USF Libraries
USF Digital Collections

Comparison of various methods to compute access density and proposing a weighted methodology

MISSING IMAGE

Material Information

Title:
Comparison of various methods to compute access density and proposing a weighted methodology
Physical Description:
Book
Language:
English
Creator:
Saxena, Meeta
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Access density
Driveways
Median openings
Signalized intersections
Un-signalized intersections
Dissertations, Academic -- Civil & Environmental Engineering -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This study aims to compare three distinct methods used to compute access density and provide a comprehensive weighted methodology to enable standardization for research and application in the future. Access density is a widely used concept that calculates the number of access points within a given distance and has been extensively applied to studies related to crash modeling, operational impact and planning. Methods used in past research show that access density is computed differently by different studies and all studies do not include all access points. The weighted methodology proposed takes into account all access points including driveways, intersections and median openings and categorizes them into geometric combinations. Each geometric combination have potential number of conflict points which include diverging, weaving, merging and crossing movements depending on the type of access point. Weights were assigned to each geometry type based on these conflict point ratio. In conclusion the study identifies and compares methods previously used to compute access density and accordingly, recommends a weighted methodology that includes all access points which can be used as a standard, universal measure all access density related studies including but not limited to safety impacts, operational impacts and planning guidelines.
Thesis:
Thesis (M.S.C.E.)--University of South Florida, 2010.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Meeta Saxena.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains X 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:
usfldc doi - E14-SFE0003454
usfldc handle - e14.3454
System ID:
SFS0027769:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Comparison of V arious M ethods t o C ompute A ccess D ensity and Proposing a Weighted Methodology By Meeta Saxena 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 Co Major Professor: Jian Lu, Ph.D. Co Major Professor: Chanyoung Lee, Ph.D. Krist i ne Williams, M.U.P Date of Approval: November 6, 200 9 Keywords: access density, driveways, median openings, signalized intersections, un signalized intersections Copyright 20 10 Meeta Saxena

PAGE 2

DEDICATION I dedicate this work to the four pillars of my life: God, my parents, and my b rother and my friends. Without you all my life would fall apart. M y parents Kishore and Chhaya who sent me to University of South Florida to pursue my Masters degree They alway one. Thanks for all the encouragement and inspiring my love for transportation. M y brother Rahul has always extended his support and invaluable guidance which motivates me to move ahead towards my career goals. Thank you very much Rahul Da. My dear friends Tejsingh, Enrique, Achilleas, Anita, Samira, Shweta, Nidhi Suhas Kranthi and Qing and all friends from CUTR you have constantly support ed encourage d and always had time to help no matter how busy you all were. Without your help I would never have achieved this. Last and most importantly I am thankful to my dearest friend God for my existence. I my Friend has always given me strength.

PAGE 3

ACKNOWLEDGMENTS I would like to express my appreciation to my graduate committee: Dr. John Lu, Dr. Chanyoung Lee, and Ms Kristine Williams. Thank You for agreeing to be a part of my graduate committee. I would like to thank Dr. John Lu for providing me with his guidance and support throughout the Master s program His knowledge and expertise is highly respected and appreciated in the field of transportation I have had the opportunity to take courses under b oth Dr. Lu and Kristine Williams and they have made a positive impact on my education by providing me with their knowledge through the teaching process Dr. Chanyoung Lee has been instrumental in providing his guidance and support throughout my research work. His mentorship and support in the pursuit of my career goals have been enormous. I would like to specially thank Dr. Huaguo Zhou who has been very helpful in suggesting this topic for my Master thesis and for the initial guidance provided in developing the study. Special thanks to t he Center for Urban Transportation Research (CUTR) for giving me the opportunity to work as a graduate research assistant while pursuing my degree. The knowledge and personal growth I have acquired at CUTR ha ve been tremendous. I would also like t o thank my graduate committee for their time and guidance. I am extremely grateful to everyone who has played a part in the completion of this thesis.

PAGE 4

i TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ..................... iii LIST OF FIGURES ................................ ................................ ................................ .................... v LIST OF EQUATIONS ................................ ................................ ................................ ............ vii ABSTRACT ................................ ................................ ................................ ............................. viii CHAPTER 1 INTRODUCTION ................................ ................................ ................................ 1 1.1 Background ................................ ................................ ................................ .......... 1 1.2 Problem Statem ent ................................ ................................ ............................... 3 1.3 Proposed Study ................................ ................................ ................................ ..... 3 1.4 Approach ................................ ................................ ................................ .............. 4 1.5 Brief Results ................................ ................................ ................................ ......... 5 1.6 Outline of Thesis ................................ ................................ ................................ .. 6 CHAPTER 2 LITERATURE REVIEW ................................ ................................ ..................... 7 2.1 Crash Modeling Related Research ................................ ................................ ....... 7 2.2 Operational Impact Related Research ................................ ................................ 11 CHAPTER 3 METHODOLOGY ................................ ................................ ............................. 14 3.1 Existing Computational Methodology ................................ ............................... 14 3.1.1 Considering Signalized Intersection and Driveways Only ...................... 15 3.1.2 Considering Both Intersections and Driveways Only .............................. 16 3.1.3 Considering All Intersections, Driveways and Median Openings ................................ ................................ ................................ .. 17 3.2 Weighted Computational Methodology ................................ ............................. 18 3.2.1 Combination of Subjective Judgment and Conflict Point Ratio ................................ ................................ ................................ ......... 18 3.2.2 Proposed Weighted Methodology Based on Detailed Geometry Types ................................ ................................ ....................... 25 CHAPTER 4 DATA COLLECTION ................................ ................................ ....................... 32 4.1 Study Location ................................ ................................ ................................ ... 32 4.2 Data Collection ................................ ................................ ................................ ... 33 4.2.1 Access Data ................................ ................................ .............................. 33 4.2.2 Crash Data ................................ ................................ ................................ 35 CHAPTER 5 DATA ANALYSIS ................................ ................................ ............................ 39 5.1 Existing Methodology ................................ ................................ ........................ 39 5.2 Weighted Methodology ................................ ................................ ...................... 40 5.3 C omparison of Existing and Weighted Methodology ................................ ........ 44

PAGE 5

ii CHAPTER 6 STATISTICAL ANALYSIS ................................ ................................ .............. 46 6.1 Comparing the Three Existing Methods of Computing Access Density ................................ ................................ ................................ ................ 47 6.2 Comparing the Two Weighted Methods of Computing Access Density ................................ ................................ ................................ ................ 48 6.3 Comparison of Existi ng and Weighted Methodology ................................ ........ 49 CHAPTER 7 CONCLUSIONS ................................ ................................ ................................ 52 REFERENCES ................................ ................................ ................................ ......................... 55 APPENDICES ................................ ................................ ................................ .......................... 60 APPENDIX A. Straight Line Diagrams ................................ ................................ ....... 61 APPENDIX B. Sample of Access Data Collected for US 19 ................................ ...... 64

PAGE 6

iii LIST OF TABLES Table 1: Number of Online Hits for Access Related Terms ................................ ............... 3 Table 2: Subjective Weights Based on Nature of Access Location ................................ .. 20 Table 3: Subjective Weights Based on Driveway / Intersection Width ............................ 22 Table 4: Details of Three Way Geometry Types in Proposed Weighted Methodology ................................ ................................ ................................ ...... 28 Table 5: Details of Four Way Geometry Types in Proposed Weighted Methodology ................................ ................................ ................................ ...... 29 Table 6: Summary of Equivalent Weights in the Proposed Weighted Methodology ................................ ................................ ................................ ...... 30 Table 7: Access Data Collected for the Selected Roadway Segment ............................... 35 Table 8: Crash Rate of the Selected Roadway Segment ................................ ................... 38 Table 9: Summary of Average Access Density B ased on Existing Methodology ........... 39 Table 10: Comparison of the Three Definitions from Literature Review ........................ 40 Table 11: Effective Weights Assigned in First Approach of Weighted Methodology ................................ ................................ ................................ ...... 41 Table 12: Access Density Calculations Based on First Methodology .............................. 42 Table 13: Access D ensity Calculations Based on Second Methodology (Proposed Methodology) ................................ ................................ ................................ ..... 43 Table 14: Summary of Correlation Results of Access Density Computational Methods ................................ ................................ ................................ .............. 44 Table 15: Access Points Breakdown for Three Existing Methods ................................ ... 48

PAGE 7

iv Table 16: Access Points Breakdown for Two Weighted Methods ................................ ... 49 Table 17: Access Point Breakdown for Definition 1 and Weighted Method 2 ................ 50 Table 18: Summary of Statistical Results ................................ ................................ ......... 51

PAGE 8

v LIST OF FIGURES Figure 1: Access Density Related Terms ................................ ................................ ............ 2 Figure 2: Conflicts in single driveway ................................ ................................ .............. 12 Figure 3 Conflicts in closely spaced driveways ................................ ................................ 12 Figure 4: Access Points Considered are Signalized Intersections and Driveway s Only ................................ ................................ ................................ .................... 15 Figure 5: Access Points Considered are Signalized Intersection and Un signalized Intersections and Driveways Only ................................ ................................ ... 16 Figure 6: Access Points Considered are All Intersections, Driveways and Median Openings ................................ ................................ ................................ ............ 17 Figure 7: Classification of Access Points for First Approach in Calculating Weighted Access Density ................................ ................................ .................. 19 Figure 8: Aerial View of Intersection Types (US 19) ................................ ...................... 21 Figure 9: Aerial View of Different Widths of Driveways (US 19) ................................ .. 22 Figure 10: Four Types of Conflicting Movements ................................ ........................... 26 Figure 11: Conflict Points Associated with Two Way Left Turn Lanes (TWLTL) ......... 31 Figure 13: Study Location ................................ ................................ ................................ 32 Figure 14: Location of Access Points along the Selected Roadway Arterial ................... 34 Figure 14: Trend of Number of Crashes from 2002 2006 ................................ ............. 36 Figure 16: Location of Crash Points along the Selected Roadway Arterial ..................... 37 Figure 16: Final Correlation Results ................................ ................................ ................. 4 5

PAGE 9

vi Figure 18: SLD from Milepost 0 to Milepost 5 ................................ ................................ 61 Figure 19: SLD from Milepost 5 to Milepost 10 ................................ .............................. 62 Figure 20: SLD from Milepost 10 to Milepost 15 ................................ ............................ 63

PAGE 10

vii LIST OF EQUATIONS Equation 1: To Calculate Access Density Based on Definition 1 ................................ .... 16 Equation 2: To Calculate Access Density Based on Definition 2 ................................ .... 16 Equation 3: To Calculate Access Density Based on Definition 3 ................................ .... 17 Equation 4: To Calculate Access Density Based on Weighted Approach 1 ..................... 22 Equation 5: To Calculate Access Density Based on Weighted Approach 2 ..................... 27 Equation 6: To Calculate Equivalent Weight for Second Approach ................................ 31

PAGE 11

viii COMPARISON OF VARIOUS METHODS TO COMPUTE ACCESS DENSITY AND PROPOSING A WEIGHTED METHODOLOGY Meeta Saxena A B STRA CT This study aims to compare three distinct methods used to compute access density and provide a comprehensive weighted methodology to enable standardization for r esearch and application in the future. Access density is a widely used concept that calculates the number of access points within a given distance and has been extensively applied to studies related to crash modeling, operational impact and planning. Metho ds used in past research show that access density is computed differently by different studies and all studies do not include all access points. The weighted methodology proposed takes into account all access points including driveways, intersections and m edian openings and categorizes them into geometric combinations. Each geometric combination have potential number of conflict points which include diverging, weaving, merging and crossing movements depending on the type of access point. Weights were assign ed to each geometry type based on the se conflict point ratio.

PAGE 12

ix In conclusion the study identif ies and compare s methods previously used to compute access density and accordingly, recommends a weighted methodology that includes all access points which can be used as a standard, universal measure all access density related studies including but not limited to safety impacts, operational impacts and planning guidelines.

PAGE 13

1 CHAPTER 1 INTRODUCTIO N 1.1 Background Access management is defined by the access management manual [33] as the systematic control of location, spacing, design and operations of driveways, median openings, interchanges and street connections to a roadway. N umber of ac cess points per mile is represented by access density of the arterial and it potentially measures the performance and conveys the safety and operational impact of the roadway As evident from the definition various access management techniques involve systematic locat ion and spacing of the access points while dealing with fundamental traffic problems of increasing number of vehicular crashes, increased travel times and resultant increased fuel consumption and vehicular emissions. Traditional approach to deal with cong estion and poor level of service is to widen the existing lanes and increase roadway capacity; however these solutions are no t always feasible due to limited land availability and huge capital investments required to purchase right of way for road widening With time any increased capacity follows the land use cycle and eventually results in reduced level of service and thus once again requiring new arterial improvements. Access management techniques break this cycle to achieve improved efficiency by effec tively managing the existing access points and developing guidelines, policies, regulations and geometric design requirements to achieve benefits.

PAGE 14

2 B roadly access points include all openings like driveways, intersections or median openings along the arteri al; however many times they are divided into various terms as e individual impact of these points Various crash prediction models which study safety impact and operating speed models which study operational impact use individual terms which are discussed in detail in chapter 2. Figure 1 : Access Density Related Terms To find the frequency of usage of these terms in technical papers, the number of hits for these individual terms was obtained from an online search done in the reputed Transportation Records (TRR) of Transportation Research Board (TRB) database accessed in August 2009. Table 1 shows the results of number of hits. ACCESS DENSITY Median density Intersection density Driveway density Midblock density Private access density Public access density

PAGE 15

3 Table 1 : Number of O nline H its for A ccess R elated T erms No Term Number of hits in TRR (TRB) 1 Access Density 1039 2 Signal Density 767 3 Median Density 458 4 Driveway density 108 This shows that in technical papers access density term is very popularly used and signal density and median density are also consistently used. Despite the importance of individual terms there are many inconsistencies in defining the computational method of access d ensity which is further discussed in 1.2. 1.2 Problem Statement Access densi ty is used to predict crashes, to study operational effects on travel times, delays vehicle emissions and to develop planning guidelines and standards D ifferent studies include access points based on their research objectives and thus lack a consistent and comprehensive approach in computing the term. In order to accurately evaluate studies related to access density, a standardized universal method is required. There is a need to and compare the various methods and propose a weighted method that can be used as a standard methodology across all access density related research studies including but not limited to safety impacts, operational impacts and planning guidelines 1. 3 Proposed Study This proposed study deals with studying the following aspects related to access density for urban arterial roadways:

PAGE 16

4 1. Conducting a thorough literature review to study the existing me thods of computing access density in crash prediction models and operational impact studies. 2. C apturing the differen ce in types of access points by giving those weights and defining a weighted computational approach which could be consistently used to compute access density. 3. Pe rforming s tatistical tests to test the improvement of the proposed weight ed methodology over the different existing methodologies. 4. Summarizing the steps to be followed for computing the proposed weighted methodology which could be easily used by researchers. Crash rate of an arterial depends on many factors other than access de nsity which are not being included in this study. This study will focus on arriving at a comprehensive methodology to compute access density. 1.4 Approach The approach adopted for the proposed study consists of the following steps: 1. Conducting a detailed l iterature review to study the existing d efinitions and computational methods of defining ac cess density. 2. Select ing an urban arterial whose crash data is available and which has a good number of access points to study safety impacts. 3. Using Google Earth 2009 to collect the access details. 4. Obtaining crash data for the selected roadway arterial 5. Use s traight line diagrams for the selected roadway arterial to join access and crash data.

PAGE 17

5 6. Classify the different access points in a detailed way to be useful in comput ing access density based on various definitions. 7. Assign weights to every access point to arrive at weighted access density 8. Computing the crash rate for the selected arterial. 9. Correlating the crash rate and access densit y obtained from different definitions in previous studies and the proposed weighted access density. 10. Performing statistical tests to capture the improvement in correlation values obtained with the proposed definition. 11. Summarizing the findings in well defined steps which can be used to calculate weighted access density. 1.5 Brief Results 1. Three commonly used definitions of computing access density which were studied are found to be statistically the same. 2. The second approach of calculating weighted access density is proposed because it shows improved correlation with the crash rate and it is statistically different from the existing three computational methods of defining access density. 3. This study does not aim to provide a conclusive methodology ; however i t aims at providing a standard methodology that can be used across all access density related studies including but not limited to safety impacts, operational impacts and planning guidelines.

PAGE 18

6 1.6 Outline of T hesis Chapter 2 talks about various previous studies and their methods of computing access density It also talks about the importance of access density term in safety and operational areas Chapter 3 .1 talks about the three identified methods of computing access density term and chapter 3.2 talks about the need for giving weights to the access points. Chapter 4 explains the study location chosen for the research and description of the access data collected from Google Earth and computation of crash rate for the selected urban arterial. Chapter 5 discussed the re sults based on the access data collected. It includes the three existing methods and the two approaches of weighted access density. Chapter 6 includes the statistical tests performed on the results obtained which are used to draw conclusions compiled in Ch apte r 7 followed by recommendations for researchers. This chapter also summarizes the steps that can be followed to arrive at weighted access density and states its advantages over other definitions.

PAGE 19

7 CHAPTER 2 LITERATURE REVIEW Literature review in this chapter consists of t wo parts. The first part covers the application of access density term in various crash prediction models. Much research has been done in this area and it is difficult to cover the whole body of literature. The second part of the literature is a review of previous studies related to operational impact of access management techniques. The methodology and framework will be further discussed in detail in Chapter 3. 2.1 Crash M odeling R elated R esearch Access density is a commonly used independent variable in various crash predicting models. There are various studies [ 1 8, 10 ] which use access density as a significant independent variable to identify the relationship between safety (crash rates / crash frequency) and various roadway characteristics. Various access data collecting methods have been used in several studies which include using ESRI ArcGIS [ 4 ] and more specifically ArcGIS orthophotos [ 12 ] s tate photo log s [ 5 ,16 ] state video log database [ 1 0 ] and state data collection documents, maps and road viewer program in addition to Google maps [ 13 ] Many studies have found that as the density of access points increases, the accident rate also increases [ 1, 2 ] which is not a surprising finding [ 15 ] given that the more number of access points there are in a segment, the higher the number of conflicting movements This trend holds true regardless of the median type [ 3 ] however inclusion of a particular

PAGE 20

8 type of median opening has an impact on crash rat e [ 3, 11 ] The before after study done by (3) investigated the crash data and found reduction in crash rates after installation of raised median. Contradictory to this finding it was observed that it is not necessary to detect significant difference s in s afety between urban highways with and without a median if the median has openings at intersections [ 10 ] A l ot of inconsistency has been observed in defining the access used in model s I t is found that various access points are excluded while co mputing access density in different studies A study by [ 7) ] defines access density as number of driveways (on both sides of the road) per mile. With an assumption that driveways include intersections it can be seen that median openings are not considered in the density calculations. Study [ 10 ] defines access density as the number of access points per km and provides no details regarding the inclusion of all access points namely driveways, intersections and median openings Study by [ 8 ] found that driveway density, un signalized public street density and median type are significantly correlated with accident frequency. A study by [ 5 ] describes access as the number of driveways and minor intersections and further categorizes driveways based on land use including residential, office, retail and industrial. Th is definition excludes access points such as median openings and major intersections. Overall these studies not only sh ow inconsistencies in computing access density but they also present hinders the purpose of such research T here is no mention of any criteria used to distinguish between dr iveways and minor intersections [ 5 ]

PAGE 21

9 As noted by the American Ass ociation of State Highway and Transportation Officials (AASHTO) ( 1994, page 793), dri veways create intersections with the street system. Th ere is a need to clarify what stand we take in incorporating driveways while computing access density. We need to consider if access points should be differentiated as driveways and intersections or whether fundamentally both must be considered as intersection s for the purpose of computing acce ss density. Based on the specif ic research objective s of different stud ies terms related to access are defined differently [ 4, 6 ] A study in a metropolitan area focused on commercial driveway density since they generate more trips compared to residential driveways [ 4 12 ] A s tudy on the safety of curbs [ 6 ] included active un signalized intersections and driveways larger than those built for a single family residential house. Stud y [ 9 ] focused on the effects of midblock access points and thus used the term access density to represent access points in between intersections. In order for this term to include all midblock access points, median opening access points in between intersections should have been included as well. In studying the effect s of median treatment and access for rural highways [ 20 ] crashes were grouped based on low access of less than 20 points; medium access with in the 20 40 point range and high access with more than 40 points. From this study too, it can be observed that t he re is no mention of inclusion of median openings as access points. A significant contribution towards assigning weights based on the type of intersection (four legged or T intersection) and presence of a traffic signal was found in only one study [ 10 ] whic h defined access density with driveway densities and public street intersections terms. Signalized intersections (four legged or T intersection)

PAGE 22

10 w ere assigned an equivalent weight of two access points because of increased rear end crashes at traffic signal s I n the case of un signalized four legged intersection s, the equivalent weight was maintained as two due to the presence of access points on both sides of the main arterial and a three way un signalized intersection was considered as only one access poin t. Each of these access points provide a gross measure of the relative amount of conflict opportunities [ 6 ] caused either by diverging, merging, conflicting or crossing movements around the access point. Access density changes may come from increases because of land development or from decrease s due to driveway consolidations or land re development.

PAGE 23

11 2.2 Operational Impact Related Research Previous studies have demonstrated that levels of access density are associated with different speed ranges. [ 14 ] sho ws a strong rel ationship between access density and the 85 th percentile speed L ower access densities ar e associated with higher speeds so the gr eater the frequency of access points, the larger the speed reduction to through traffic [ 29, 30 ] Access density is defined by [ 14 ] as the number of access points (driveways and intersections) per mile. Most other studies have used the term access density without clear ly stating the access points included in the term. While studying the operational impacts this study [ 14 ] measure s the access density between features that could control the speeds along the section (e.g., signal, etc.) however ; c ounting of median access po ints is not included in this definition. Studies have shown that un signalized access density, s ignalized access density and median type certainly affect corridor operations and there is a need to understand how these operational impacts change with change in geometric characteristics [ 25 ] Although there is an abundance of research on signalized intersections and th eir effects on through traffic, little study has explored the effects of access points on the operations of urban streets between signalized i ntersections. There is a need to explore various methods by which the operational impact of all access points can be studied. A h igh frequency of closely spaced access points can have a substantial impact on through traffic

PAGE 24

12 [ 24 ] Figures 2 and 3 show the i ncrease in the number of conflict points when driveways are closely spaced. Figure 2 : Conflicts in single driveway Figure 3 Conflicts in closely spaced driveways Source: Statewide U rban Design and Specifications Design Manual C hapter 5.1 While estimating vehicular emissions by capturing traffic variations, access density was found as a significant variable [31, 32]. 32 As the number of unrestricted vehicular property access increase there is an increase in vehicle conflict points which has shown result in increases in delays, crash rates and vehicle emissions. Thus accurate measurement of access points and managing them will help save fuel and reduce vehicle emissions which in turn will ensure efficient progression of through vehicles along major roadways, help maintain desired progression speeds and reduce propensity for start and stop traffic operations due to vehicle t urning conflicts [32] As can be seen from the literature presented above, past research indicates that there is a high level of inconsistency in incorporating all access points to compute access density

PAGE 25

13 The difference in the operational definition of a ccess points is highly dependent on each individual study making it difficult to further research in the area in a consistent and standardized manner. The literature indicates the re is a lack of an universal methodology which encompasses all access points and can be applied to all access density research studies.

PAGE 26

14 CHAPTER 3 METHODOLOGY The a ccess density term and its application were briefly introduced in C hapter 1. In this chapter part 3.1 will describe the three most common methods by which access density has been calculated in the existing studies P art 3.2 of thi s chapter further discuss two weighted methodologies for computing access density. The first approach in giving weights to access points involved a combination of subjective judgment and weight assigned based on conflict point ratio of the geometry of access opening. This attempt demonstrated certain limitations such as multiple weighting of same access p oint which are discussed later in more detail. I mproving on this first weighted approach a second weighted approach is proposed which is statistically different from existing computational methods and show s improved correlation to crash data. Based on all the existing and weighted methodologies defined access data was collected using G oogle E arth and tables were generated to arrive with access numbers for each definition T he next step is to check the correlation of access points with crash points. Further statistical non parametric tests are conducted to test the hypothesis that all methods of calculating access density are same. 3.1 Existing Computational Methodology As seen in chapter 2, past researches point towards existing inconsistency in incorporating all access points to compute access density Depending on the research objective studies have included only certain access points important for their purpose and

PAGE 27

15 use o f the study area like commercial driveways, public access driveways, residential driveways and others. However while studying the computation methodologies in existing studies three scenario of computing access points are compared as described in following parts. 3.1.1 Considering S ignalized I ntersection and D riveways O nly Figure 4 : Access P oints C onsidered are S ignalized I ntersections and D riveways O nly Many research studies ha have proved that number of signalized intersections per mile have a huge impact on safety which results in increased crash incidence. Driven by the importance of the signalized density many studies include midblock driveway points in addition to signalize d intersections in their access density computation methods. As a result the un signalized intersection access points and median access points are not considered in the method. This is the first scenario of calculating access density from existing methodol ogies and is For computational use the following equation is used to arrive at access density corresponding to definition 1

PAGE 28

16 Equation 1 : To C alculate Access Density Based on Definition 1 3. 1 2 Considering B oth I ntersection s and D riveways O nly Figure 5 : Access P oints C onsidered are S ignalized Intersection and U n signalized I ntersections and D riveways O nly In th e methodology adopted in this scenario, no distinction is made b ased on presence of traffic signal thus all intersections (signalized or un s ignalized ) and driveways in both directions of travel are included in access points while computing access density The median openi ngs are however not considered even in this methodology. This is the second scenario of calculating access density from exis ting methodologies and is referred For computational use the following equation is used to arrive at access density corresponding to definition 1. Equation 2 : To Calculate Access Density Based on Definition 2 Access density = (No of intersections (signalized or un signalized) +No of driveways) Length of Segment Access density = (No of signalized intersections + No of driveways) Length of Segment

PAGE 29

17 3.1.3 Considering All I ntersections, D riveways and M edian O penings Figure 6 : Access Points Considered are All Intersections, Driveways and Median O penings This is the last methodology which include s all the possible access points on an arterial segment. All components of access points namely signalized intersections, un signalized intersections, median openings and driveways on both sides of the arterial are included. This is the last scenario of calculating access density from existing methodologies and is 3 statistical tests and conclusions. For computational use the following equation is used to arrive at access density corresponding to definition 3 Equation 3 : To Calculate Access Density Based on Definition 3 Access density = ( All intersections + No of driveways + No of median openings) Length of Segment

PAGE 30

18 3.2 Weighted Computational Methodology The third scenario defined in previous part includes all the access points on the arterial however literature has shown the differences in conflict points for different access openings. It can be clearly understood that impact of single driveway versus imp act of a four way intersection could be quite different. The width of driveway opening could potentially have different impact on safety and operational aspects of roadway T o capture the relative difference in these impacts, an attempt has been made to as sign weights to different access points based on certain criteria and use it to calculate weighted access density. Two main approaches have been adopted to arrive at the weighted access density and further statistical test are performed to test the improve ment from existing methodologies. 3.2.1 Combination of S ubjective Judgment and C onflict P oint R atio The first approach to classify the access points i s based on combination of subjective judgment and conflict point ratio of the geometry of the access opening. Various features related to access points are studied and relevant features are identified which could be given weights Type of intersection (fou r way or three way), presence or absence of traffic lights, type of median openings and width of driveway openings are identified to assign weights. S ubjective judgment is used to assign weights based on driveway width and nature of access points as shown in tables 2 and 3 Figure 7 represents the classification of access points which are identified to assigned weights in this first approach in calculating weighted access density.

PAGE 31

19 Figure 7 : C lassification of A ccess P oints for First Approach in Calculating Weighted Access Density

PAGE 32

20 Conflict points are highly correlated with crash rates and are considered as the basis to assign weights to four wa y and three way intersections. As stated in literature review there is an increased possibility of rear end crashes at signalized f our way and three way intersections Intersection area witness higher crashes due to vehicular movements from two driveways a nd is subjectively weighted higher than single driveway access point. Additionally presence of traffic signal at an intersection results in even higher number of conflicts for proving a subjective weight they are rated even higher. Median openings witness conflicting movements due to change in directions and thus assigned an intermediate weight of 2 between driveway weight of 1 and signalized intersection weight of 3. Un signalized intersections placed intermediate and assigned equivalent weighted of 2. Table 2 : Subjective W eights B ased on N ature of A ccess L ocation No Description Subjective Weights 1 Driveways 1 2 Median Openings 2 3 Un signalized Intersection 2 4 Signalized Intersection 3 Ratio of conflict points between four way intersection and three way intersections are considered to capture the difference in access facilitated by their geometrical

PAGE 33

21 configuration. Thus 32: 9 is reduced to 3.3:1 which reasonably captures the desired safety impact Figure 8 shows a n Four way intersection and three way intersection in the study area of US 19. Four way intersection with 32 conflict points Three way intersection with 9 conflict points Figure 8 : Aerial View of Intersection Types (US 19) Another observation was effect of width of driveway opening to safety impact of the arterial segment. Generally wider driveways are an indication of large amount of traffic flow at the access point which indicates larger chance of the crash possibility. Vo lume of driveway opening could be more appropriate to assign weight for increased crash possibility but it is very difficult to obtain the volume data of various driveway accesses and hence measuring the ir width and correspondingly applying weights worked out to b e practical. Broadly the following categories of driveway widths and their corresponding equivalent weight were used in this first approach of arriving at weighted access density. Figure 9 shows driveways with different widths.

PAGE 34

22 Table 3 : Subjective Weights B ased on Driveway / Intersection W idth Driveway / Intersection Width Equivalent Weight 0 to 25 (feet) 0.5 25 to 50 (feet) 1.0 50 to 75 (feet) 1.5 75 to 100 (feet) 2.0 > 100 (feet) 2.5 Figure 9 : Aerial View of Different W idths of D riveways (US 19) The following e quation summarizes the assignment of weights to compute access density adopted in approach 1. Equation 4 : To Calculate Access Density Based on Weighted Approach 1 Access Density = (3 x Signalized Intersection + 2 x Un signalized Intersection + 2 x Medians + 1 x Driveways) / no. of miles

PAGE 35

23 Where : Signalized Intersection = 3.3 (If four way intersection) = 1.0 (If three way intersection) Driveways = 0.5 when width = 0 25 ft = 1.0 when width = 25 50 ft = 1.5 when width = 50 75 ft = 2.0 when width = 75 100 ft = 2.5 when width > 100 ft feet First approach was a preliminary attempt to arrive at weighted methodology which assigned weights to different access points based on above mentioned criteria however it has drawbacks. One of the major drawbacks of thi s approach was use of subj ective judgment which needs to be supported by other findings. Also there was excessive multiple weighting of any access point based on different aspects For example, t he driveway openings are weighted 1 and further weighted depen ding on their width. Similarly median openings and intersection openings were weighted depending on pres ence /absence of traffic signal and type of intersection. While conflict points were used to assign weights to four way intersections and three way int ersections it is seen that the numbers of conflict points are very sensitive to change in median type associated with the intersections. These detailed conflict points were not considered in this approach.

PAGE 36

24 There are many types of directional median openings with different number of conflict points which have different impacts on the traffic movement. First approach fails to deal with giving subjective weights to two way left turn lanes (TWLTL) which is an ad ditional drawback. A notable inconsistency in the classification of access points as in table 7 is distinction between the terms driveways and intersections F undamentally every driveway creates an intersection with the main arterial. Thus we need to clari fy what stand we take in incorporating intersections while computing access density. We need to decide if access points should be differentiated as driveways and intersections or fundamentally considered as intersection only for access density computation purpose. This method talks about one way of assigning subjective weights to various access points however several combinations of weights could be tried to observe changes in correlation values to crash data. Although driveways are assigned a subjective weight 1 we can argue that the traffic movements on driveways are un controlled as compared to signalized intersections and thus driveways could be potentially more unsafe and should be weighted higher than signalized intersections. Thus effectively we cou ld recalculate the access density with different weights and see how best they correlate crash rate. A detailed sensitivity analysis could result in more appropriate and statistically significant weights however the drawbacks of subjective judgment, multip le weighting and incapability of explaining special access openings like TWLTL do not favor use of this methodology.

PAGE 37

25 Keeping in mind these drawbacks and concerns regarding use of intersection and driveway terminologies a second approach with no subjective judgment is defined The following part 3.2.2 deals with the second approach of assigning weights by classifying all the possible access points into well defined types. 3.2.2 Proposed W eighted M ethodology B ased on D etailed G eometry T ypes Considering the l imitations noticed in the first approach to assign weights to different components while defining access density, there was a need to come up with an objective method of assigning weights. One major inconsistency while computing access density arises when Intersections are driveways and thus for access density calculation purpose they should not be split in terms of driveway density and intersection density. Once driveways and intersections are not considered as separate terms one can define all possible geometry types as a combination of intersections and median openings. Intersections can further be thr ee way (driveway on only one side) and four way (driveway on both sides ). With this as basis and elimi nating any subjective component, the basic geometry types with all possible permutation and combinations of access locations (say four way or three way) and median types (raised / un divided / directional / TWLTL ) are defined in detailed Once the geometry was identified conflict points associated with the geometry are worked out The following conflict points are obtained for each geometry type.

PAGE 38

26 Figu re 10 : Four Types of Conflicting Movements T able 4 describes basic five type s of three way geometry type and Table 5 describes basic five types four way geometry type. Different type of median arrangements are incorporated in thes e types and based on the conflict points of the geometry they are assigned equivalent weight All the three major components of access density namely, driveways, intersections and median openings are incorporated in these geometry types. Type s 1 to type 5 are the most common geometries associated with three way intersection (driveways are considered as intersection). Similarly types 6 to type 10 are the commonly found geometries associated with four legged intersections. Type 1 which represents a typical single access opening with full median access is considered as base condition and assigned a n equivalent weight of 1. The equivalent weights of all other types are calculated with type 1 as base and are summarized in table 6 (Page 31).

PAGE 39

27 Tables 4 an d table 5 describe the geometry of each access type, their median opening number of conflict points and calculated equivalent weights. Once the weights are obtained the next step is to find the number of these types in the selected roadway arterial. Usin g the equivalent weights and number of each type, f ollowing equation can be used to arrive at the weighted access density. Equation 5 : To Calculate Access Density Based on Weighted Approach 2 Weighted access density = [(1 x # T ype 1) + (2.2 x # Type 2) + (0.2 x # Type 3) + (0.6 x # Type 4) + (0.6 x # Type 5) + (3.6 x # Type 6) + (0.4 x # Type 7) + (0.8 x # Type 8) + (0.8 x # Type 9) + (0.1 x # Type 10) + (W x1 x #Type X 1 ) + (W x2 x # Type X 2 )] / Length of the segment.

PAGE 40

28 Table 4 : Details of T hree W ay G eometry T ypes in P roposed W eighted M ethodology TYPE 1 TYPE 2 TYPE 3 TYPE 4 TYPE 5 Number of lanes = 2 Median type = Undivided Access = single entrance Number of lanes = 2 Median type = Undivided Access = closely spaced entrance Number of lanes = 2 Median type = Raised Access = single entrance Number of lanes = 2 Median type = Raised Access = left turn egress only from intersection or driveway Number of lanes = 2 Median type = Raised Access = left turn ingress only into driveway or driveway Conflict Points = 9 Conflict Points = 20 Conflict Points = 2 Conflict Points = 5 Conflict Points = 5 Weighted Access Equivalent = 1 Weighted Access Equivalent = 2.2 Weighted Access Equivalent = 0.2 Weighted Access Equivalent = 0.6 Weighted Access Equivalent = 0.6

PAGE 41

29 Table 5 : Details of F our W ay G eometry T ypes in P roposed W eighted M ethodology TYPE 6 TYPE 7 TYPE 8 TYPE 9 TYPE 10 Number of lanes = 2 Median type = Undivided Access = typical four way intersection or driveway Number of lanes = 2 Median type = Raised Access = typical four way intersection or driveway Number of lanes = 2 Median type = Raised Access = left turn egress only from intersection or driveway Number of lanes = 2 Median type = Raised Access = left turn ingress only into driveway Number of lanes = 2 Median type = Raised Access = left turn into driveways from both direction lanes Conflict Points = 32 Conflict Points = 4 Conflict Points = 7 Conflict Points = 7 Conflict Points = 10 Weighted Access Equivalent = 3.6 Weighted Access Equivalent = 0.4 Weighted Access Equivalent = 0.8 Weighted Access Equivalent = 0.8 Weighted Access Equivalent = 1 .1

PAGE 42

30 Table 6 : Summary of Equivalent W eights in the P roposed W eighted M ethodology Equivalent Weight Type 1 1 Type 2 2.2 Type 3 0.2 Type 4 0.6 Type 5 0.6 Type 6 3.6 Type 7 0.4 Type 8 0.8 Type 9 0.8 Type 10 1 .1 *Types correspond to the description provided in table 4 and table 5 A lthough these ten types constitute the primary set of geometry types commonly found in urban arterials there is always room for other access types With advancement in median treatments and adoption of newer configurations of medians one would have to calculate their conflict points and reduce it using the equation and obtain its equivalent weight This second approach of assigning weights to access points provides flexibility to incorporate additional geometr y types and arriving at the weighted access equivalent by diving their conflict points by

PAGE 43

31 Equation 6 : To Calculate Equivalent Weight for Second Approach Equivalent weigh t = Number of conflict points in the given geometry 9 Although the study area of US 19 does not have any Two Way Left Turn Lane (TWLTL) we can come up with its equivalent weight by calculating the number of conflict points as shown in the Figure 1 1 As TWLTL has 30 conflict points and based on the formula above it is assigned e quivalent access weight of 3.33. Figure 11 : Conflict P oints A ssociated with Two W ay L eft T urn L anes (TWLTL) Thus the weighted access density can be obtained which can then be used to correlate with the crash rates. These correlations are the key to draw conclusions.

PAGE 44

32 CHAPTER 4 DATA COLLECTION 4.1 Study Location Figure 12 : Study Location The selected roadway segment is the US 19 in the Pinellas County, Florida The most commonly seen access points along this major arterial are driveways with commercial land us e Length of enti re arterial is of 32 miles which is a 3 + 3 multilane arterial in FDOT d istrict 7. The posted average spee d limit on US 19 is 55 mph. It is classified under urban and other principal arterial. For purpose of studying the access points, 15 miles stretch of road in South

PAGE 45

33 Pinellas was included. Appendix A shows the straight line diagram of the selected 15 miles obtained from FDOT. These straight line diagrams were used to connect the access data w ith crash data using mile post as the common variable. 4.2 Data Collection 4.2.1 Access Data Access d etails for the selected roadway segment are obtained from Google Earth and Google Maps (See appendix B) The access deta ils are represented in GIS maps as seen in figure 15. About 420 access points are identified in the 15 miles of selected roadway.

PAGE 46

34 Figure 13 : Location of A ccess P oints a long the Selected Roadway Arterial

PAGE 47

35 Table 7 : Access D ata C ollected for the Selected R oadway S egment Mile Post Four way intersection Three way intersection Driveway Median Start End Signalized Un signalized Signalized Un signalized N bound S bound Total 0.87 1 1 0 0 0 1 4 0 1 2 4 4 0 4 27 16 1 2 3 2 5 0 2 17 27 0 3 4 2 4 0 11 31 28 1 4 5 5 8 0 4 42 48 2 5 6 3 11 0 3 19 25 0 6 7 2 7 0 6 32 27 0 7 8 1 1 0 3 19 18 1 8 9 4 5 0 1 28 26 1 9 10 3 0 0 0 24 18 4 10 11 1 0 0 0 15 7 5 11 12 1 3 0 0 11 7 2 12 13 1 0 0 0 14 11 0 13 14 0 0 0 0 7 10 0 14 15 0 0 0 0 11 3 0 4.2.2 Crash Data Data for the number of crashes and annual average daily traffic ( AADT ) along arterial is collected for five years from 200 2 to 2006 It is very important to have good quantity and fairly consistent crash data before us ing it for correlating with access points. The five year data is checked to make sure it is free from any obvious abnormalities in occurrences along the mileposts of the selected roadway. Figure 13 shows that numbers of crashes in the five years do es not show lot of variation in trend and hence the summation of crashes can be used to arrive at crash rates of the selected roadway segment.

PAGE 48

36 Figure 14 : Trend of N umber of C rashes from 2002 2006 A total of 3264 number of crashes were used to arrive at crash rate of the selected roadway segment. These are located over the entire stretch of the urban arterial under study. 0 100 200 300 400 500 600 700 800 900 1000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Sum Of Crashes 2006 2005 2004 2003 2002

PAGE 49

37 Figure 15 : Location of Crash Points along the Selected Roadway Arterial

PAGE 50

38 Crash rates are calculated for each mile as a function of number of crashes, volume and length in terms of crashes per million vehicle miles traveled. Where : 1. R is crash rate for the section (in crashes per MVMT) 2. A is the number of report ed crashes 3. T is time frame of the data (years) 4. V is AADT of the road section 5. L is the length of the segment (miles) Table 8 : Crash Rate of the S elected Roadway Segment Mile Post No Of Crashes A ADT Crash Rate ( crashes per M VMT) Start End 0.87 1 26 26500 4.14 1 2 147 27345 2.45 2 3 192 34140 2.57 3 4 217 42500 2.33 4 5 354 42435 3.81 5 6 286 42420 3.08 6 7 305 45500 3.06 7 8 294 45500 2.95 8 9 439 48305 4.15 9 10 405 67775 2.73 10 11 100 69500 0.66 11 12 171 69500 1.12 12 13 277 69500 1.82 13 14 35 44450 0.36 14 15 16 44000 0.17

PAGE 51

39 CHAPTER 5 DATA ANALYSIS 5.1 Existing Methodology Data was a naly zed b y applying the proposed methodology using the collected access and crash data obtained from previous chapters Following table summarizes the resultant average access densities based on the three existing methods as discussed in chapter 3.1. The entire data set of access points and crash points has been broken down into 15 points of one mile. T able 9 : Summary of A verage A ccess D ensity B ased on E xisting M ethodology Mile Post Definition 1 Definition 2 Definition 3 Start End 0.87 1 4 4 4 1 2 23 31 32 2 3 28 35 35 3 4 27 42 43 4 5 61 73 75 5 6 13 27 27 6 7 31 44 44 7 8 28 32 33 8 9 38 44 45 9 10 39 39 43 10 11 21 21 26 11 12 11 14 16 12 13 24 24 24 13 14 17 17 17 14 15 14 14 14 Average access densities: 2 7 33 3 4 points/ mile points/ mile points/ mile Correlation of access density and crash rates is performed to arrive at correlation values for each definition. The R square values obtained are summarized in the following table

PAGE 52

40 Table 10 : Comparison of the T hree D efinitions from L iterature R eview No Access Points included Access Density Correlation Relative improvement (%) Comments 1 Both direction driveways + Signalized Intersections 27 0.42 Nil Base condition 2 Both direction driveways + Signalized Intersections + Un signalized Intersections 33 0.50 19 % Improvement from base condition 3 Both direction driveways + Signalized Intersections + Un signalized Intersections + Median Openings 34 0.48 14% Improvement from base condition 5 2 Weighted Methodology There are two methods which a re adopted to arrive at weighted access density The first method is a combination of subjective and objective judgment which assigned the weights to the access components as discussed in chapter 3.2 Table 11 shows the effective weights which are used in access density calculations.

PAGE 53

41 Table 11 : Effective W eights A ssigned in First Approach of W eighted M ethodology Four way intersection Three way intersection Driveway Median Signalized Un signalized Signalized Un signalized 0 25 25 50 50 75 75 100 > 100 Subjective 3 2 3 2 0.5 1 1.5 2 2.5 2 Objective 3.3 3.3 1 1 NET 9.9 6.6 3 2 0.5 1 1.5 2 2.5 2 Tables 12 and 13 show the calculations to arrive at average weighted access density by the two approaches. Once the weighted access density is calculated correlation of weighted access density and crash rates are performed.

PAGE 54

42 Table 12 : Access D ensity C alculations B ased on F irst M ethodology Mil e Post Four way intersection Three way intersection Driveway Median Opening Access Density Start End Signalized Un signalized Signalized Un signalized 0 25 25 50 50 75 75 100 > 100 0.87 1 9.9 0 0 0 0.5 2 0 2 0 0 14.4 1 2 39.6 26.4 0 8 4.5 20 1.5 0 0 2 102 2 3 19.8 33 0 4 5.5 18 4.5 2 0 0 86.8 3 4 19.8 26.4 0 22 13 21 1.5 0 0 2 105.7 4 5 49.5 52.8 0 8 10 37 10.5 0 2.5 4 174.3 5 6 29.7 72.6 0 6 4 18 4.5 0 0 0 134.8 6 7 19.8 46.2 0 12 7 21 3 2 0 0 111 7 8 9.9 6.6 0 6 3.5 13 3 2 0 2 46 8 9 39.6 33 0 2 1.5 27 3 0 0 2 108.1 9 10 29.7 0 0 0 3 16 6 0 0 8 62.7 10 11 9.9 0 0 0 0 11 6 4 0 10 40.9 11 12 9.9 19.8 0 0 0.5 12 6 0 0 4 52.2 12 13 9.9 0 0 0 0.5 15 6 0 2.5 0 33.9 13 14 0 0 0 0 1.5 8 4.5 2 0 0 16 14 15 0 0 0 0 0.5 10 3 0 0 0 13.5 AVERAGE ACCESS DENSITY BY APPROACH 1 = 78

PAGE 55

43 Table 13 : Access D ensity C alculations B ased on S econd M ethodology (Proposed Methodology) Mile Post Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Type 8 Type 9 Type 10 Access Density Start End CONFLICTS 9 20 2 5 5 32 4 7 7 10 WEIGHTS 1.0 2.2 0.2 0.6 0.6 3.6 0.4 0.8 0.8 1.1 Wt Wt Wt Wt Wt Wt Wt Wt Wt Wt 0.87 1 0 0 0 0.0 3 0.7 0 0 0 0.0 1 3.6 0 0.0 0 0 0 0.0 0 0.0 4.2 1 2 4 4 1 2.2 13 2.9 0 0 0 0.0 8 28.4 3 1.3 0 0 1 0.8 0 0.0 39.7 2 3 2 2 1 2.2 18 4.0 0 0 0 0.0 7 24.9 5 2.2 0 0 0 0.0 0 0.0 35.3 3 4 11 11 0 0.0 28 6.2 0 0 0 0.0 6 21.3 2 0.9 0 0 1 0.8 0 0.0 40.2 4 5 4 4 0 0.0 34 7.6 0 0 0 0.0 13 46.2 12 5.3 0 0 2 1.6 0 0.0 64.7 5 6 3 3 0 0.0 10 2.2 0 0 0 0.0 11 39.1 5 2.2 0 0 0 0.0 0 0.0 46.6 6 7 6 6 0 0.0 10 2.2 0 0 0 0.0 9 32.0 13 5.8 0 0 0 0.0 0 0.0 46.0 7 8 3 3 0 0.0 9 2.0 0 0 0 0.0 2 7.1 8 3.6 0 0 1 0.8 0 0.0 16.4 8 9 1 1 0 0.0 9 2.0 0 0 0 0.0 9 32.0 12 5.3 0 0 0 0.0 1 1.1 41.4 9 10 0 0 0 0.0 9 2.0 0 0 0 0.0 1 3.6 12 5.3 0 0 0 0.0 4 4.4 15.3 10 11 0 0 0 0.0 8 1.8 0 0 1 0.6 1 3.6 3 1.3 0 0 1 0.8 3 3.3 11.3 11 12 0 0 0 0.0 12 2.7 0 0 0 0.0 3 10.7 0 0.0 0 0 0 0.0 2 2.2 15.6 12 13 0 0 0 0.0 17 3.8 0 0 0 0.0 1 3.6 3 1.3 0 0 0 0.0 0 0 .0 8.7 13 14 0 0 0 0.0 13 2.9 0 0 0 0.0 0 0.0 2 0.9 0 0 0 0.0 0 0 .0 3.8 14 15 0 0 0 0.0 12 2.7 0 0 0 0.0 0 0.0 1 0.4 0 0 0 0.0 0 0.0 3.1 AVERAGE ACCESS DENSITY BY APPROACH 2 = 28

PAGE 56

44 5 3 Comparison of E xisting and W eighted M ethodology Table 14 : Summary of Correlation R esults of A ccess D ensity C omput ational M ethods Various Definitions Access points included Access Density Correlation values with crash rate Relative Improvement (%) Definition 1 Both direction driveways + Signalized Intersections 27 0.42 Base condition Definition 2 Both direction driveways + Signalized Intersections + Un signalized Intersections 33 0.50 19% Definition 3 Both direction driveways + Signalized Intersections + Un signalized Intersections + Median Openings 34 0.48 14% Weighted Method 1 Weights based on subjective judgment + Conflict point 78 0.58 38% Weighted Method 2 Weights based on geometry types based on only conflict points 28 0.59 40% Weighted methods for computing access density show an improvement of about 40% from the definition 1 of existing methods which includes signalized intersections and midblock driveway openings only. Provi ding weights to access points certainly improve s the correlation values between access density and crash rates. There is not much difference between correlation values of

PAGE 57

45 weighted method 1and weighted method 2 values; however weighted method 2 has clear benefits as compared to the drawbacks found in weighted method 1. Figure 16 : Final Correlation Results 0.42 0.5 0.48 0.58 0.59 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Definition 1 Definition 2 Definition 3 Weighted Method 1 Weighted Method 2 Correlation values of access density with crash rates

PAGE 58

46 CHAPTER 6 STATISTICAL ANALYSIS While comparing the access density values obtained by the five methodologies explained in previous chapters it is very important to check if they are significantly different from each other and i f the improvement is worth the effort. Descriptive statistics of numerical methods and graphical representation can be used to summarize and interpret data. N umerical methods have an advantage over graphical representation methods as they provide precise and objective determined values that can be easily manipu lated, interpreted and compared Thus numerical method has been used to statistically test the results obtained in chapter 5. Hypothesis testing is used to determine whether the differences in access density have arisen by chance or whether some other factor is responsible for the difference. To formulate a hypothesis testing, two competing statistical hypothesis namely null hypothesis and alternative hypothesis are formulated. Null hypothesis is an assertion about a population parameter that is assumed to be true until there is sufficient statistic evidence to conclude otherwise. Based on the objective to determi ne if the various definitions are same or different a null hypothesis is defined. The purpose of defining the hypothesis is that when it gets rejected it shows that the definitions are different

PAGE 59

47 Kruskal Wallis is used to te st the effect of a ccess densi ty obtained by various definitions. This test ha s been performed three times to test the following aspects. 1. To test if the existing three methods of computing access density are statistically different. 2. To test if the two weighted methods of computing access density are statistically different. 3. To test if the definition 3 (which counts all access points) and weighted method 2 (which in comparison has the best correlation) are statistically different. This non parametric test was performed with 9 5 % leve l of confidence. For performing a Kruskal Wallis test n ul l hypothesis to be tested i s formulated. The observations to be tested are then ranked for each group and sum of their ranks is obtained. Test statistic is then calculated using the sum of ranks of e ach group. Critical chi square value is obtained for k 1 degrees of freedom. Test statistic is further compared to critical statistic and inferences are drawn about the null hypothesis. 6.1 C omparing the T hree E xisting M ethods of C omputing A ccess D ensity A n on p arametric Kruskal Wallis t est is conducted with the following steps to arrive at conclusions. 1. Null Hypothesis (Ho) The three existing methods of computing access density are the same. 2. Calculated test statistic = 2. 0199 3. Level of confidence = 9 5 % 4. C ritical Statistic = 5.9915 5. As Test Statistic < Critical Statistic, we have no evidence to reject Null Hypothesis Conclusion : The three methods of computing access density are not significantly different from each other.

PAGE 60

48 Table 15 : Access P oint s B reakdown for T hree E xisting M ethods Observation No Definition 1 Definition 2 Definition 3 1 4 4 4 2 23 31 32 3 28 35 35 4 27 42 43 5 61 73 75 6 13 27 27 7 31 44 44 8 28 32 33 9 38 44 45 10 39 39 43 11 21 21 26 12 11 14 16 13 24 24 24 14 17 17 17 15 14 14 14 6.2 Comparing the T wo W eighted M ethods of C omputing A ccess D ensity A n on p arametric Kruskal Wallis t est is conducted with the following steps to arrive at conclusions. 1. Null Hypothesis (Ho) T wo weighted methods of computing access density are the same. 2. Calculated t est Statistic = 8.19 3. Level of confidence = 9 5 % 4. Critical Statistic = 3.8415 5. As Test Statistic > Critical Statistic, we reject Null Hypothesis Conclusion : The t wo weighted methods of computing access density are significantly different from each other.

PAGE 61

49 Table 16 : Access P oints B reakdown for T wo W eighted M ethods Observation No Weighted Definition 1 Weighted Definition 2 1 14.4 4.2 2 102 39.7 3 86.8 35.3 4 105.7 40.2 5 174.3 64.7 6 134.8 46.6 7 111 46.0 8 46 16.4 9 108.1 41.4 10 62.7 15.3 11 40.9 11.3 12 52.2 15.6 13 33.9 8.7 14 16 3.8 15 13.5 3.1 6.3 Comparison of Existing and Weighted Methodology A n on p arametric Kruskal Wallis t est is conducted with the following steps to arrive at conclusions. 1. Null Hypothesis (Ho) Access density calculated by weighted method is same as a ccess density calculated by existing methods 2. Calculated test statistic = 4 .723 3. Level of confidence = 9 5 % 4. Critical Statistic = 3.8415 5. As Test Statistic > Critical Statistic, we reject Null Hypothesis Conclusion : The two methods of computing a ccess density are significantly different from each other with LOC =9 5 %

PAGE 62

50 Table 17 : Access Point Breakdown for Definition 1 and Weighted Method 2 Observation No Definition 3 Weighted Definition 2 1 4 4.2 2 32 39.7 3 35 35.3 4 43 40.2 5 75 64.7 6 27 46.6 7 44 46.0 8 33 16.4 9 45 41.4 10 43 15.3 11 26 11.3 12 16 15.6 13 24 8.7 14 17 3.8 15 14 3.1 Thus in conclusion the three Kruskal Wallis statistically confirms the main result that the weighted access density calculated by second approach is significantly different from existing definition and it not a matter of chance that weighted methodo logy access density values appear to be improved.

PAGE 63

51 Table 18 : Summary of Statistical Results No Null Hypothesis: Kruskal Wallis Result 1 Three existing methods of computing access density are the same No evidence to reject null hypothesis 2 Two weighted methods of computing access density are the same Reject null hypothesis 3 Definition 3 of existing methods and definition 2 of weighted methods have the same access density Reject null hypothesis

PAGE 64

52 CHAPTER 7 CONCLUSION S Being able to define a systematic approach to incorporate the different access points while computing access density for an urban arterial segment is an important area in arterial management. Access density appears to be a simple terminology, however inconsistent usage of the term results in inconsistent numerical v alue which makes it difficult to compare access density results from t wo different studies. There are many other factors like number of lanes, traffic volume on driveways, type of channelization which affects the crash rate which are not included in the st udy and main focus is on computational methods of access density. This study compares the different existing computational methods of calculating access density and further proposed a weighted method that is capable of capturing impacts of different types of access points. With this weighted methodology, we are a step closer to improved and clearly defined definition which could be applied in areas of safety, operational and planning areas. Non parametric statistical test s w ere performed to test if the improvement between the existing and proposed methodologies is significantly different. There was no evidence to show that three existing methods of defining access density are different however the proposed weighted methodology was found to be significant ly different and correlation values indicate an improvement with reference to explaining the crashes on the selected urban arterial.

PAGE 65

53 Assigning subjective weights to various access types does improve the correlation of access density value with crash rates This study has adopted one set of weights however more combinations of different weights could be tried to arrive at improved weights; however assigning subjective weights arises various questions Statistically the difference in the access density calcu lated by proposed methodology and that of existing methodology is significantly different. Recommended Weighted Methodology For convenience, the computational method to arrive at weighted access density can be simplified in four simple steps: 1. Develop a st raight line diagram and keep categorizing the access points into one of the types as defined in Tables 4 and 5. 2. Any access points which do not fall into the 10 types defined should be considered as a special case (say type X 1 and X 2 and so on) and quickly capture the geometry to arrive at number of conflicts points corresponding to the geometry. 3. Use the formula of dividing the conflict points by 9 to arrive at the weight (say W x1 and W x2 and so on) for the special access opening. 4. Weighted access density ca n now be calculated as: (Weight for type X) x (number of type X access) Length of the segment Where X ranges from 1 to 10 as described in tables 4 and 5. Advantages of using this weighted methodology is that it requires easy steps as described above. Unlike the first weighted approach i t can be applied to t wo way left turn lanes (TWLTL) and does

PAGE 66

54 not result in multiple assignment of weights to same access points. Most importantly t here is t ransparency in assigning weights and flexibility of accommodatin g special access types. Recommendations 1. A consistently used standard method should exist to compute access density. 2. All three components should be considered i. Both direction driveways ii. Type of intersections (signalized + un signalized) iii. Type of median openings 3. To incorporate the impact of different access points a weighted approach to compute access density must be adopted. 4. Eliminate subjective judgment and use conflict points as a measure to assign weights. This study has adopted one set of weights ho wever more combinations of different weights could be tried to arrive at improved weights based on subjective judgment. The recommended approach of assigning objective weights based on geometry types has clear advantages based on the study findings.

PAGE 67

55 REFERENCE S 1. Chimba D (2004). Sensitivity Analysis on Traffic Crash Prediction Models by Using STATA ," New Economic Paper, NEP ALL 2004 08 23. 2. Effects of Two Way Left Turn Lane on Roadway Safety 3. 4. David Plazak, Adam Garms, and 5. Zuxuan Deng, John N. Ivan, Per Grder ( 2006 the Severity of Head On Crashes: Two 6. Jongdae B aek and Joseph E. Hummer ( 2008 7. Said M. Easa, Qing Chong You ( 200 9 Dimensional Two 8. James A. Bonneson, Patrick T. Mccoy ( 1997 9. Transpor tation Research Record Transportation Research Board, No. 1665. pp. 75 83.

PAGE 68

56 10. Brown, H. and A. Tarko (1999). "The Effects of Access Control on Safety on Urban Arterial Streets", Transportation Research Record Transportation Research Board, No. 1665, pp. 68 74. 11. Impact of Raised Medians and Driveway Density: Experiences from Texas and Transportation Research Record Transportation Research Board, No. 1931, pp. 108 116. 12. Grant G. Schultz, PTOE Kordel T. Braley, ( 2002 13. Transportation R esearch Board 2008 Annual Meeting CD ROM 14. Fitzpatrick, K., P. Carlson, M. A. Brewer, M. D. Woolridge and S. P. Miaou Transportation Research Board National Research Council, Washington, D.C. 15. Transportation Research Record National Research Council, Washington, D.C. 16. Li,J., Abdelwahab, W., and Brown, G., (1994 Journal of Civil Engineering. 21, 1012 1024 17. Exploration of relationship between operating speed and roadway featur es on tangent section 18. Research Record, Transportation Research Board, No. 1950, pp. 65 72

PAGE 69

57 19. Papayann C019: Presented at the Urban Street Symposium June 28 30, 1999, Dallas, Texas. 20. ffects of Rural Highway Transportation Research Record Transportation Research Board, No. 1931, pp. 99 107 21. Presented at the Transportati on Research Board 7th National Access Management Conference August 13 16, 2006, Park City, Utah 22. Presented on Speed on Roadway Segments National Forum on Speeding June 15 16, 2005. 23. K. Fitzpatrick, P.J. Carlson, M.D. Wooldridge, M.A. Brewer, (2000) Project Summary Report 1769 S (1769 3) 24. Transportation Research Circular E C081: A Research Program for Improvement of the Hig hway Capacity Manual (2005) 25. Research Need Statement for Operational Impacts of Access Management, TRB (2008) 26. Quantification of the benefits of access management for Kentucky Overview and Status Briefing June 2, 2006 27. Taylorsville Road Scoping Study. Summ ary of Findings and Recommendations (2007) Final Report by PB Americas Inc. 28. Access Management Manual TX DOT accessed at onlinemanuals.txdot.gov/txdotmanuals/acm/acm.pdf on August 2009.

PAGE 70

58 30. ion Among Trasnportation Research Board, National Research Council, Washington, D.C., 2000. 31. Spacing th Annual meeting: CD_ROM, Transportation Research Board, Washington, D.C., 1999 32. K.S. Nesamania,_, Lianyu Chub, Michael G. McNallyc, R. Jayakrishnan Estimation of Vehicular Emissions Atmospheri c Environment 41 (2007) 2996 3008 33. on Tangent Sections of Two Transportation Research Record Transportation Research Board, No. 1737, pp. 50 57 34. Transportation Research Record National 35. Institute, Report Number PTI 9606 (December 1996).Speed to Roadway G 36. Journal of Transportation Engineering Vol. 125, No. 3 (May/June 1999). 37. Speed Characteristics Review of Special Report 89 Highway Research Board, Washington, DC (1966). 38. Frontage Road Level of Report FHWA/TX 97/1393 4F, Federal Highway Administration, Washington, DC (August 1996).

PAGE 71

59 39. Preston, H. Statistical Relationship between Vehicular Crashes and Highway Access accessed online at www.dot.state.ut.us/main/uconowner.gf?n=200704111 444381 in August 2009. 40. Basic Technical Criterion for Selecting and Ranking Raised Curb Median Prepared By City of Cape Coral Department of Public Works Transportation Division Traffic Engineering Section, September 15, 2006.

PAGE 72

60 APPENDICES

PAGE 73

61 APPENDIX A Straight Line Diagrams Figure 17 : SLD from Milepost 0 to Milepost 5

PAGE 74

62 APPENDIX A (Continued) Figure 18 : SLD from Milepost 5 to Milepost 10

PAGE 75

63 APPENDIX A (Continued) Figure 19 : SLD from Milepost 10 to Milepost 15

PAGE 76

64 APPENDIX B S ample of Access Data Collected for US 19 ROAD ID Street name Mile Post South Bound Driveway Width South Bound Driveway Use Median Width Type Of Median North Bound Driveway Width North Bound Driveway Use Signalized Approach II Types 15150000 54 ave 0.87 93.83 Inter 8.22 4 (s) 93.83 Inter yes 6 15150000 0.91 28.77 G Div 3 15150000 0.93 30.98 G Div 3 15150000 0.95 21.87 P C Div 3 15150000 1 23.35 P C 7.61 4 33.2 Res 1 15150000 1.01 7.61 4 33.2 Res 1 15150000 1.04 Div 30 Res 3 15150000 50 ave 1.1 77.51 Inter 5.37 4 26.15 C 6 15150000 1.12 div 20.31 c 3 15150000 1.15 19.52 Res Div 3 15150000 1.19 24.02 P C 5.32 4 28.89 C 6 15150000 1.24 Div 17.31 C 3 15150000 1.26 Div 20.2 C 3 15150000 1.29 27.57 P C Div 28.85 C 7 15150000 1.31 29.1 H 15.74 4 30.05 C 6 15150000 1.34 Div 30.05 C 3 15150000 1.35 Div 23.72 C 3 15150000 1.35 25.3 H Div 3 15150000 46 ave 1.37 30.18 Inter 7.72 4 (s) 19.68 Inter yes 6 15150000 1.39 30.05 C Div 26.57 C 7


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 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003454
035
(OCoLC)
040
FHM
c FHM
049
FHMM
090
XX9999 (Online)
1 100
Saxena, Meeta.
0 245
Comparison of various methods to compute access density and proposing a weighted methodology
h [electronic resource] /
by Meeta Saxena.
260
[Tampa, Fla] :
b University of South Florida,
2010.
500
Title from PDF of title page.
Document formatted into pages; contains X pages.
502
Thesis (M.S.C.E.)--University of South Florida, 2010.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
3 520
ABSTRACT: This study aims to compare three distinct methods used to compute access density and provide a comprehensive weighted methodology to enable standardization for research and application in the future. Access density is a widely used concept that calculates the number of access points within a given distance and has been extensively applied to studies related to crash modeling, operational impact and planning. Methods used in past research show that access density is computed differently by different studies and all studies do not include all access points. The weighted methodology proposed takes into account all access points including driveways, intersections and median openings and categorizes them into geometric combinations. Each geometric combination have potential number of conflict points which include diverging, weaving, merging and crossing movements depending on the type of access point. Weights were assigned to each geometry type based on these conflict point ratio. In conclusion the study identifies and compares methods previously used to compute access density and accordingly, recommends a weighted methodology that includes all access points which can be used as a standard, universal measure all access density related studies including but not limited to safety impacts, operational impacts and planning guidelines.
590
Advisor: Jian Lu, Ph.D.
653
Access density
Driveways
Median openings
Signalized intersections
Un-signalized intersections
690
Dissertations, Academic
z USF
x Civil & Environmental Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.3454