage bay; and g/c: green cycl e ratio. For actuated controlled signal, the g/c ratio here is defined as the maximum green arrow time for left turn phase divided by the average cycle length of the signalized intersection.

PAGE 60

Site 1 is located in the city of Tampa, at Fowler Avenue and 56 th Street. Fowler Avenue is a principle arterial road with three lanes in each direction. The studied driveway is located on Fowler Avenue and it serves a shopping plaza with a Publix Supermarket and many small businesses. The median opening across the driveway is a directional median opening, which restricts left turn egress movements from driveway. The speed limit of the selected road segment is 50 mph. Figure 4.1 Site 1 Fowler Avenue and 56 nh Street Site 2 is located in the city of Tampa, at Fowler Avenue and 22 nd Street. Fowler Avenue is a principle arterial road with four lanes in each direction. The studied driveway is located on Fowler Avenue and it is one of the driveways that serve the University Mall. The driveway has two lanes for egress of vehicles with one lane dedicated to DLT vehicles and the other one dedicated to right turning vehicles. The median opening across the driveway is a full median opening which allows almost all turning movements at un-signalized intersection. The speed limit of selected segment is 45mph. 50

PAGE 61

Figure 4.2 Site 2 Fowler Avenue and 22 nd Street Site 3 is located in the city of Tampa, at Hillsborough Avenue and Webb Avenue. Hillsborough Avenue is a principle arterial road, which has three lanes in each direction. The studied driveway is located on Hillsborough Avenue. The driveway serves a parking lot for a plaza includes a major bank and some small businesses. The median opening across the driveway is a full median opening. The speed limit of this road is 45mph. Figure 4.3 Site 3 Hillsborough Avenue and Webb Street Site 4 is located in the city of Tampa at Dale Mabry Highway and North Dale Street. Dale Mabry Highway is a major highway divided by a raised median. This 51

PAGE 62

highway has three lanes for each direction. The studied driveway is on Dale Mabry Highway and it is one of the driveways that serve a major shopping plaza that includes many small businesses and retail stores. The median opening across the driveway is a directional median opening, which restricts left turn egress movements from the driveway. The speed limit is 45 mph at selected segment. Figure 4.4 Site 4 Dale Mabry Highway and North Dale Street Site 5 is located in the city of Tampa at Bruce B. Downs Boulevard and Fletcher Avenue. Bruce B. Downs Boulevard is a major arterial road with three lanes in each direction. The driveway is one of the driveways that serve Target Plaza that consists of; Target, Eckerd and U Save supermarkets, fast food restaurants and many small businesses. The driveway is on Bruce B. Downs Boulevard and has two separate lanes for DLT and right-turn movements. There is a full median opening located across the driveway. The speed limit in this segment is 45 mph. 52

PAGE 63

Figure 4.5 Site 5 Bruce B. Downs Boulevard and Fletcher Avenue Site 6 is located in the city of Saint Petersburg at 34 th Street and 54 th Street. 34 th Street is major arterial with three lanes for northbound and southbound traffic. The driveway that was studied is one of the driveways that serve a major shopping plaza consists of a Publix Supermarket, some retail stores and many small businesses. The median across the driveway has a full median opening. The posted speed limit here is 45 mph. Figure 4.6 Site 6 34 th Street and 54 th Street 53

PAGE 64

Site 7 is located in the city of Saint Petersburg at 34 th Street and 22 nd N. Street. 34 th Street is major arterial with three lanes for northbound and southbound traffic. It is divided by a raised median. The driveway that was studied is one of the driveways that serve a major shopping plaza consists of a Kash N Karry Supermarket, some retail stores and many small businesses. The median across the driveway has a full median opening. The RTUT movements are completed with a U-turn at the 34 th Street and 22 nd N. Street signalized intersection. The posted speed is 45 mph. Figure 4.7 Site 7 54 th Street and 22 nd N. Street Site 8 is located in the city of Tampa at Dale Mabry Highway and Maple Dale Street. At this segment, Dale Mabry Highway is divided by a raised median with three lanes in each direction. The selected driveway is on Dale Mabry Highway and it serves the parking lot for Sams Club Retail Store. The median opening in this site is different from that of other sites. The median opening here permit left egress from driveway while left-turn-in movements from major road are prohibited. The speed limit in this segment is 45 mph. 54

PAGE 65

Figure 4.8 Site 8 Dale Mabry Highway and Maple Dale Street 4.2 Data Collection In this study, equipments used for data collection include 5 video cameras, VCRs, batteries, inverters, and TVs. In order to cover the whole right turn followed by U-turn procedure and the signalized intersection parameters, the two-story scaffoldings were installed in the field. Figure 4.9 shows that cameras were set up at the top of a 15-feet high scaffolding. The basic cameras locations in the field are shown in figure 4.12. 55

PAGE 66

Figure 4.9 Equipments Setup in the Field Figure 4.10 Equipments Setup in the Field 56

PAGE 67

Figure 4.11 Equipments Setup in the Field Figure 4.12 the Basic Camera Locations in the Field A typical data collection day generally starts at 7:00 in the morning. Before start recording, all video cameras were synchronized so that the data extracted from different videotapes can be matched. Data collection usually was conducted during weekday 7:00 57

PAGE 68

58 AM to 7:00PM. More than 30 hours data were collected in each site. Data were not collected during inclement whether or when there were unusual traffic conditions in the road. 4.3 Data Reduction The collected videotapes were reviewed in office. In this project, the reduction of field data is very hard and timing consum ing since there were more than 300 hours videotapes need to be reviewed. Each tape needs to be review for five to six times in order to get different kinds of data needed for further analysis. Each vehicle coming from the driveway making DLT or RTUT was track ed. Since all video cameras have already been synchronized in field, data collected by different video cameras can be matched. By reviewing videotapes, the follo wing information was recorded: 1) Waiting delay: waiting delay of DLT and RTUT vehicles at driveway; waiting delay of DLT vehicles at median opening; and waiting delay of RTUT vehicles at signalized intersection; 2) Travel time: the total travel time of DLT and RTUT vehicles; 3) Traffic volume: major-road through-traf fic volume, left-turn-in volume from major-road, left-turn-out volume from driv eway, and right turn followed by U-turn volume; and 4) Signal parameters: green arrow time, cy cle length, queue discharge time, queue discharge headways for left-turning and U-turning vehi cles, and left turn volume from inside left turn lane. Total delay of each vehicle at driveway is measured from a vehicle stops at the waiting queue until it exits the stop line. The definition of delay here consists of queue

PAGE 69

59 time and service time. This definition is a little bit different from the definition of average control delay in HCM, since vehicles decele ration and acceleration were not considered when estimating delay in this project. The wa iting delay of left-turns at a median opening was measured by recording the time from the vehicle stops at the median until it leaves the median. The waiting delay of U-turning vehicles at signalized intersection was recorded as the time from the vehicle stops at the inside left turn lane until it starts making a U-turn. By tracking each individual vehi cle, the total travel time of each DLT or RTUT vehicle can also be recorded. The reduction of field data is based on five-minute time interval. In each time interval, the average total delay and travel time for DLT and RTUT vehicles were recorded. In addition, traffic volume data, including major-road through-traffic volume, left-turn-in volume from major-road, left-t urn-out volume from driveway, right turn followed by U-turn volume, and left-turn volume from inside left-turn lane, were also measured based on this time interval.

PAGE 70

60 CHAPTER 5 OPERATIONAL EFFECTS 5.1 General In this project, the operational evaluati on of right turns fo llowed by U-turns at signalized intersection as alternatives to di rect left turns consis ts of four parts: 1) The comparison of the average delay of DLT and RTUT under specific volume and roadway geometric conditions. This obj ective was achieved by developing delay models for these two movements 2) The comparison of the average total travel time of DLT and RTUT under specific volume and roadway geometric conditions. Th is goal was realized by building travel time models for two movements; 3) The estimation of the percentage of driv ers selecting RTUT when both choices are available. A binary logistic regression model was developed to achieve this objective; 4) The estimation of U-turns effects on signa lized intersection capacity by applying the adjustment factor for U-turn movement s on left-turn saturation flow rate. In addition, in this chapter, the operations of two wide ly used U-turn approaches, U-turns at median opening in advance of signa lized intersection and U-turns at signalized intersection, were also comp ared based on the models developed in this study and those from 2001 project.

PAGE 71

61 Field data collected from selected sites were used to build these models. As mentioned before, the reducti on of field data was based on five-minute time interval. When specifying models, the original data at five-minute intervals were aggregated to fifteen-minute intervals because the data at fifteen-minute intervals were found to have better statistical characteristics. In this study, statistical analysis was performed by the use of SPSS software. 5.2 Average Delay Delay is an important MOE of traffic ope rations. The HCM use control delay as the criteria to evaluate LOS of signalized in tersections. Control delay was defined as the proportion of total delay attr ibuted to control measur es, which includes initial deceleration delay, queue move-up time, stop ped delay, and final acceleration delay. Delay defined in this study doesnt include vehicles deceleration and acceleration time; because it is very difficult to measure decel eration and acceleration in the field without sophisticated tracking equipment. In this study delay is defined as the total elapsed time from the time a vehicle stops at the end of th e queue to the time the vehicle departs from the stop line. The average delay of DLT move ments consists of delay at driveway and delay at median opening. The average delay of RTUT includes dela y at driveway and delay at signalized intersection. 5.2.1 Delay Model for Direct Left Turn Data collected from those sites with full median openings were used to build delay model for direct left turn move ments. This kind of sites include site 2, 3, 5, 6, and 7. As mentioned earlier in this chapter, the origin al data set at five-m inute intervals were aggregated to fifteen-minute intervals when specifying models.

PAGE 72

Figure 5.1 Traffic Flows Affecting the Delay of DLT Figure 5.1 illustrated the conflicting volumes affect the delay of DLT movement. Statistical analysis showed that both linear and exponential forms are suitable to describe the relationship between the average delay of DLT movement and conflicting volumes. However, the exponential form was found to have better theoretical and statistical characteristics. The delay model was described as Equation 5-1. 04321aLTINaDLTVaSPLITaTVaLeTD (5-1) Where, TD L = average total delay of DLT (sec/ veh), TV = flow rate of major-road through-traffic (vph), DLTV = flow rate of DLT from a driveway (vph), LTIN = flow rate of left-turn-in from major roads (vph), SPLIT = percentage of upstream through traffic flow rate, SPLIT=TV1/ (TV1+TV2), and a 0 a 1 a 2 a 3 a 4 = parameters Totally there are 464 observations at fifteen-minute intervals being used to estimate the delay model for DLT movement. The dependent variable (average total delay of DLT) refers to average total waiting delay per vehicle making a left turn during a 62

PAGE 73

63 fifteen-minute period. The independ ent variables, including left-turn-in flow rate, through traffic flow rate, and DLT flow rate, ar e equal to four time s traffic volume at fifteen-minute intervals. Multiple regression analysis was carried out to determine the best model by testing different independent variables. The st atistical characteristics of collected data are given in Table 5.1. The final regression results are listed in Table 5.2. Table 5.1 Descriptive Statistic s of the Collected Data Average Total Delay Flow Rate of TV SPLIT Flow Rate of DLT Flow Rate of LTIN N Valid 465 465 465 465 465 Missin g 0 0 0 0 0 Mean 38.1 3437 .5096 43.6229 48.2896 Median 36.4 3441 .5038 40.0000 44.0000 Mode 25.0 3200 .44(a) 36.00 36.00 Std. Deviation 16.1 614.0 .04886 22.3922 8 20.7647 9 Variance 260.5 377112.6 .002 501.414 431.176 Range 81.7 3080 .22 112.00 116.00 Minimum 9.6 1884 .39 8.00 8.00 Maximum 91.3 4964 .62 120.00 124.00 Table 5.2 Regression Results for Delay Models of DLT Model Summary (b) Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .648(a) .420 .415 .33665 Notes: Predictors (Constant), LTIN, VOL UME, DLTV, SPLIT; Dependent Variable: lnTD ANOVA (b) Model Sum of Squares df Mean Square F Sig. 1 Regressio n 37.681 4 9.420 83.122 .000(a)

PAGE 74

Table 5.2 Regression Results for Delay Models of DLT (Continued) ANOVA (b) Residual 52.132 460 .113 Total 89.813 464 Notes: Predictors (Constant), LTIN, VOLUME, DLTV, SPLIT; Dependent Variable: lnTD Coefficients (a) Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constan t) 2.364 .184 12.851 .000 TV .0004 .000 .525 11.766 .000 SPLIT -.935 .411 -.104 -2.277 .023 DLTV .004 .001 .210 4.702 .000 LTIN .004 .001 .185 4.198 .000 Notes: Dependent Variable: lnTD As shown in Table 5.2, all independent variables are significant at 95 percent level of confidence. The R square value is 0.42, which implies that the selected independent variables can explain 42% of variations in dependent variable. The residual plot for each independent variable was obtained from the results of regression analysis. It was found that the residual for each independent variable was randomly scattered about the x-axis line, which indicated that the model was correctly specified. According to these parameter estimates, the final developed regression equation was: LTINDLTVSPLITTVLeTD004.0004.0935.00004.063.10 (5-2) 64

PAGE 75

65 Where, TD L = average total delay of DLT (sec/ veh), TV = flow rate of major-road through-traffic (vph), DLTV = flow rate of DLT from a driveway (vph), LTIN = flow rate of left-turnin from major roads (vph), and SPLIT = percentage of upstream through traffic flow rate, SPLIT=TV1/ (TV1+TV2) In Equation 5-2, the coefficients of DLTV (0.004) and LTIN (0.004) are much greater than the coefficient of TV (0.0004). Th is implies that DLT and LTIN flow rate have greater impact on the delay of DLT than that of major-road through-traffic. The independent variable SPLIT has a negative co efficient, indicating that the downstream through-traffic flow rate (TV2) has a grea ter impact on the delay than corresponding upstream flow rate (TV1). This is because when the median space is occupied by other maneuvers, left-turn vehicles must wait at the driveway even if suitable gaps are available at the upstream through-traffic stream. Based on Equation 5-2, curves for the aver age delay of DLT under different traffic volume conditions can be developed. Figure 5.2 shows a group of curves for average delay of DLT assuming the left-turn-in flow rate from the major road was 100 vph, split was 0.5, and the flow rate of DLT was made equal to 50, 100, and 150 vph, respectively. The x-axis represents the flow rate of twodirectional through-traf fic on the major road. The y-axis represents the average total delay of DLT.

PAGE 76

LTIN=100 vph, SPLIT= 0.5 Figure 5.2 Curves for the Average Total Delay for DLT (LTIN=100 vph, Split=0.5) 5.2.2 Delay Model for Right Turn Followed by U-turn In this study, the total delay of RTUT includes delay at driveway and delay at signalized intersection. Past studies generally indicated that RTUT vehicles delay at signalized intersection will be affected by signalization conditions and demand flow rate, including g/c ratio, cycle length, and left-turn flow rate from inside exclusive left-turn lane. Variables expected to affect RTUT delay at driveway include through traffic flow rate, split and RTUT flow rate. Another variable that could affect RTUT delay at driveway is the distance from driveway to downstream signalized intersection. As illustrated in Figure 5.3, when this distance is shorter than the left turn deceleration lane on the major road, many drivers will select a suitable simultaneous gap in all through lanes and then make a direct entry into the left turn deceleration lane. When the distance 66

PAGE 77

is medium or long, the driver wants to make a right turn doesnt have to wait for a simultaneous gap in all three through lanes since they can easily select a suitable gap, turn into the right-side lane, accelerate to an appropriate speed, and then weave to the exclusive left turn lane. Therefore sometimes increase the distance from driveway to downstream U-turn bay will reduce RTUT delay at driveway. In practice, it is difficult to define a suitable distance from driveway to downstream intersection or median opening since there are a lot of different factors need to be considered. More details considering this topic will be included in another project. Figure 5.3 Two Different Weaving Patterns The average total delay model for RTUT movement is described as follows: 07654321/aLaCaCGaLTVaRUVaSPLITaTVaRUeTD (5-3) Where, TD RU = average total delay of RTUT (sec/ veh), TV = flow rate of major-road through-traffic (vph), RUV = flow rate of RTUT from a driveway (vph), G/C = g/c ratio for exclusive left turn phase, 67

PAGE 78

68 C = Cycle length (sec); cycle length used in pretimed signal control, or average cycle length for actuated control; LTV = left-turn flow rate from inside left turn lane; L = the distance from driveway to downstream signalized intersection, SPLIT = percentage of upstream through traffic flow rate, SPLIT = TV1/ (TV1+TV2), and a 0 a 1 a 2 a 3 a 4 a 5 a 6 a 7 = parameters The dependent variable in this model is the average total waiting delay per vehicle making a right turn followed by a U-turn at downstream signalized intersection during a fifteen-minute interval. In this study, g/c ratio is defined as the green arrow time for left-turn phase divided by the cycle length of selected signa l. If the study site is an actuated signal with varying cycle and phase length, g/c ratio is de fined as the maximum green arrow time for left-turn phase divided by average cycle length. This definition here is different from that of some other studies which generally use average green arrow time when defining g/c ratio of actuated controlled signal. However, this study found that the maximum green arrow time is a better indi cator of actuated cont rolled signal capacity, and has better statistical characteristic when incorporated into delay and travel time models. A total of 610 observations at fifteen-mi nute intervals were used to perform the regression analysis. The statistic al characteristics of collecte d data are given in Table 5.3. The final regression results are listed in Table 5.4.

PAGE 79

69 Table 5.3 Descriptive Statistic s of the Collected Data TD VOLU ME SPLI T G/C C L RUV LTV N Valid 610 610 610 610 610 610 610 610 Missin g 0 0 0 0 0 0 0 0 Mean 351.04 3575.66 .475 4 .15517 122.2 577.93 18.2721 107.7 Median 310.50 3516.00 .472 8 .15000 106.7 525.00 16.0000 104.0 Mode 226(a) 3332(a) .41(a ) .150 106.67 525 16.00 100.00 Std. Deviation 168.61 2 493.825 .066 .014 19.8100 3 149.401 5.75935 42.2 Variance 28430. 0 243862. 7 .004 .000 392.4 22320.7 33.170 1778.7 Range 1482 3320 .31 .169 57.68 600 40.00 268.00 Minimum 134 1588 .33 .106 106.67 300 12.00 8.00 Maximum 1616 4908 .63 .275 164.35 900 52.00 276.00 Notes: Multiple modes exist. The smallest value is shown Table 5.4 Regression Results for Delay Models of RTUT Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .683(a) .466 .460 .17992 Notes: Predictors (Constant), L, VOLUME, RUV, SPLIT, C, LTV, G/C ANOVA (b) Model Sum of Squares df Mean Square F Sig. 1 Regressi on 17.032 7 2.433 75.164 .000(a) Residual 19.487 602 .032 Total 36.519 609 Notes: Predictors (Constant), L, VOLUME RUV, SPLIT, C, LT V, G/C; Dependent Variable: lnTD.

PAGE 80

Coefficients (a) Mod el Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 3.358 .096 35.157 .000 VOLUME .000 .000 .327 9.959 .000 SPLIT .427 .123 .115 3.479 .001 G/C -3.483 .653 -.199 -5.333 .000 C .006 .000 .474 13.013 .000 RUV .003 .001 .076 2.386 .017 LTV .002 .000 .407 11.169 .000 L -.001 .000 -.346 -9.208 .000 Notes: Dependent Variable: lnTD As shown in Table 5.4, all independent variables are significant at 95 percent level of confidence. The R square value is 0.466, which implies that the selected independent variables can explain 46.6% of variations in dependent variable. The residual plot for each independent variable was obtained from the results of regression analysis. It was found that the residual for each independent variable was randomly scattered about the x-axis line, which indicated that the model was correctly specified. According to these parameter estimates, the final developed regression equation was: LCCGLTVRUVSPLITTVRUeTD00056.00059.0/483.3002.0003.0427.000016.073.28 (5-4) Where, TD RU = average total delay of RTUT (sec/ veh), TV = flow rate of major-road through-traffic (vph), RUV = flow rate of RTUT from a driveway (vph), G/C = g/c ratio for exclusive left turn phase, C = Cycle length (sec); cycle length used in pretimed signal control, or average cycle length for actuated control; 70

PAGE 81

71 LTV= left-turn flow rate from inside left turn lane; L = the distance from driveway to downstream signalized intersection, SPLIT = percentage of upstream through traffic flow rate, SPLIT=TV1/ (TV1+TV2), and a 0 a 1 a 2 a 3 a 4 a 5 a 6 a 7 = parameters As shown in Equation 5-4, the coefficien t of TV is very small (0.00016), which implies that the average total delay of RTUT is not very sensitive to the change in flow rate of through-traffic. The coefficient for G/C is negative, which suggests that providing a large g/c ratio for left-turn phase will redu ce RTUT delay at signal. Obviously, a long cycle length will result in long waiting delay fo r vehicles in left-tur n lane. Therefore the coefficient for C is positive. Another finding fr om this equation is that the increase of distance from driveway to downstream signalized intersection reduces average total delay of RTUT. As mentioned earlier in this chap ter, when this distance is long enough, the driver wants to make a right turn doesnt have to wait for a simultaneous gap in all three through lanes since they can easily select a su itable gap, turn into the right-side lane, accelerate to an appropriate speed, and then weave to the exclusive left turn lane. Therefore, in this condition, the RTUT drivers will have less delay at driveway as compared with the condition in which the distance is too short. Based on equation 5-4, different curves can be developed under different volume and roadway geometric conditions. Curves in Figure 5.4 are developed assuming the g/c ratio is 0.15, cycle length is 120 sec, SPLIT is 0.5, left-turn flow rate from inside left turn lane is 100 vph, and the distance form driveway to downstream signal is 560 ft. In this figure, the x-axis represents th e flow rate of major-road thr ough-traffic; the y-axis refers

PAGE 82

to the average total waiting delay per vehicle making a right turn followed by a U-turn at downstream signalized intersection during a fifteen-minute interval. SPLIT=0.5, LTV=100vph, G/C=0.15, C=120sec, L=560ft Figure 5.4 Curves for the Average Total Delay for RTUT (SPLIT=0.5, LTV=100vph, G/C=0.15, C=120sec, L=560ft) 5.2.3 Delay Comparison of DLT and RTUT One of the major objectives of this project is to compare delay of DLT and RTUT under specific traffic and roadway geometric conditions. In practice, DLT will have less delay than RTUT when conflicting volumes are very light. When conflicting volumes increase, however, the delay of DLT increases greatly because of the restrained median storage and the gap acceptance characteristics of DLT movement. Therefore, RTUT could have longer delay than DLT under high volume conditions. The crux here is to find the break point. That is to determine under what volume conditions, would DLT have more 72

PAGE 83

delay than RTUT. To achieve this objective, the curves in Figure 5.2 and 5.4 were combined together and illustrated in Figure 5.5. Figure 5.5 Comparison of Average Delay of Two Movements As shown in Figure 5.5, the breakpoints for delay of these two movements can be found as follows: 1) When both DLT and RTUT flow rates are equal to 50 vph, the average total waiting delay of RTUT is greater than that of DLT until the major-road through-traffic flow rate is greater than 5500 vph; 2) When both flow rates are equal to 100 vph, RTUT has less delay than DLT when the through-traffic flow rate is more than 5200 vph; and 3) When both flow rates are equal to 150 vph, RTUT will suffer less delay when the through-traffic flow rate is about 5000 vph. 73

PAGE 84

5.2.4 Delay Comparison of two U-turn Approaches In this chapter, the operations of two widely used U-turn approaches, U-turns at median opening in advance of signalized intersection and U-turns at signalized intersection, were also compared based on the models developed in this study and those from 2001 project. Curves for the average delay of RTUT at median opening and RTUT at downstream stream signalized intersection were developed. Figure 5.6 shows a group of curves for average delay of RTUT assuming the g/c ratio is 0.15, cycle length is 120 sec, SPLIT is 0.5, left-turn flow rate from inside left turn lane is 100 vph, the distance form driveway to downstream signal is 560 ft, and the flow rate of RTUT was made equal to 50, 100, and 150 vph, respectively. The x-axis represents the flow rate of two-directional through-traffic on the major road. The y-axis represents the average total delay of RTUT. Figure 5.6 Comparison of Average Delay of Two U-turn Approaches 74

PAGE 85

As shown in this figure, providing right turn followed by U-turn at downstream signalized intersection will suffer longer delay than the condition in which U-turn is accommodated at downstream median opening. 5.3 Average Total Travel Time In this project, the average total travel time of DLT is defined as the sum of average total waiting delay and the time for DLT vehicles crossing the through lanes. The average total travel time for RTUT includes the average total waiting delay, the running time from vehicle leaves driveway until it stops at exclusive left turn bay, plus the travel time from U-turn bay back to median opening at driveway. 5.3.1 Travel Time Model for DLT Data collected from those sites with full median openings were used to build delay model for direct left turn movements. The dependent variable is the average total travel time for DLT movements at fifteen-minute intervals. The independent variables include the flow rate of major-road through-traffic, split, the flow rate of left-turn-in traffic from a major roadway, and the flow rate of DLT. (5-5) 04321aLTINaDLTVaPLITaTVaLeTT Where, TT L = average total travel time of DLT (sec/ veh); TV = flow rate of major-road through-traffic (vph); DLTV = flow rate of DLT from a driveway (vph); LTIN = flow rate of left-turn-in from major roads (vph); SPLIT = percentage of upstream through traffic flow rate, and a 0 a 1 a 2 a 3 a 4 = parameters 75

PAGE 86

76 A total of 459 observations at fifteen-mi nute intervals were used to perform the regression analysis. The statistic al characteristics of collecte d data are given in Table 5.5. The final regression results are listed in Table 5.6. Table 5.5 Descriptive Statistic s of the Collected Data TT VOLUM E SPLIT DLTV LTIN Valid 459 459 459 459 459 N Missing 0 0 0 0 0 Mean 44.0293 3437.75 74 .5099 43.5483 48.1830 Median 42.0769 3441.33 .5043 40.0000 44.0000 Mode 33.00 3200.00( a) .44(a) 36.00 36.00 Std. Deviation 16.0580 6 615.046 91 .04858 22.2870 5 20.7773 6 Variance 257.861 378282. 704 .002 496.713 431.699 Range 82.50 3080.00 .22 112.00 116.00 Minimum 14.50 1884.00 .39 8.00 8.00 Maximum 97.00 4964.00 .62 120.00 124.00 Notes: Multiple modes exist. The smallest value is shown Table 5.6 Regression Results for Travel Time Models of DLT Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .654(a) .428 .423 .33473 Notes: Predictors (Constant), LTIN, VOLUME, DLTV, SPLIT ANOVA(b) Model Sum of Squares df Mean Square F Sig. 1 Regressio n 38.041 4 9.510 84.882 .000(a) Residual 50.867 454 .112 Total 88.908 458

PAGE 87

Coefficients(a) Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constan t) 2.369 .184 12.862 .000 TV .000 .000 .532 11.811 .000 SPLIT -.996 .416 -.110 -2.395 .017 DLTV .004 .001 .212 4.772 .000 LTIN .004 .001 .195 4.441 .000 Notes: Dependent Variable: lnTD As shown in Table 5.6, all independent variables are significant at 95 percent level of confidence. The R square value is 0.428, which implies that the selected independent variables can explain 42.8% of variations in dependent variable. The independent variable SPLIT has a negative coefficient, which suggests that the downstream through-traffic flow rate (TV2) has a greater impact on the total travel time than corresponding upstream flow rate (TV1). According to these parameter estimates, the final developed regression equation was: LTINDLTVSPLITTVLeTT0041.00042.0996.000038.069.10 (5-6) Where, TT L = average total travel time of DLT (sec/ veh); TV = flow rate of major-road through-traffic (vph); DLTV = flow rate of DLT from a driveway (vph); LTIN = flow rate of left-turn-in from major roads (vph); SPLIT = percentage of upstream through traffic flow rate. Based on Equation 5-6, curves for the average total travel time of DLT can be developed. Figure 5.7 shows a group of curves for average total travel time of DLT 77

PAGE 88

assuming the left-turn-in flow rate from the major road was 100 vph, split was 0.5, and the flow rate of DLT was made equal to 50, 100, and 150 vph, respectively. The x-axis represents the flow rate of two-directional through-traffic on the major road. The y-axis represents the average total travel time of DLT. LTIN=100 vph, SPLIT=0.5 Figure 5.7 Curves for the Average Total Travel Time for DLT 5.3.2 Travel Time Model for RTUT The average total travel time for RTUT includes the average total waiting delay, the running time from vehicle leaves driveway until it stops at exclusive left turn bay, plus the travel time from U-turn bay back to median opening at driveway. The same datasets for the delay models were used to develop the travel time model for DLT and RTUT. The statistical characteristics of collected data are given in Table 5.7. The final regression results are listed in Table 5.8. 78

PAGE 89

79 Table 5.7 Descriptive Statistic s of the Collected Data TT VOLUM E SPLIT RUV LTV Valid 610 610 610 610 610 N Missing 0 0 0 0 0 Mean 99.6765 3567.96 .4752 18.3858 109.602 2 Median 98.3661 3514.67 .4728 16.0000 104.000 0 Mode 124.99 3440 .41(a) 16.00 100.00 Std. Deviation 18.8543 4 496.254 .06558 5.81191 43.3931 3 Variance 355.486 246267. 629 .004 33.778 1882.96 4 Range 99.72 3328 .31 40.00 268.00 Minimum 57.87 1580 .33 12.00 8.00 Maximum 157.58 4908 .63 52.00 276.00 Notes: Multiple modes exist. The smallest value is shown Table 5.8 Regression Results for Travel Time Models of RTUT Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .676(a) .457 .450 .12992 Notes: Predictors (Constant), LTV, SPL IT, RUV, VOLUME, C, G/C, L, SPEED ANOVA(b) Model Sum of Squares df Mean Square F Sig. Regressi on 8.550 8 1.069 63.321 .000(a) Residual 10.144 601 .017 1 Total 18.695 609 Notes: Predictors (Constant), LTV, SPLIT RUV, VOLUME, C, G/C, L, SPEED ; Dependent Variable: lnTT

PAGE 90

80 Coefficients(a) Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 4.926 .221 22.247 .000 VOLUME .000 .000 .356 10.309 .000 SPLIT .192 .090 .072 2.135 .033 G/C -.691 .504 -.058 -1.370 .171 C .006 .000 .684 13.886 .000 L .000 .000 -.287 -5.466 .000 SPEED -.032 .006 -.324 -5.247 .000 RUV .002 .001 .075 2.309 .021 LTV .001 .000 .355 9.222 .000 Notes: Dependent Variable: lnTT The t-stat indicated that most of the i ndependent variables ar e significant at 95% level of confidence except G/C, which is significant at 80% level of confidence. The coefficient of SPLIT is positive, which indi cates that the upstream through-traffic flow rate (TV1) has a greater impact on the trav el time than corresponding downstream stream flow rate (TV2). The coefficient of L is very small (-0.00046), which suggests that the travel time of RTUT is not very sensitive to the distance from driveway to downstream signalized intersection. As mentioned early in this chapter, the RTUT will take less delay at driveway when the downstream signal is lo cated at a suitable dist ance from driveway. Therefore, when L increased, th e total travel time of RTUT could decrease because of the reduced delay at drivew ay. This conclusion may not be hold when this distance is getting too long. In this condition, it will take long time for RTUT to traverse the weaving section, and the total travel time of RTUT could get increased. The maximum L of the selected sites is 900 ft. It is not easy to find a suitable distance from driveway to downstream signalized intersecti on or U-turn median openi ng, because a lot of factors

PAGE 91

need to be considered besides the delay and travel time of RTUT. Another project will be conducted concerning this topic and more details surrounding the selection of this distance will not be explained in this study. The empirical equation based on the regression results is as follows: SPEEDLCCGLTVRUVSPLITTVRUeTT032.000032.00061.0/691.00014.00023.0192.000013.08.137 (5-7) where, TT RV = average total travel time of RTUT (sec/ veh); TV = flow rate of major-road through-traffic (vph); RUV = flow rate of RTUT from a driveway (vph); LTV = left-turn flow rate from inside left turn lane; G/C = g/c ratio for exclusive left turn phase; C = Cycle length (sec); cycle length used in pretimed signal control, or average cycle length for actuated control; SPLIT = percentage of upstream through traffic flow rate; L = The distance from driveway to downstream signalized intersection (ft), SPEED = speed limit along the arterial (mph) Based on Equation 5-7, curves for the average total travel time of RTUT can be developed. Figure 5.8 is an example assumes that the g/c ratio is 0.15, cycle length for downstream signal is 120 sec, SPLIT is 0.5, left-turn flow rate from inside left turn lane is 100 vph, and the distance form driveway to downstream signal is 560 ft. Three different curves represent different volume conditions in which flow rate of RTUT is 50, 100, and 150vph respectively. The x-axis represents the flow rate of two-directional through-traffic on the major road. The y-axis represents the average total travel time of DLT. 81

PAGE 92

Figure 5.8 Curves for Average Total Travel Time for RTUT (SPLIT=0.5, LTV=100vph, G/C=0.15, C=120sec, L=560ft, and SPEED=45mph) 5.3.3 Travel Time Comparison of DLT and RTUT Given the travel time models for DLT and RTUT, the average total travel time of these two movements can be compared under different traffic and roadway geometric conditions. Figure 5.8 is an example assuming that g/c ratio is 0.15, cycle length for downstream signal is 120 sec, SPLIT is 0.5, left-turn flow rate from inside left turn lane is 100 vph, and the distance form driveway to downstream signal is 560 ft. In this figure, the curves from Figure 5.7 and 5.8 were combined together and shown as follows: 82

PAGE 93

Figure 5.9 Comparison of Average Total Travel Time of Two Movements As shown in Figure 5.10, the breakpoints for average total travel time of these two movements can be found as follows: 1) When both DLT and RTUT flow rates are equal to 50 vph, the average total travel time of RTUT is greater than that of DLT until the major-road through-traffic flow rate is greater than around 6600 vph; 2) When both flow rates are equal to 100 vph, RTUT has less delay than DLT when the through-traffic flow rate is more than about 6300 vph; and 3) (3) When both flow rates are equal to 150 vph, RTUT will suffer less delay when the through-traffic flow rate is about 6000 vph. 5.3.4 Travel Time Comparison of Two U-turn Approaches Travel time model for RTUT developed in this project was compared with the model from 2001 project; therefore the operations of two widely used U-turn approaches, 83

PAGE 94

including U-turn at median opening in advance of signalized intersection and U-turn at signalized intersection, can be compared. Figure 5-11 is an example. Curves based on Equation 5-7 and 2001 model were shown in this Figure. Figure 5.10 Comparison of Average Total Travel Time of Two U-turn Approaches As illustrated in this figure, providing right turn followed by U-turn at downstream signalized intersection will suffer longer travel time than the condition in which U-turn is accommodated at downstream median opening. 5.4 Amount of RTUT under Both Choices The 2001 project indicated that when there is a suitable U-turn median opening downstream, some drivers prefer to make a RTUT rather than a DLT to avoid conflict with all other movements at the median opening. This decision is encouraged when the median storage space is occupied by other maneuvers or when there is a large left-turn-in 84

PAGE 95

85 volume from the major-road. Therefore, in th is condition, the driv ers selection of a RTUT or a DLT will be affected by traffic volume conditions. For the condition in which U-turn is accommodated at signalized intersection, drivers choice behavior is different. As i ndicated earlier, a driver making a U-turn at signalized intersection doesnt need to wait and find a suitable gap form right-side through-traffic since U-turn is made at protected signal phase. Therefore some drivers may prefer this option with the perception that it is safer. However, there are also some drivers dont like making a U-turn at signalized intersection since they think it takes long time to wait for the signal to turn green. In this study, a binary logistic regression model was developed to estimate under what traffic and roadway geometric conditions would more drivers select RTUT rather than DLT. The reason for choosing logistic regression lies in the bounded nature of dependent variable (The percen tage of drivers selecting RT UT rather than DLT always varies between 0 and 1). In this model, the ratio of RTUT was defined as the number of RTUT divided by the sum of DLT and RTUT at fifteen-minute intervals as shown in Equation 5-8. RATIO= (# of RTUT)/ (# of RTUT+ # of DLT) (5-8) Data collected from site 2, 3, 5, 6, and 7 were used to build this model, because only these sites permit both of DLT and RTUT movements. A total of 381 observations at fifteen-minute intervals were used to develop this model. Only intervals when there are both DLT and RTUT were chosen to perfor m the regression analysis. The regression results are given in Table 5.9

PAGE 96

86 Table 5.9 Regression Results for Ratio of RTUT Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .712(a) .507 .502 .45102 Notes: Predictors (Constant ), SPLIT, L, TVOLUME, LTIN ANOVA(b) Model Sum of Squares df Mean Square F Sig. 1 Regressi on 78.734 4 19.683 96.764 .000(a) Residual 76.485 376 .203 Total 155.219 380 Notes: Predictors (Constant), SPLIT, L, TVOLUME, LTIN; Depe ndent Variable: lnP Coefficients(a) Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constan t) -1.162 .233 -4.993 .000 L -.003 .000 -.931 -15.475 .000 LTIN .014 .004 .864 14.034 .000 TV 6.262E-0 5 .000 .051 1.217 .224 SPLIT 1.933 .359 .202 5.377 .000 Notes: Dependent Variable: lnP T-stat indicated that most of the inde pendent variables are significant at 95% confidence level except the flow rate of thr ough traffic, which is significantly at 75% level of confidence. The coefficient for SPL IT is positive, which indicates that the upstream through-traffic flow rate (TV1) has a greater impact on drivers decision than

PAGE 97

corresponding upstream flow rate (TV2). In practice, when the downstream signal turns red, DLT vehicles waiting at median opening can easily find a gap from downstream traffic and then make a left turn. This is why downstream traffic does not significantly affect drivers selection of RTUT as compared with upstream through-traffic flow rate. As shown in this model, the coefficient of the independent variable L, which represents the distance from driveway to downstream signalized intersection, is negative. In the real world, some drivers prefer to make a RTUT rather than a DLT when the downstream signal is close to the driveway. This conclusion may not be hold when this distance is too short, for example, shorter than the length of left-turn storage bay. In this condition, drivers who want to make a RTUT are often blocked by through traffic already queued at the traffic signal. Drivers must cross the intersection, and then make a U-turn at a median opening downstream of the signalized intersection. This is another U-turn approach which needs to be considered and evaluated. More information regarding the distance from driveway to signalized intersection and U-turns after signal will be incorporated into other projects and are not detailed in this study. Based on the regression analysis, the final equation for estimating percentage of drivers selecting RTUT was: LSPLITTVLTINRATIORATIO003.0933.100006.0014.0162.1)1ln( (5-9) Where, RATIO = percentage of RTUT at fifteen-minute intervals; TV = flow rate of major-road through-traffic (vph); LTIN = flow rate of left-turn-in from major roads (vph); SPLIT = percentage of upstream through traffic flow rate; 87 L = the distance from driveway to signalized intersection (ft).

PAGE 98

Curves were developed based on Equation 5-9, which assumes the split is 0.5, the distance from driveway to downstream signalized intersection is 560ft, and the flow rate of major road through traffic is made equal to 2000, 4000, and 6000 vph, respectively. The x-axis represents the flow rate of left-turn-in from major road. The y-axis represents the percentage of drivers choosing right turn followed by a U-turn at downstream signalized intersection rather than making a direct left-turn from driveway when both choices are available. SPLIT=0.5, L=560ft Figure 5.11 Ratio of RTUT vs. Left Turn in Volume and Through Volume Based on the figure, it is clear that: 1) When the flow rate of major road through traffic is equal to 2000 vph, more drivers will select RTUT when the flow rate of left-turn-in from major road is about 130 vph or higher; 88

PAGE 99

2) When the flow rate of major road through traffic is equal to 4000 vph, the ratio is getting close to 50 percent when the flow rate of left-turn-in from major road is about 120 vph or higher; and 3) When the flow rate of major road through traffic is equal to 6000 vph, the ratio is greater than 50 percent when the flow rate of left-turn-in from major road is larger than 110 vph. 5.5 U-turns Effects on Signalized Intersection Capacity In this project, U-turn effects on signalized intersection were estimated by applying the adjustment factor for U-turn movements on left-turn saturation flow rate. The pilot survey conducted at early stage of this project indicated that U-turning vehicles have some adverse impacts on intersection capacity, and this effects increase with the percent of U-turning vehicles from inside left-turn lane. In addition, the pilot survey found that when there are U-turning vehicles in the exclusive left-turn lane, the queue discharge patterns do not display an easily identifiable steady maximum rate. Therefore traditional headway research, which assumes that discharge flow rate reaches saturation state after fourth or fifth discharged vehicle, can not be used in estimating the adjustment factor for U-turning vehicles. A procedure was developed to estimate the relationship between the percentage of U-turning vehicles in the left-turn lane and the average queue discharge time for each vehicle. The average queue discharge time was defined as the queue discharge time divided by the number of vehicles in the discharged queue as shown in equation 5-10: luNNTh (5-10) 89

PAGE 100

90 where, h = average queue discharge time for each vehicle (sec); T =queue discharge time (the time from the be ginning of green until the rear axle of the last vehicle in queue cros ses the stop line) (sec); N u = the number of U-turning vehicles in queue; and N l = the number of left-turning vehicles in queue. The study team selected three intersecti ons with exclusive left-turn lanes and protected signal phasing, and recorded discharge time for 260 queues, including 571 U-turning vehicles and 1441 left-turning vehi cles. These data are used to build an empirical model which estimates the relations hip between the percentage of U-turning vehicle and the average queue discharge time. Analysis showed that a cubic polynomial regression model is ap propriate to describe the rela tionship. Figure 5.12 presents the distribution of collected data. The regressi on results are listed in Table 5.10. The average queue discharge time model was de scribed as Equation 5-11.

PAGE 101

Figure 5.12 Average Queue Discharge Time versus the Percentage of U-turning Vehicles in Queue Table 5.10 The Regression Results for the Average Queue Discharge Time Model Summary Mode l R R Square Adjusted R Square Std. Error of the Estimate 1 .681(a) .464 .458 .20339 Notes: Predictors (Constant), Put2, Put, Put3 ANOVA(b) Model Sum of Squares df Mean Square F Sig. 1 Regressi on 9.183 3 3.061 73.998 .000(a) Residual 10.590 256 .041 Total 19.773 259 Notes: Predictors (Constant), Put2, Put, Put3; Dependent Variable: h 91

PAGE 102

Coefficients(a) Model Unstandardized Coefficients Standardiz ed Coefficien ts t Sig. B Std. Error Beta 1 (Constant ) 2.149 .026 81.472 .000 Put .002 .003 .185 .611 .542 Put3 -3.034E07 .000 -.293 -.540 .589 Put2 7.771E-0 5 .000 .786 .984 .326 Notes: Dependent Variable: h 1487.20018.000008.00000003.023UTUTUTPPPh (5-11) where, h = average queue discharge time for U-turn and left-turn mix flow (sec); P UT = percentage of U-turn vehicles from inside left-turn lane (%); luuUTNNNP and a 1 a 2 a 3 a 4 = parameters. The dependent variable in this model is the average queue discharge time for each vehicle. Considering the intercept, which represents the base average queue discharge time assuming no U-turning vehicles in left-turn flow, this model provides a reasonable value of 2.1487 sec. It is important to note that the definition of average queue discharge time in this model is different from that of the saturation headway, which is very difficult to be measured when there are U-turning vehicles in discharging queue. The R square value of this model is about 0.46. 92

PAGE 103

Based on the definition of the adjustment factors for turning movements, the U-turn adjustment factor for the left-turn saturation flow rate can be estimated by the following equation: 1487.20018.000008.00000003.01487.2360036002300UTUTUTUTPPPhhhhf (5-12) where, f UT = adjustment factor for U-turn movement; h= average queuing discharge time for U-turn and left-turn mix flow; h 0 = base average queuing discharge time for left-turn only flow (sec); and P UT = percentage of U-turn vehicles from inside left-turn lane (%); By using the Equation 5-12, the U-turn adjustment factor for different percent of U-turning vehicles are calculated and listed as follows: Table 5.11 Adjustment Factor for U-turn movements P UT (%) 5 10 20 30 40 50 60 70 80 90 100 f UT 0.99 0.99 0.97 0.95 0.92 0.89 0.87 0.84 0.81 0.78 0.76 From this table, it is clear that U-turning vehicles have considerable effects on the traffic flow in a left-turn lane, especially when the percent of U-tuning vehicles is high (>40%). Therefore, when estimating the left-turn lane capacity, it is essential to account for the effects of U-turning vehicles. This effect can be quantified by applying the adjustment factors for U-turn movements on left-turn saturation flow rate, as those listed in Table 5.11. 93

PAGE 104

94 5.6 Summary Four major conclusions are made in this chapter including: First, delay and travel time models for DLT and RTUT can be used to determine under what traffic flow rate conditions (maj or road, left-turn-in, and driveway) DLT would experience more delays or tr avel time as compared to RTUT. Second, by comparing the delay and travel time models which were developed in this chapter with those from the 2001 project, it is clear th at providing RTUT at median opening in advance of signalized intersection will suffer less delay and travel time as compared with the condition in which U-turns are accommodated at signalized intersection. Third, the driver selection of a RTUT or a DLT on the basi s of accessibility considerations is affected by both traffi c flow and roadway geometric conditions. A binary logistic regression m odel was developed to estimate the relationship between the percentage of RTUT and e xplanatory variables; and Last, the effects of U-turns on left-turn la ne capacity can not be ignored especially when there is a large percentage of U-tu rning vehicles (>40%). A cubic polynomial regression model was developed to estimate the relationship between the average queue discharge time for each vehicle and the percent of U-turning vehicles in left-turn lane. Adjustment factors for varying percents of U-turning vehicles in left-turn lane are established by usi ng this model.

PAGE 105

CHAPTER 6 SUMMARY AND CONCLUSIONS 6.1 Summary Florida uses restrictive medians and directional median openings in the State Highway System to manage left turn egress movements from driveways and side streets. By installing raised curb medians and replacing full median openings with directional median openings in some places, direct left turn movements are substituted by making a right-turn followed by a U-turn at downstream median opening or signalized intersection. This report is one of the two reports that evaluated the safety and traffic operational effects of a widely used access management treatment: U-turns at downstream signalized intersection as alternatives to direct left turns. This research focused on evaluating the traffic operational impacts of RTUT and DLT movements. The primary objectives of this part of the study were to explore methodologies for evaluating the operational effects of U-turns at signalized intersection as alternatives to direct left turns and to provide information on the potential impacts of these alternatives under various conditions. To achieve these objectives, field measurement was conducted on eight selected street segments in Tampa Bay area. A total of more than 300 hours of traffic data were collected using video cameras. While reviewing videotapes, each vehicle coming from the driveway 95

PAGE 106

96 making DLT or RTUT was tracked. Delay and tr avel time for each DLT or RTUT vehicle were recorded. Other information reduced from videotapes includes traffic volume, signal parameters, and queuing discharge ti me in inside left-turn lane. Delay and travel time models were deve loped based on collected field data. The delay and travel time of DLT and RTUT were determined as a function of conflicting volumes, signalization conditions, and road way geometric conditions. Curves were developed based on regression results depict ing operational differences between making a DLT versus making a RTUT. The curves demonstrated the point at which a driver making a right turn followed by a U-turn at downstream signalized intersection experiences less delay and travel time than a dr iver attempting to make a direct left turn through a median opening onto a major road. In this project, the operations of two widely used U-turn approaches: providing U-turn at median opening in advance of signalized intersection and U-turn at signalized intersection, were also compared by comp aring the delay and travel time models developed in this chapter with those from the 2001 project. Based on the comparison, it is clear that providing RTUT at median opening in advance of signalized intersection will experience less delay and travel time as comp ared with the condition in which U-turns are accommodated at signalized intersection. Drivers selection of RTUT or DLT may be affected by some traffic characteristics such as through-traffic volume, left-turnin volume, and so on. In addition, field measurement found that drivers choice of RTUT is also affected by the distance from driveway to downstream signalized intersecti on. A binary logistic regression model was developed to estimate how many drivers would like to make a RTUT rather than a DLT

PAGE 107

97 under specific traffic and roadway geometric conditions when U-turns are provided at downstream signalized intersecti on. The findings indicated that the left turn in volume had significant impacts towards increasing the amount of RTUT. Additionally, the RATIO model also indicated that fewer drivers will select RTUT wh en the distance from driveway to downstream signalized intersection is relatively long. In order to evaluate U-turns effects on signalized intersection capacity, a procedure was developed to apply an adjustment factor for U-turn movements on left-turn saturation flow rate. A third-degree polynomial regression model was developed to estimate U-turn effects on queue discharge time. The dependent variable of this model is the average queue discharge time for each vehicle. The independent variable is the percentage of U-turning vehicles of left-turn lane. Adjustme nt factors for varying percents of U-turning vehicles in left-turn lane are establ ished by the use of this model. 6.2 Conclusions This study developed a procedure to estim ate the operational effects of a widely used access management treatment: U-turns at signalized intersection as alternatives to direct left turns. Though this study, conclusi on can be made that U-turns at signalized intersection could have better operational performance than di rect left turns under certain traffic and roadway geometric conditions. More specifically, th e findings of this study include following. First, the curves based on delay and travel time models indicated that under high major road and driveway volume conditions, vehicles making a direct left turn will experience longer delay and travel time than t hose make a right turn followed by a U-turn at downstream signalized intersection.

PAGE 108

98 Second, when major road and driveway volume are not very high, vehicles making a right turn followed by a U-turn at downstr eam signalized intersection will suffer longer delay and travel time than those make a dire ct left turn. The break point, which indicates the specific volume condition in which DLT will have more delay than RTUT, can be estimated by using the delay and travel ti me models developed in this project. Third, considering the selection of differ ent U-turn approaches, providing U-turns at median opening in advance of signalized intersection will experience less delay and travel time as compared with the conditi on in which U-turns are accommodated at signalized intersection. Forth, the percentage of RTUT movements increases with left-turn-in flow rate from major-road and major-road through-tr affic flow rate; and decreases with the distance from driveway to downs tream signalized intersection. Finally, when estimating the left-turn lane capacity, it is essen tial to account for the effects of U-turning vehicles especially when the percent of U-turning vehicles is large (>40%). This effect can be quantified by ap plying the adjustment factors for U-turn movements on left-turn saturation flow rate. In this project, a cubic polynomial regression model was developed to estimate the relati onship between the average queue discharge time for each vehicle and the percent of U-turn ing vehicles in left-turn lane. Adjustment factors for varying percents of U-turning ve hicles in left-turn lane are established by using this model. 6.3 Recommendations The findings of this study are helpful in providing local and state transportation agencies with recommendations for the design and selection of medi an treatments in six

PAGE 109

99 to eight-lanes urban or subur ban arterial roads. The poten tial median treatments include the installation of a restrictive median, closing median openings and replacing full medina opening with a directional one. Delay and travel time models provide a tool to help address public concerns related to the opera tional impacts of U-turns and would be particularly helpful in identifying the circumstances where the right turn followed by U-turn takes less time than the direct left turn. Adjustment factors for varying percent of U-turning vehicles developed in this study can be directly used in estimating th e capacity of a signalized intersection where U-turns are accommodated. The percentage of drivers selecting RTUT was found to be greatly affected by Left-turn-in volume from major road. For exam ple, when the flow rate of major road through traffic is equal to 4000 vph, more driv ers (>50%) tend to select RTUT when the flow rate of left-turn-in from major ro ad is higher than 120 vph. In addition, the left-turn-in volume also has a dramatic im pact on the delay of left turn out from driveways. Field measurement found that DLT dr ivers often refuse to yield to left-turn-in vehicles especially when the vehicles delay at driveway is increa sing. Sometimes it will cause accident. These findings indicated that le ft-turn-in volume could be an important indicator when considering replacing a fu ll median opening with a directional one. This study found that providing U-turn s at median opening in advance of signalized intersection will experience less delay and travel time as compared with the condition in which U-turns are accommodated at signalized intersection. This conclusion may be helpful when selecting a suitable Uturn approach. It was also found in this project that providing U-turns at signalized intersection will not only increase U-turning vehicles delay and travel time, but also d ecrease the signalized intersection capacity.

PAGE 110

100 From the authors point of view, U-turns at a median opening in advance of signalized intersection has more potential operational benefits than U-turns at signalized intersection. However, it is also important to note that the operational performance of RTUT or DLT is not the only criterion for design and selecting median treatment. When selecting a median treatment, safety should have the first priority in decision making. Several issues were not addressed in th is study including operational effects of U-turns at four-lane road, th e selection of the optimum distance from driveway to downstream signalized intersection or median opening, and the operati onal evaluation of another widely used U-turn approach: right turns followed by U-turns at a median opening after downstream signalized intersecti on. It would to address these issues in future research.

PAGE 111

101 REFERENCES 1. Access Management Manual, (2003). Transportation Research Board, Washington, D.C. 2. Access Management, Location and Design, Participant Notebook. (2000). National Highway Institute, NHI Course No. 133078. 3. J. Lu, S. Dissanayake, H. Zhou and X. Yang, (2001). Operational Evaluation of Right Turns followed by U-turns as an Al ternative to Direct Left Turns, University of South Florida. 4. J. Lu, S. Dissanayake a nd N. Castillo, (2001). Safety Evaluation of Right-Turns followed by U-turns as an Alternative to Direct Left Turns: Conflict Data Analysis, University of South Florida. 5. J. Lu, S. Dissanayake and L. Xu, (2001). Safety Evaluation of Right-Turns followed by U-turns as an Alternative to Direct Left Turns: Crash Data Analysis, University of South Florida. 6. Median Handbook, (1997). Florida Departme nt of Transportation. 7. Florida Statute 335 The Access Management Act (1993). 8. Rule Chapter 14-96: State Highw ay System Connection Permits, Department of Transportation. 9. W. E. Frawley, (1993). Access management in Florida, Texas Transportation Institute. 10. Rule Chapter 14-97: State Highway Syst em Access Management Classification System and Standards, Department of Transportation. 11. Gary H. Sokolow. (1993). Practical Consideration fo r Beginning a Comprehensive Access Management Program. Florida Department of Transportation. The first national access management conference.

PAGE 112

102 12. T. F. Barry, (2003). Median Opening and Access Management Decision Process Systems Planning. 13. R. P. Roess, W. R. McShane and E. S. Prassas, (1998). Traffic Engineering. Second Edition. 14. F.V. Webster, (1958). Traffic Signal Settings. Road Research Technical Paper No.39. London. 15. G. New, (1965). Approximation Methods for Queu es with Application to the Fixed-Cycle Traffic Light SIAM Review, Vol.7. 16. A. Miller, (1968). Settings for Fixed-C ycle Traffic Signals ARRB Bulletin3, Australian Road Research Board, Victoria, Australia. 17. A. Miller, (1968). The Capacity of Signalized Intersections in Australia. ARRB Bulletin3, Australian Road Research Board, Victoria, Australia, 18. D.R. McNeil, (1968). A solution to the fixed-cycle tr affic light problem for compound poission arrivals. Journal of Applied Probability 5, 624-635. 19. D. Heidemann, (1994). Queue Length and Delay Distribut ions at Traffic Signals. Transportation Research B 28B, 377-389. 20. Highway Capacity Manual (2000) Transportation Research Board, National Research Council, Washington, DC. 21. R. Tapio Luttinen, Delay at Signalized Intersections: A Comparison of Capcal 2, DanKap, and HCM 2000. TL Consulting Engineers, Ltd. 22. F. Dion, H. Rakha and Y. Kang. (2004). Comparison of delay estimates at under-saturated and over-saturated pre-timed Signalized Intersections. Transportation Research Part B 38 99-122. 23. M. Qureshi, A Comparison of Control Delay by SimTraffic and CORSIM for Actuated Signals versus Observed Control Delay. Department of Civil Engineering. University of Missouri Rolla. 2003 TRB meeting 24. A.E. Radwan and C. S. Kumares, (1980). Gap Acceptance and Delay at Stop Controlled Intersections on Multi-lane Divided Highways, ITE Journal, pp.38-44

PAGE 113

103 25. Highway Capacity Manual Special Report 209. (1994). Transportation Research Board, National Research Council, Washington, DC. 26. Z. Tian, M. Kyte, and J. Colyar, (1997). Field Measurement of Capacity and Delay at Unsignalized Intersections, ITE Journal, pp.22-26. 27. W. Brilon, and N Wu, (1999). Capacity at Unsignalized Two-stage Priority Intersections. Transportation Research Part A 33, pp.275-289. 28. B. W. Robinson, Z. Tian, W. Kittelson, M. Vandehey, M. Kyte, W. Brilon, N. Wu, and R. Troutbeck, (1999). Extensions of Theoretical C apacity Models to Account for Special Conditions. Transportation Research Part A 33, pp. 217-233. 29. J. C. Adams, and J. E.Hummer, (1993). Effects of U-turns on Left-turn Saturation Flow Rates. Transportation Research Record 1398. 30. S. Tsao and S. Chu. A Study on Adjustment Factors for U-turns in Left-turn Lanes at Signalized Intersections, Journal of Advanced Transportation, Vol. 29, No. 2, pp. 183-192. Successful Median Modifi cations Project-Case Study. 31. F.V. Webster, and B.M. Cobbe, (1966). Traffic Signals. Ministry of Transport. Road Research Technical No.56. HMSO London.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001478787
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040811s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000431
035
(OCoLC)56563592
9
AJS2477
b SE
SFE0000431
040
FHM
c FHM
090
TA145 (ONLINE)
1 100
Fan, Jingjing.
0 245
Operational evaluation of right turns followed by U-turns at signalized intersections as an alternative to direct left turns
h [electronic resource] /
by Jingjing Fan.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (M.S.C.E.)--University of South Florida, 2004.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 113 pages.
520
ABSTRACT: Direct left turn (DLT) movements have a lot safety and operation problems. They increase conflicts, delay and reduce capacity of the roadway. Increasing, many states and transportation agencies are considering using right-turns followed by u-turns as an alternative to direct left turns. Recently, Florida state are planning to close median openings in order to reduce the safety and operational problems caused by DLT. The left-turn egress movement would then to be made by turning right to the arterial road and then making a u-turn at an available downstream median opening or a signalized intersection. FDOT sponsored a project in 2001 (Methodology to quantify the Effects of Access Management Treatments on Roadway Operations and Safety) which has provided useful information on operational effect of right-turn followed by u-turn at median opening as an alternative to the direct left turn from driveways. However, the impact of the treatment---right-turn followed by u-turn at signalized intersection as an alternative to direct left turn is still not clear. A operational comparison of this treatment would be valuable. Field data collection was conducted to quantify the operational effects of right turn followed by u-turn at signalized intersection versus direct left turns. Totally eight sites around the Tampa Bay area were selected and more than 300 hours field data including delay, travel time, traffic volume, traffic control, headway, signal timing and geometric data were collected. Data reduction was conducted after that and several statistic models were developed evaluating total travel time, travel delay of these two movements and the ratio of drivers choice of RTUT versus DLT. As the conclusion, the study found that RTUT followed by U-turn at signalized intersection has longer delay and longer travel time as compared to DLT and the existence of long cycle downstream signal will discourage driver's selection of RTUT. There conclusions are very important in that they provided decision makers with useful information regarding the selection of the suitable U-turn location.
590
Adviser: Jian John Lu.
653
DLT.
delay.
travel time.
RTUT.
690
Dissertations, Academic
z USF
x Civil Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.431