Republic of Iraq Ministry of Higher Education and Scientific Research Al-Mustansiriya University College of Engineering Highway and Transportation Engineering Department Modeling of Traffic Queues and Delay at CBD in Sulaymaniyah City A THESIS SUBMITTED TO THE HIGHWAY AND TRANSPORTATION ENGINEERING DEPARTMENT COLLEGE OF ENGINEERING AL-MUSTANSIRIYA UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN HIGHWAY AND TRANSPORTATION ENGINEERING By: SAAD MOHSEN KHALIL (B.Sc. 2010/ Highway and Transportation Engineering) Supervised by Asst. Prof. Dr. Gandhi G. Sofia Dhu al-Qi'dah, 1434 Dr. Ali Jabbar Kadhim September, 2013 ِ ِ ِ ِ الرِحيِم من ح الر اهلل م س ب َّ َّ ْ ْ ينِآ َمنهواِمنْ هِْكِ َو َاَّلي َِنِ ُآوتهواِ اّللِ َاَّل َِ يَ ْرفَعِِ َ هِ ونِخَبِيرِِ اّللِب َماِتَ ْع َمله َِ الْع َِْلِ َد َر َجاتِِ َو َ هِ الع ِظيم ص َ دق اهللُ َ َ (المجادلة)11- Dedication To My Family CERTIFICATE OF THE EXAMINING COMMITTEE We certify that we have read this thesis entitled “Modeling of Traffic Queues and Delay at CBD in Sulaymaniyah City ” and as an examining committee, examined the student (Saad Mohsen Khalil) in its content and in what is connected with it, and that in our opinion it meets the standard as a thesis for the Degree of Master of Science in Highway and Transportation Engineering. Signature: Name: Asst. Prof. Dr. Asma Th. Ibraheem (Chairman) Date: /11 / 2013 Signature: Name: Asst. Prof. Dr. Zainab A. AL-Kaissi (Member) Date: /11 / 2013 Signature: Name: Dr. Mohammed B. Abduljabar (Member) Date: /11 / 2013 Signature: Asst. Prof Dr. Gandhi G. Sofia (Supervisor) Date: /11 / 2013 Signature: Dr. Ali J. Kadhim (Supervisor) Date: /11 / 2013 Approved by the Dean of College of Engineering Signature: Assist. Prof. Dr. Jamal Saeed Abd AL-Amier Date: /11 / 2013 Supervisor Certificate We certify that the preparation of this thesis titled (Modeling of Traffic Queues and Delay at CBD in Sulaymaniyah City) has been prepared by “Saad Mohsen Khalil " under our supervision at Al-Mustansiriya University college of Eng. Highway and Trans. Eng. Dept., in partial fulfillment of the requirement for the degree of Master of Science in Highway and Transportation Engineering. Signature: Name: Asst. Prof. Dr. Gandhi G. Sofia Date: / 9 / 2013 Signature: Name: Dr. Ali Jabbar Kadhim Date: / 9 / 2013 In view of the available recommendations; I found this thesis adequate for debate by the examining Committee Signature: Name Dr. Ali Jabbar Kadhim (Head of the Highway and Transportation Department) Al-Mustansiriya University Date: / 9 / 2013 Acknowledgements First and foremost, all thanks to GOD for giving me the strength, determination, and courage to implement this goal. All wisdom and knowledge comes from his almighty, and I could not have reached this stage without his divine guidance. Cordial thanks are due to my supervisors Asst. prof. Dr. Gandhi G. Sofia and Dr. Ali J. Kadhim for their guidance, patience, and encouragement through the preparation of this study. Special thanks are devoted to Traffic Directorate of Sulaymaniyah Governorate for helping me during this research work. Thanks are presented to Eng. Ghadah Ghassan, Eng. Mays Rabih and Asst. Lecturer Israa Saeed at Al-Mustansiriya University for their interest and support. Finally, I would like to thank and acknowledge my family. They have always supported everything I have ever attempted; each gives me the time and space I needed. I Abstract Abstract Sulaymaniyah city is suffering from congestion, vehicle delays on most streets and intersections especially during peak hours, causing a lot of traffic problems and a negative impact on economy of the country. The increasing number of vehicles and limited new road construction will increase the problem in the network. The correct prediction of delay and queue length become important to make the most efficient use of the existing road system by providing better traffic operation and control. The objective of this study is the evaluation of traffic performance at selected network, through the evaluation of intersections performance and development of delay time and queue length models with reference to traffic prevailing conditions at Sulaymaniyah city. This is followed by suggestions of some improvement proposals including changing timing plan, geometric improvement and changing intersection type. To carry out the above objective, traffic data is collected at the network in CBD area in Sulaymaniyah city. Video camera is used to collect field data because it provides permanent record of data with minimum manpower. These data are abstracted from video records using EVENT program, and processed by prepared EXCEL sheets, while, the spot speed data for each entire link in the network are collected using pavement marking method. SYNCHRO/SIMTRAFFIC software V.20 is used for evaluation and analysis of intersections, best timing plan and coordination. Suggested proposals are evaluated for the target year 2017 by using mentioned software. Regression models, developed to estimate delay and queue length, show good correlation with field values. The ratio of total width of exit roadway of intersection for traffic departing straight forward to total width of lane groups II Abstract departing to the same exit roadway at the same phase at stop line (We/Ws) appear to affect the delay. The ratio has a negative effect on delay time when the ratio is less than 1.0. Delay is estimated by HCM 2000 delay model and compared with field delay. Based on results obtained from analysis, HCM 2000 delay model has been modified to represent the field delay. It has been observed that, HCM 2000 delay model consistently over estimates delay at degree of saturation more than 1.0. It has been suggested from the analysis that, theoretical HCM 2000 delay model may be reduced by 81 % to better reflect field conditions. Recommended delay model shows good correlation with field data within the limits of this study. Improvement strategies applied to the traffic flow in the study area of Sulaymaniyah city, through the application of the calibrated software (SYNCHRO/ SimTraffic). The improvement strategies divides into two parts. The first part includes Signal Timing Optimization and Coordination. The second part includes the geometric improvement. This improvement has very good effect on the measure of effectiveness (MOE) for the study network. Finally, the measure of effectiveness obtained from SYNCHRO/ SimTraffic optimization runs after the last improvement strategy for the target year 2017. III L ist of Contents L ist of Contents Acknowledgement ................................................................................................ I Abstract ................................................................................................................II List of Contents ................................................................................................. IX List of Tables ...................................................................................................VIII List of Figures .....................................................................................................X List of Notations .............................................................................................. XII Chapter One Introduction 1.1 General .......................................................................................................... 1 1.2 Sulaymaniyah city...........................................................................................3 1.3 Objectives of the Study ................................................................................. 4 1.4 Research Methodology .................................................................................. 4 1.5 Scope of the Work ......................................................................................... 4 Chapter T wo L iterature Review 2.1 General............................................................................................................6 2.2 Urban Streets ……………………………………………………… ……….6 2.3 Intersections....................................................................................................7 2.3.1 Signalized Intersections...............................................................................9 2.3.1.1 Signalized Intersection Flow Characteristics .......................................10 2.3.1.2 Performance Measures...........................................................................11 2.3.1.3 Discharge Headway, Lost time, and Saturation Flow............................11 2.3.1.4 Capacity of Signalized Intersections.....................................................14 2.3.1.5 Level of service of Signalized Intersections.........................................18 IV L ist of Contents 2.3.1.6 Delay at Signalized intersection..........................................................20 2.3.1.6.1 Development of Delay Models........................................................22 2.3.1.7 Queue at Signalized Intersection.........................................................30 2.3.1.7.1 Existing Queue Models...................................................................34 2.3.2 Two Way Stop Controlled intersection (TWSC intersection) ….............39 2.3.2.1 Delay at Two Way Stop Controlled intersection ..................................39 2.4 Calibration of The Traffic Models ………...…………………………....… 41 2.5 Local Traffic Improvement Studies…………………………………………42 2.5.1 Mosul City……………………………………………………………….42 2.5.2 Baghdad City…………………………………………………………….43 Chapter T hree M ethodology and Data Collection 3.1 General ........................................................................................................ 44 3.2 Study Area Description ............................................................................... 44 3.3 Data Collection Methodology ..................................................................... 47 3.4 Methods of data collection............................................................................49 3.4.1 Manual methods.....................................................................................49 3.4.2 Video Recording Method.......................................................................49 3.5 Geometric Data ............................................................................................50 3.6 Traffic Data ..................................................................................................50 3.6.1 Traffic Volume Data ............................................................................ 50 3.7 Speed Data ................................................................................................... 54 3.8 Saturation Flow Data ................................................................................... 56 3.9 Traffic Signal Data ...................................................................................... 56 3.10 Minimum Green Time for Pedestrian ....................................................... 57 3.11 Data Abstraction ....................................................................................... 58 V L ist of Contents 3.12 Queue Length .............................................................................................60 3.13 Data Processing ......................................................................................... 60 3.14 Delay Measurement.................................................................................... 61 3.15 Software Used for Data Analysis............................................................... 64 Chapter Four Data Presentation and A nalysis 4.1 General ........................................................................................................ 67 4.2 Evaluation of Existing Traffic Flow .............................................................67 4.3 SYNCHRO Model Calibration .................................................................... 71 4.4 Development of Statistical Models for Delay and Queue length .................75 4.4.1 Process of Models Building....................................................................75 4.4.2 Identification of Dependent and Predictor Variables..............................75 4.4.3 Data Analysis..........................................................................................76 4.4.3.1 Selecting Sample Size……………………………………………..76 4.4.3.2 Scatter Plots.....................................................................................76 4.4.3.3 Outliers............................................................................................77 4.4.3.4 Testing of Normality.......................................................................78 4.4.3.5 Multicollinearity.............................................................................80 4.4.4 Regression Modeling.............................................................................82 4.4.4.1 Stepwise Regression Procedure......................................................83 4.4.4.2 Developed Models..........................................................................83 4.4.4.2.1 Delay model.............................................................................84 4.4.4.2.2 Left turn Queue Model............................................................85 4.4.4.2.3 Through Queue Model............................................................86 4.5 Models Validation.........................................................................................87 4.6 Models Analysis............................................................................................90 VI L ist of Contents 4.7 Compression of Observed Delay with That HCM2000 Delay.................. 91 4.8 Improvements ...............................................................................................93 4.8.1 Signal Timing Optimization and Coordination ........................................ 93 4.8.2 Geometric Improvement .......................................................................... 97 4.9 Forecasted Network ....................................................................................101 Chapter Five Conclusions and R ecommendations 5.1 Conclusions ............................................................................................... 103 5.2 Recommendations ..................................................................................... 104 5.3 Recommendations for Further Research ....................................................104 R eferences References ........................................................................................................106 A ppendices A Sample of Spot Speed Data..............................................................................A B Work Sheet for Field Measurement of Delay...................................................B C Sample of EVENT Program Output ................................................................C D Sample of SYNCHRO Output .........................................................................D E Traffic and Geometric Condition of Signalized intersections …………....…..E VII List of Tables List of Tables 2-1 Principal Factors Affecting Saturation Flow ……….…………………...18 2-2 Level of Service Criteria for Signalized Intersections ….………………19 2-3 Some Traffic Improvement Studies in Baghdad City Sub-networks….....43 3-1 Geometric Features for the Selected Intersections …………...…………52 3-2 Recommended Trap Length ………………………………..…………..56 3-3 Phase Order, Phase and Cycle Length for Signalized Intersections …...57 3-4 Average Observed Queue Length ………………………..…………….60 3-5 Summary of Abstracted Data ………………………………..…………60 3-6 Acceleration/Deceleration Delay Correction Factor (CF) – second…….…64 4-1 Total Delay and Level of Service for Isolated Intersections Produced by SYNCHRO for Existing Condition ………………..…......................................70 4-2 Comparison of Actual and SYNCHRO Delay Prediction for Signalized Intersections………………………………………………………………..…...73 4-3 Case Processing Summary for Outliers …………………..…………….....77 4-4 K-S test and (S-W) test Results.....................................................................79 4-5 Correlation Coefficient Matrix for Delay Model.…………..….……........81 4-6 Correlation Coefficient Matrix for Queue Length of Left Turn Movement…………………………..……………………………………....…81 4-7 Correlation Coefficient Matrix for Queue Length of Through Movement …………………………………………………………………….82 4-8 Stepwise Regression Models Summary for Delay Model..……….….…...84 4-9 Data Range and Statistical Characteristics for Delay Model ……...…….85 4-10 Stepwise Regression Models Summary for Queue Length of Left Turn Movements ...……………………………………………………………….….85 VIII List of Tables 4-11 Data Range and Statistical Characteristics for Queue Length Model of Left Turn Movement ………….……………………………………………………86 4-12 Stepwise Regression Models Summary for Queue of Through Turn Movements …………..…………………………………………………….…..86 4-13 Data Range and Statistical Characteristics for Queue Length Model of Through Movement …………………...……………………………………….87 4-14 Regression Results for the Three Models ………………………..…….88 4-15 ANOVA Test for the Regression Model ……………..………………….92 4-16 Measure of Effectiveness for the Improved Intersections with the Optimum Cycle Length Produced by SYNCHRO ……………………………96 4-17 Measure of Effectiveness for the Improved Intersections Produced by SYNCHRO (Geometric Improvement(………….………………98 4-18 Measure of Effectiveness Produced by SYNCHRO after Last Improvement Strategy for the Base Year 2010, and Target Year 2017…..….102 IX List of Figures List of Figures 1-1 The Study Area at Sulaymaniyah City …………………………...…..… 3 1-2 Research Methodology …………………………………………………..5 2-1 Traffic Control Type 2-2 Recommended Traffic Control Type with Different Heavy Vehicles ...………………………………...………………8 Percentage ………………………………………………………………..8 2-3 Fundamental Attributes of Flow at Signalized Intersection…….....……10 2-4 Condition at Traffic Interruption ……………………………………….12 2-5 Headways at Traffic Interruption ……………………………………....13 2-6 Saturation Flow and the Related Signal Timing Parameters …………..17 2-7 Illustration of Delay Measures …………..……………………………..21 2-8 Queuing Diagram for Signalized Intersection …………………………32 2-9 Graphical Representation of the queue accumulation and discharge process………………………………………………………………………….36 3-1 The Study Area in Sulaymaniyah City …….…………………………..45 3-2 Palace Intersection Layout …………………………..…………………46 3-3 Sara Intersection Layout ………………………………………………46 3-4 Great Mosque Intersection Layout ………………….…………………47 3-5 Traffic and Data Collection Procedure ……………..…………………48 3-6 Hourly Traffic Volume for Average Typical Weekdays ……………....53 3-7 Average Spot Speed for Each Link in the Study Area…….……….…..55 4-1 Network Layout as Printout by SYNCHRO ……………………...…….68 4-2 Lane Settings by SYNCHRO …………………………………….…….69 4-3 SYNCHRO Average Delay Values versus Field Delay Values for All Approach ……………………………………………………….…...…………74 X List of Figures 4-4 SYNCHRO Delay Values versus Field Delay Values after Calibrating the Basic saturation flow rate for All Approach…………………...………………74 4-5 Observed versus Predicted Delay……………………………………….89 4-6 Observed versus Predicted Queue Length for the Model of Left Turn Movement …………………………………………………………………...…89 4-7 Observed versus Predicted Queue Length for the Models of Through Movement …………………………………………………………….………..90 4-8 HCM Delay versus Field Delay …………………….…………………..92 4-9 Optimization Steps in SYNCHRO ……………………………………...94 4-10 Improvement Proposal for Sara Intersection……….……………………100 4-11 Improvement Proposal for Great Mosque Intersection…..……………..100 4-12 Improvement Proposal for Palace Intersection…………...……………..101 XI List of Notations List of Notations adj Adjustment term for model AWSC All Way Stop Controlled AR Annual rate of traffic increase (%) b1, b2 Regression coefficients c Adjusted lane group capacity (veh/hr) C Cycle length (sec) CBD Central Business District CCG Canadian Capacity Guide ci Capacity of lane group I (vph) CI Controller interface CL Lane group capacity per lane (veh/hr) cm,x Capacity of movement x (veh/hr) d Average overall delay per vehicle (sec) d1 Uniform delay (sec/veh) d2 Incremental, or random, delay (sec/veh) d3 Residual demand delay(sec/veh) dM Average delay per vehicle(sec) do Overflow delay (sec) du Uniform delay (sec) fp Progression adjustment factor FVS Fraction of vehicles stopping g Effective green time (sec) G Green time (sec) G(t) The mean queue length at time t (veh) XII List of Notations g/C Green ratio Ge Effective green signal duration (sec) gi Effective green time for lane group i (sec) GIS A geographic information system Gp Minimum green time (sec) h Saturation flow rate headway (sec) HCM Highway Capacity Manual I Interval between vehicle-in-queue counts (sec) I Upstream filtering/metering adjustment factor k Incremental delay factor KB Second-term adjustment factor related to early arrivals L The segment length(m), l1 Start-up lost time (sec) l2 The clearance lost time (sec) lamda Effective green ratio LOS Level of service Lp Crosswalk length (m), MLR Multiple linear regression MOE Measure of effectiveness MSE Mean square error n Traffic analysis period (year) Nped Number of pedestrians crossing during an interval (p) in (sec) P Proportion of vehicles arriving during the green interval PF Adjustment factor for the effect of the quality of progression in coordinated systems PF2 Adjustment factor for effects of progression pnmv Percentage NMV in traffic XIII List of Notations q Arrival flow rate (veh/hr) q Average number of vehicles in queue (veh) Q Average queue length (veh) Q Maximum distance in vehicles over which queue extends from stop line on average signal cycle (veh) q Vehicle arrival rate (PCU/sec) Q1 First-term queued vehicles (veh), Q1CCG Estimates of average queue reach (veh) Q2 Second-term queued vehicles (veh) Q2CCG Maximum queue reach during the congestion period (vehicles) Qb Initial queue at the start of period T (veh) QbL Initial queue at start of analysis period (veh) QM Maximum queue length (veh) QQ Average queue length while queue is present (veh) r Effective red time (sec) R2 Correlation coefficient s Mean service rate (veh/hr) s Saturation flow for lane group i (vph) Si The measured spot speed of vehicle (km/hr) Sp Average speed of pedestrians (m/sec) T Duration of analysis period (h), t Duration of unmet demand in T (h), T Evaluation time (hr) T Length of analysis period (hr) Tc Length of cycle (sec) te Evaluation time (minutes) TFF Traffic forecast factor XIV List of Notations ti Incremental headway for the ith vehicle (sec) ti The time required for vehicle (i) to transverse the section (sec) tQ Time duration of queue (sec) TWSC Two-way stop- controlled intersection u Delay parameter v Arrival rate (veh/hr) v Mean arrival rate (veh/hr) Viq Sum of vehicle-in-queue counts, vehicle VL Lane group flow rate per lane (veh/hr), vphg Vehicles per hour of effective green time Vstop Volume counts of stopping vehicle Vtot Volume counts of total vehicles vx Flow rate for movement x (veh/hr), w Width of lane group at stop line (m) WE Effective cross walk width (m) We Total width of exit roadway of intersection (m) Ws Total width of lane groups departing to the same exit roadway (m) x Degree of saturation X Volume to capacity ratio of lane group XL Ratio of flow rate to capacity (vL/cL ratio). Y Change interval (sec) β0 Constant term in the model β1 Coefficient corresponding to q β2 Coefficient corresponding to c β3 Coefficient corresponding to x β4 Coefficient corresponding to lamda β5 Coefficient corresponding to pnmv XV List of Notations Δ Minimal time distance between vehicles (sec) λ Arrival rate (veh/sec) λ Proportion of the cycle that is effective green (g/C) ρ Flow ratio (-) σ2µ Variance of service time (sec) µ Service rate (veh/sec) XVI INTRODUCTION Chapter One Introduction Chapter One Introduction 1.1 General The continuous traffic growth through developed areas and the difficulties in building infrastructure have caused many traffic problems. In 1999, the number of registered vehicles in Sulaymaniyah city was (32468) vehicles. This number increased to (101,878) vehicles in 2006, corresponding to an annual increment in the number of vehicles of 17.75 %. After 2006, the growth rate becomes 3.4 %. (Traffic directorate of Sulaymaniyah governorate, 2013). Therefore, a need for traffic engineering studies for careful monitoring of operating condition existing highway facilities, developing a new model for such local conditions, evaluation of the traffic flow quality are essential for developing improved strategies to mange the increase in traffic demand. Over the past years, the quality and accuracy of the simulation models have been improved. In addition, new concepts and new technologies continue to add to this field. Simulation models have different characteristics: static or dynamic, deterministic or stochastic, microscopic or macroscopic. Each simulation model was developed with different backgrounds and different algorithms. Each model has its positive and negative points and it is not ideal for every situation. Therefore, the analyst should select the program that satisfies the project objectives and the prevailing traffic conditions for the selected study area. Usually, little information is provided about the default parameters embedded in the simulation models such as their source, their appropriateness and 1 Chapter One Introduction how they should be modified under different traffic conditions. Therefore, the analyst should ensure that appropriate changes to these parameters are made based on the field measured traffic data and not only on the engineering judgment (Turley, 2007). Before using any analytical or simulation model, there is a need for calibration of the model in order to obtain a close match between the simulated results and the field measurements. The parameters or factors that can be calibrated vary from one model to another. Thus, when the analyst selects a model; she/he should know which parameters or factors might be used as a yard stick in the calibration process and which measures of effectiveness (MOEs) must be considered for such calibration. Usually, the MOEs are used for calibration including intersection delay, number of stopped vehicles, platoon dispersion and average or maximum queue length. For many years, measurement of the level of performance of signalized intersections has primarily focused solely on vehicle delay. Beside delay, other performance measures such as the number of vehicle stops and the spatial extent of queues on intersection approaches have also been found to play an important role in the performance evaluation of signalized intersections of developing countries, where most of the times the intersection remains near or at oversaturated condition. Hence, stop and go situation is very frequent at intersections. These measures not only relate to the level of service provided to the drivers, but also to the level of fuel consumption and air pollution generated by the vehicles traversing the signalized intersections. In particular, while vehicle stop estimates play an important role in determining vehicle fuel consumption and emissions on intersection approaches, queue length estimates are important not 2 Chapter One Introduction only for the design of pocket lanes, but also to ensure that traffic signal operations do not result in vehicle queues that spillback onto upstream intersections. 1.2 Sulaymaniyah City Sulaymaniyah is a city in Kurdistan Region of Iraq. It lies at 45˚, 26ˋ longitude and 35˚, 33ˋ latitude, departed at about 355 km north east away from Baghdad. The Azmer Range, Goyija Range and the Qaiwan Range in the north east of the city , Baranan Mountain in the south and the Tasluja Hills in the west surround it. The study area, as located in the central business district area (CBD) represents the core area of the city, as shown in Figure (1-1). Figure (1-1): The Study Area at Sulaymaniyah City.(Google maps,2013) 3 Chapter One Introduction The new growth in Sulaymaniyah traffic congestion has been recognized as a serious problem in the city, which affects the economy, travel time, driver behavior, and causes discomfort to drivers and visitors. The number of vehicles increases rapidly without considerable increase in the capacity of the road network, which leads to increase delay and lower the level of service (LOS). 1.3 Objectives of the Study The main objectives of this study are, as follows: 1. Evaluating the traffic performance at the studied area, and suggesting the required traffic and/or geometric solutions to alleviate the congestion problem at the selected network. 2. Developing statistical model for the prediction of vehicle delay at signalized intersection. 3. Developing statistical model for the prediction of queue length at signalized intersection. 4. Evaluating the adequacy of applying selected software on local traffic conditions, and adjusting models parameters to be applicable with local conditions. 1.4 The Study Methodology The study methodology can be presented in Figure (1-2). 1.5 Scope of the Work The general scope adopted in this study can be summarized, as follows: 1. Chapter one gives a brief idea about the present work. 2. Chapter two reviews the literature which is related to previous work in delay and queue length. A description of the selected computer program (SYNCHRO 8) is also presented. 4 Chapter One Introduction 3. Chapter three describes the study area and the methodology of data collection and abstraction to be used in the selected traffic program. 4. Chapter four illustrates results analysis, discussion and improvement strategies. 5. Chapter five contains conclusions, recommendations, recommendations for further researches. Site Selection Traffic and Geometric Data Collection Using Video Recording, Satellite Image, Manual Method and Moving Car Data Abstraction and Processing Using EVENT & Prepared EXCEL Sheets Development of Delay Model for Signalized Intersection Development of Queue Model for Signalized Intersection Applying Software for Simulation Compare Observed with Simulated Delay Statistical Tests Improvement Alternatives Draw Recommendations Figure (1-2): Research Methodology. 5 and LITERATURE REVIEW Chapter Two Literature Review Chapter Two Literature Review 2.1 General In this chapter, a description of urban streets, signalized intersections, capacity and level of service at signalized intersection, development of delay and queue models at signalized and unsignalized intersections are presented with the researches related to this field of study. 2.2 Urban Streets In the hierarchy of street transportation facilities, urban streets (including arterials and collectors) are ranked between local streets and multilane suburban and rural highways. The difference is determined principally by street function, control conditions and the character and intensity of roadside development (TRB, 2005). Urban minor arterials carry large traffic volumes within and through urban areas. The principal objective of an urban minor arterial should be mobility with limited or restricted service to local development. Urban arterials are capable of providing some access to abutting property. Such access service should, however, be only incidental to the arterial’s primary function of serving major traffic movements (AASHTO, 2004). The degree of mobility provided by urban street is assessed in terms of travel speed which affects highway capacity (HCM) (TRB, 2005), as follows: 6 Chapter Two Literature Review 1. The street environment includes the through-traffic stream. The speed is influenced by three main factors, the geometric characteristics of the facility, the character of roadside activity and adjacent land uses. 2. The interaction among vehicles is determined by traffic density, the proportion of trucks and buses and turning movements. 3. Traffic control (including signals and signs) forces a portion of all vehicles to slow or stop. As a result, these factors also affect quality of service. 2.3 Intersections An intersection is a general area where two or more highways join or cross at grade, within which are included the roadway and roadside facilities for traffic movements in that area. Each highway radiating from an intersection and forming part of it, is an intersection leg. (SORB, 2005) (Garber and Hoel, 2010) defined the intersection as an area, shared by two or more roads whose main function is to provide for the change of route direction. Intersections vary in complexity from a simple intersection, which has only two roads crossing at a right angle to each other, and more complex intersection, where three roads or more cross within the same area . The intersection is required to control conflicting and merging streams of traffic so that delay is minimized. This is achieved through choice of geometric parameters that control and regulate the vehicle paths through the intersection. These determine priority so that all movements take place with safety (Rogers, 2008). Each intersection type aims at providing vehicle drivers with a road layout that will minimize confusion. The need for flexibility dictates the choice of most suitable junction type. The selection process requires the economic, environmental and operational effects of each proposed option to be evaluated 7 Chapter Two Literature Review (Rogers, 2008). The determination of intersection control type is illustrated in Figure (2-1). Figure (2-1): Traffic Control Type (Rogers, 2008). Polus and Vlahos (2005) recommended ranges of major road one-way traffic volumes for which the construction of the at-grade intersection control is preferred, as shown in Figure (2-2). Figure (2-2): Recommended Traffic Control Type with Different Heavy Vehicles Percentage (Polus and Vlahos, 2005). 8 Chapter Two Literature Review 2.3.1 Signalized Intersections Traffic control signals that are properly designed, located, operated, and maintained will have one or more of the following advantages (FHWA, 2009): 1. Providing the orderly movement of traffic. 2. Increase the traffic-handling capacity of the intersection if: A. Proper physical layouts and control measures are used. B. The signal operational parameters are reviewed and updated (if needed) on a regular basis (as engineering judgment determines that significant traffic flow and/or land use changes have occurred) to maximize the ability of the traffic control signal to satisfy current traffic demands. 3. Reducing the frequency and severity of certain types of crashes, especially right-angle collisions. 4. Using coordinated signals provide for continuous or nearly continuous movement of traffic at a definite speed along a given route under favorable conditions. 5. Using the interrupt heavy traffic at intervals to permit other traffic, vehicular or pedestrian, to cross. Traffic control signals, even when justified by traffic and roadway conditions, can be ill-designed, ineffectively placed, improperly operated, or poorly maintained. Improper or unjustified traffic control signals can result in one or more of the following disadvantages (FHWA, 2009): 1. Excessive delay. 2. Excessive disobedience of the signal indications. 3. Increased use of less adequate routes as road users attempt to avoid the traffic control signals. 4. Significant increases in the frequency of collisions (especially rear-end collisions). 9 Chapter Two Literature Review 2.3.1.1 Signalized Intersection Flow Characteristics For a given approach at signalized intersection, three signal indications are seen: green, yellow, and red. The indication may include a short period during which all indications are red, referred to as an all-red interval, which with the yellow indication forms the change and clearance interval between two green phases. Figure )2-3( presents some fundamental attributes of flow at signalized intersection. The diagram represents a simple situation of one-way approach to signalized intersection having two phases in the cycle (HCM, 2000). Figure (2-3): Fundamental Attributes of Flow at Signalized Intersection (Source: HCM 2000) 10 Chapter Two Literature Review The diagram is divided into three parts. The first part shows a time-space plot of vehicles on the northbound approach to the intersection. Intervals for the signal cycle are indicated in the diagram. The second part repeats the timing interval, and labels the various time intervals of interest with the symbols. The third part is an idealized plot of flow rate past the stop line, indicating the saturation flow rate. 2.3.1.2 Performance Measures Common measures by which the performance of an intersection may be evaluated include (1) delay, (2) stops and (3) queue length. Each of these may be expressed as values, which represent totals or averages for the entire intersection or for particular approaches or movements within the intersection. Averages are often expressed on a per vehicle basis. Other measures, which have been used to characterize the performance, are throughput and total travel time (Mcshane and Roess, 1990). Delay, specifically the control delay is the measure used in the signalized intersection methodology of the 2000 HCM and the primary measure used in the number of signalization optimization procedure. Performance measures are critical part of all intersection design methodologies. 2.3.1.3 Discharge Headway, Lost time, and Saturation Flow Lee and Chen (1986) studied the entering headways in small city Lawrence, Kansas and six factors were examined. Entering headway values from total of 1,899 traffic queues were recorded by using video camera equipment. He found that: 1. Signal type has little influence on entering headway at signalized intersections. 2. Time of the day (AM or PM) has little influence on entering headways. 11 Chapter Two Literature Review 3. The inside lane of an approach has slightly lower entering headways than does outside lane. 4. The entering headways at approaches with speed limits of 20 mph (32 kph) are significantly higher than those at approaches with higher speed limits. For approaches with speed limits higher than 30 mph (48 kph), the influence of speed limit on the headway is noticeable. 5. In general, streets that have higher speed limits and less roadside friction have lower entering headway values. Figure (2-4) illustrates a queue of vehicles at signalized intersection. When the signal turns green, the queue begins to move into the intersection. The first headway is defined as the time between the initiation of the green signal and the first vehicles front bumper crossing the stop line. This first headway will be comparatively long, as the first driver must see the light turn green, then react and accelerate the vehicle into intersection. The second headway, measured as the time between the first and second vehicles bumper's crossing the stop line, will be somewhat smaller, as the second driver's reaction to the green overlaps the first driver's reaction. The third reaction would be smaller than the second, and so on. Figure (2-4): Condition at Traffic Interruption (Source: HCM 2000) 12 Chapter Two Literature Review Eventually (usually between the fourth and sixth headway), vehicles entering the intersection have fully accelerated by the time they reach the stop line, and approximately equal headways would then be observed. Figure (2-5) represents a plot of average headway of vehicles entering the intersection versus the position of the vehicle in the queue. Figure (2-5): Headways at Traffic Interruption (Source: HCM 2000). The constant headway achieved once a stable moving queue is established, is called the saturation headway. If it is assumed that, each vehicle entering the intersection consumes h sec, then the number of vehicles that can enter the intersection in a lane may be computed as (HCM,2000): 𝑠 = 3600 𝐸𝑞. (2 − 1) ℎˉ 13 Chapter Two Literature Review where: 𝑠 = saturation flow rate (vehicle per hour of green per lane). ℎˉ = saturation flow rate headway (sec). The saturation flow rate(𝑠), is the number of vehicles that could enter the intersection in a single lane if the signal is always green for that lane, and the vehicles never stop. Traffic stream at signalized intersection does stop periodically. When the traffic stream starts, the first several vehicles consume more than h )sec/veh(. Thus, sum of the incremental headways (above h sec/v) for the first several vehicles is called start-up lost time (HCM,2000). 𝑬𝒒. (2 − 2) 𝒍𝟏 = ∑ 𝒕𝒊 𝒊 where: 𝑙1 = start-up lost time (sec). 𝑡𝑖 = incremental headway for the ith vehicle (sec). The clearance lost time 𝑙2 , is the time between the last vehicle from one approach entering the intersection and the initiation of the green signal for conflicting movements. The capacity at signalized intersection is based on saturation flow rate; the lost time and the signal timing (HCM, 2000). 2.3.1.4 Capacity of Signalized Intersections The 2000 Highway Capacity Manual defines the capacity of facility as " the maximum hourly rate at which persons or vehicles can reasonably be expected to traverse point or uniform section of a lane or roadway during a given time period under prevailing roadway, traffic and control conditions. 14 Chapter Two Literature Review The capacity of a traffic signal approach may be defined as the maximum number of vehicles that can pass when the signal turns to green under the prevailing traffic and weather conditions. Therefore, the capacity of a signal controlled junction is limited by the capacities of the individual approaches to the junction (Sofia, 1998). The factors which affect capacity are (Al-Azzawi, 2003): 1-Physical and operation factor: Includes parking condition, width of approaches, one -way or two-way operation and number of lanes. 2-Traffic characteristics: Consist of traffic signals (length of cycle time and green to cycle time ratio for each approach) and marking of approach lanes. 3-Enviromental factors: Include degree of utilization of an individual approach, variation of demand during the peak hour and the location of the intersection within the metropolitan area. 4-Control measures: Involve turning movements and vehicles composition Cars, buses and trucks). 5- Area type: the saturation flow in central business district (CBD), where intersection geometry, pedestrian flow, and roadside friction are more restrictive, is less than the saturation in other areas. Capacity of signalized intersections is based on the concept of saturation flow rate, which is defined as the maximum rate of flow that can pass through a given lane group under prevailing traffic and roadway conditions, assuming that a given lane group has 100% of real time available as an effective green time and is expressed in units of vehicles per hour of effective green time (vphg) (TRB, 2005). Figure (2-6) shows that, the average rate of flow is lower during the first few seconds (while vehicles are accelerating to normal running speed) and also during the amber period (as some vehicles decide to stop and others continue to 15 Chapter Two Literature Review move on). It is convenient to replace the green and amber periods by an "effective green" period, during which the flow is assumed to take place at the saturation rate, combined with a "lost" time during which no flow takes place. This is a useful concept because capacity is then directly proportional to effective green time. The capacity of the lane group is (HCM,2000): 𝑐=𝑠∗ 𝑔 𝐶 𝐸𝑞. (2 − 3) where: c : capacity of lane group (vph), s: saturation flow for lane group (vph), 𝑔: effective green time for lane group , 𝐶: cycle length. 16 Chapter Two Literature Review Figure (2-6): Saturation Flow and the Related Signal Timing Parameters (Akcelik, 2009). There are also several factors affecting saturation flow which can be summarized in Table (2-1) 17 Chapter Two Literature Review Table (2-1): Principal Factors Affecting Saturation Flow (Al-Azzawi, 2003). Factors Element Affecting Saturation Flow • Approach width • Width of lanes • Number of lanes Geometric Condition • Grade • Turning Radius • Length of turn bay • Signal timing and phasing arrangements • Peaking characteristics Operating Condition • Parking activities • Bus stop operations • Traffic composition • Turning movements Traffic Characteristics • Pedestrian activity • Weather • Driver behavior • Area population Environmental and other Factors • Roadway surface conditions • Adjacent land use 2.3.1.5 Level of Service of Signalized Intersections Level of service is defined in terms of the average total vehicle delay of all movements through an intersection. Vehicle delay is a method of quantifying several intangible factors, including driver discomfort, frustration and lost travel time. Specifically, levels of service criteria are stated in terms of average delay per vehicle during a specified time period (for example, the PM peak hour). Vehicle delay is a complex measure based on many variables, including signal phasing (i.e., progression of movements through the intersection), signal cycle length and traffic volumes with respect to intersection capacity (TRB, 2000). 18 Chapter Two Literature Review Table (2-2) shows level of service criteria for signalized intersections, as described in the Highway Capacity Manual (TRB, 2000). Table (2-2): Level of Service Criteria for Signalized Intersections (TRB, 2000). Level of Service A B C Average Control Delay (sec/veh) D >35-55 E F >55-80 >80 ≤10 >10-20 >20-35 General Description (Signalized Intersections) Free Flow Stable Flow (slight delays) Stable Flow (acceptable delays) Approaching unstable flow (tolerable delay, occasionally wait through more than one signal cycle before proceeding) Unstable Flow (intolerable delays) Forced flow (jammed) Highway Capacity Manual (2000) describes service quality in the following terms: • Speed and travel time: One of the most easily perceived measures of service quality is speed, or travel time. On freeway, speed is very evident measure of quality, while on surface streets systems, the driver is very sensitive to total travel time. • Density: Density is not often used in traffic analysis. Density describes the proximity of vehicles to each other in the traffic stream and reflects ease of maneuverability in the traffic stream, as well as psychological comfort of drivers. • Delay: Delay can be described in many ways. It represents excess or additional travel time due to travel time of controls. • Other measures: A variety of other measures are used to describe service quality. 19 Chapter Two Literature Review In some cases, the used measures are not directly discernible to drivers or passengers. Such measures generally rely upon volumes or flow rates. 2.3.1.6 Delay at Signalized Intersection Delay is one of the key parameters that is utilized in the optimization of traffic signal timings. Furthermore, delay is a key parameter in computing the level of service provided to motorists at signalized intersections. Delay, however, is a parameter that is difficult to estimate because it includes the delay associated with decelerating to a stop, the stopped delay and the delay associated with accelerating from a stop (Youn, 2000). Delay can be quantified in many different ways. The most frequently used forms of delay are defined below: • Stopped time delay • Approach delay • Travel time delay These delay measures can be quite different, depending on conditions at the signalized intersection. Figure (2-7) shows the differences among stopped time, approach and travel time delay for single vehicle traversing a signalized intersection. The desired path of the vehicle is shown, as well as the actual progress of the vehicle, which includes a stop at a red signal. 20 Chapter Two Literature Review Figure (2-7): Illustration of Delay Measures (McShane, 2004). The delays experienced on the arterial signalized streets are mainly associated with the intersections where conflicting movements are separated and controlled by traffic signals. These traffic signals can operate under an isolated control strategy, with the signal settings of each signal set independently of the settings of adjacent signals. The delay is defined as the difference in travel time when a vehicle is unaffected by the controlled intersection and when a vehicle is affected by the controlled intersection. This delay includes lost time due to deceleration and acceleration as well as stopped delay. Thus, intersection delay estimates are directed toward estimating total delay or simply stopped delay (Youn-Soo Kang, 2000). 21 Chapter Two Literature Review 2.3.1.6.1 Development of Delay Models Webster, (1958) developed a model for estimating the delay incurred by motorists at under-saturated signalized intersections that become the basis for all subsequent delay models. The mathematical form of the model is: 𝑑= C (1−𝜆)2 2(1−𝜆𝑋) + X2 2v(1−X) C 1 3 − 0.65 ( 2 ) [x 2+51] v Eq. (2 − 4) where: d = average overall delay per vehicle (seconds), = proportion of the cycle that is effective green (g/C), C = cycle length (seconds), v = arrival rate (vehicles/hour), c = capacity for lane group (vehicles/hour), g = effective green time (seconds). X= lane group v/c ratio or degree of saturation In Equation (2-4), the first term represents the average delay to the vehicles assuming uniform arrivals. The second term estimates the additional delay due to the randomness of vehicle arrivals. This additional delay is attributed to the probability that sudden surges in vehicle arrivals may cause the temporary oversaturation of the signal operation. The third term, finally, is an adjustment factor that is introduced in the model to correct the delay estimates and that develops semi-empirically. Following this classical work, numerous studies were conducted in the field of estimating delays at signalized intersections. As a result of these studies, a number of delay models based on deterministic queuing theory were proposed to suite different field conditions. Among these, the most notable are the models 22 Chapter Two Literature Review developed by Miller (1963) and Akcelik (1981) in Australia, the models developed for use in HCM 1985 (TRB, 1985), HCM 1994 (TRB, 1994) and HCM 2000 (TRB, 2000) in United States and the model developed by Teply et al. (1995) in Canada. William (1977) presented simple, accurate technique for measuring vehicular delay on the approach to a signalized intersection. Precise definitions were established for four measures of performance: stopped delay, time-inqueue delay, approach delay and percentage of vehicles stopping. Approach delay was selected as being the most representative of intersection efficiency. The values; thus, obtained were statistically compared with true values from time-lapse photography. The point sample, stopped delay procedure and the percentage of vehicle stopping method were selected as the most suitable methods for practical use and were performed in the field at three sites. Correction factors were developed to allow the field results to more accurately estimate the true values of stopped delay and percentage of vehicle stopping. Interrelationships among the four measures of performance were established so that approach delay can be estimated from a value of stopped time. Lin (1989) evaluated the reliability of the HCM 1984 procedure, based on field data, and discussed needed modifications. Stopped delay was measured for single lane movements at seven intersections. To compare the HCM estimates with observed delays, the cycle lengths, green durations, yellow durations, and saturation flow rates were also recorded using video cameras with built-in stopwatches. The evaluation reveals that, the procedure tends to overestimate stopped delay at reasonably well-timed signal operations. The discrepancies between the HCM estimate and the observed delays can be very large even when correct cycle length and green durations are used as inputs. Given actual 23 Chapter Two Literature Review cycle lengths and green durations, the procedure's ability to correctly identify the level of service was found to be good. He found that large discrepancies between HCM estimates and some of the observed delays could be reduced significantly if no progression adjustment is applied to the estimates obtained from HCM delay equation. Teply (1989) examined two approaches for measuring delay-- a timespace diagram and queuing diagram -- and explained various problems related to each. He concluded that, while delay cannot be measured precisely, it could be useful engineering tool if it is calculated properly. Some of his findings were: • Uniform delay formulas slightly underestimate overall delay because they neglect a portion of the acceleration delay. • The fact that uniform delay formulas and delay surveys based on queue counts do not account for the rate of arrival at the end of the queue, and the rate of discharge at the front of queue does not resort the resulting delay values. • Delay surveys based on stopped queue count produce stopped delay values. In situations with low volumes and short red interval, these techniques may overestimate stopped delay to the point of exceeding overall delay values. • The ratio between measured values of overall delay and stopped delay is not constant. Burrow (1989) and Akcelik (1990) presented generalized delay expressions for existing time-dependent delay models due to their similar forms. In general, they include a uniform delay term and an overflow or an incremental delay term given by Equation (2-5). where: 𝑑 = 𝑑 𝑢 + 𝑑𝑜 𝐸𝑞. (2 − 5) d = average total delay (seconds), 24 Chapter Two Literature Review 𝑑𝑢 = uniform delay (seconds) and, 𝑑𝑜 = overflow delay (seconds). In HCM 2000, average delay per vehicle for a lane group is given by equation (TRB, 1998): 𝑑 = 𝑑1 ∗ 𝑃𝐹 + 𝑑2 + 𝑑3 𝑑1 = 0.5C 𝑔 2 [1 − ] 𝐶 Eq. (2 − 6) 𝐶 𝑑2 = 900 𝑇 [(𝑋 − 1) + √(𝑋 − 1)2 + 𝑃𝐹 = Eq. (2 − 7) 𝑔 [1 − 𝑀𝑖𝑛 (1, 𝑋) ] 8𝐾𝐼𝑋 ] 𝑐𝑇 (1 − 𝑃)𝑓𝑝 where: 1− Eq. (2 − 8) Eq. (2 − 9) 𝑔 𝐶 d = average overall delay per vehicle (seconds/vehicles), d1 = uniform delay (seconds/vehicles), d2 = incremental, or random, delay (seconds/vehicles), d3 = residual demand delay to account for over-saturation queues that may have existed before the analysis period (seconds/vehicles), PF = adjustment factor for the effect of the quality of progression in coordinated systems, C = traffic signal cycle time (seconds), 25 Chapter Two Literature Review g = effective green time for lane group (seconds), X = volume to capacity ratio of lane group, c = capacity of lane group (vehicles/hour), k = incremental delay factor dependent on signal controller setting (0.50 for pretimed signals; vary between 0.04 to 0.50 for actuated controllers), I = upstream filtering/metering adjustment factor (1.0 for an isolated intersection), T = evaluation time (hours), P = proportion of vehicles arriving during the green interval, 𝑓𝑝 = progression adjustment factor. In cases where X > 1.0 for a 15-min period, the following period begins with an initial queue. This initial queue is referred to as Qb, in vehicles. Qb is observed at the start of the red period and excludes any vehicles in queue due to random, cycle-by-cycle fluctuations in demand (overflow queue due to cycle failures). When Qb ≠0, vehicles arriving during the analysis period will experience an additional delay because of the presence of an initial queue. The magnitude of this additional delay depends on several factors, including the size of the initial queue, the length of the analysis period, and the volume to capacity ratio during the analysis period. The initial queue delay term is designated which is a term in the delay model given in Equation (2-10). where: 𝑑3 = 1800 𝑄𝑏 (1+𝑢)𝑡 Eq. (2 − 10) 𝑐𝑇 Qb = initial queue at the start of period T (veh), c = adjusted lane group capacity (veh/h), 26 Chapter Two Literature Review T = duration of analysis period (h), t = duration of unmet demand in T (h), and u = delay parameter. Sofia (1998) studied the driver behavior and traffic flow in Mosul city. He develops delay model using regression analysis. Equations (2-11),(2-12) show the obtained regression models d = 0.39 C – 39 g/C +22.6 v/c -6.8 (R2=0.93) for v/c ≤ 1.0 d = 0.44 C – 83.5 g/C +20.7 v/c -3.0 (R2=0.91) for v/c > 1.0 Eq. (2 − 11) Eq. (2 − 12) Shamsul and Asif (2007) modified Webster’s delay formula under non lane based mixed road traffic condition by adding an empirical adjustment term with the sum of first and second terms, which has been calibrated based on field observations of delays. Hence, the general pattern of the modified Webster’s delay formula will be, as follows: 𝑑= 𝐶(1−𝜆)2 2(1−𝜆𝑋) + 𝑋2 2𝑞(1−𝑋) + 𝑎𝑑𝑗 Eq. (2 − 13) 𝑎𝑑𝑗 = 𝛽𝑜 + 𝛽1 ∗ 𝑞 + 𝛽2 ∗ 𝑐 + 𝛽3 ∗ 𝑥 + 𝛽4 ∗ 𝑙𝑎𝑚𝑑𝑎 + 𝛽5 ∗ 𝑝𝑛𝑚𝑣 Eq. (2 − 14) where: adj = Adjustment term for model. q = Vehicle arrival rate (PCU/sec). 𝑐= Cycle length (seconds). 𝑥= Degree of saturation. 27 Chapter Two Literature Review lamda = Effective green ratio. pnmv = Percentage NMV in traffic. β0 = Constant term in the model. β1 = Coefficient corresponding to q. β2 = Coefficient corresponding to 𝑐. β3 = Coefficient corresponding to 𝑥. β4 = Coefficient corresponding to 𝑙𝑎𝑚𝑑𝑎. β5 = Coefficient corresponding to 𝑝𝑛𝑚𝑣. In Equation (2-13), the adjustment term (𝑎𝑑𝑗) is to be calculated from delays observed in field. To make use of this term in the calibration, it must be based on the field value of delay and calculated from the Equation (2-15). 𝑎𝑑𝑗 = 𝑑 − 𝐶(1−𝜆)2 2(1−𝜆𝑋) + 𝑋2 Eq. (2 − 15) 2𝑞(1−𝑋) In Equation (2-15), 𝑑 is the actual delay observed in field. If the left hand side of the above equation is taken as the dependent variable, it needs to be regressed against a set of independent variables. Sierpiński(2007) proposed model for signalized intersections with queueing theory analytical models usage. The model is a sum of the average waiting time from the Clayton’s model and the average waiting time from the M+/G+/1queueing model that uses the compressed queueing processes. The proposed delay model has a form as Equation (2-16). 28 Chapter Two Literature Review 𝑑𝑀 = 𝐺𝑒 ] 𝑇𝑐 𝐺 2[1− 𝑒 𝜌] 𝑇𝑐 𝑇𝑐 [1− + 1 2 𝜆.𝜎µ2 +𝜆[𝜇−Δ] 2(1−𝜌) (1 − 𝜇Δ) Eq. (2 − 16) where: 𝑑𝑀 = average delay per vehicle(s). 𝑇𝑐 = length of cycle (s). 𝐺𝑒 = effective green signal duration (s). = flow ratio (-). = arrival rate (veh/s). 𝜇= service rate (veh/s). 𝜎𝜇2 = variance of service time (s). Δ= minimal time distance between vehicles (s). This formula makes a generalization of the Webster’s model (1958). Webster used the M/D/1 queuing model and his formula has got a form, as follow: 𝑑= 𝐺 2 𝑇𝑐 [1− 𝑇𝑒 ] 𝑐 𝐺 2[1− 𝑒 𝜌] 𝑇𝑐 + 𝜌2 2.𝜆(1−𝜌) 𝑇𝑐 − 0.65 [ 2 ] . 𝜌 𝜆 where: 𝑑 = average delay per vehicle (s). 𝑇𝑐 = length of cycle (s). 1 3 29 𝐺𝑒 ] 𝑇𝑐 [2+5 Eq. (2 − 17) Chapter Two Literature Review 𝐺𝑒 = effective green signal duration (s). = flow ratio (-). 𝜆= arrival rate (veh/s). 2.3.1.7 Queue at Signalized Intersection: Queue length is one of the most important performance measures of an intersection. Using queue length, other arterial performance measures, such as intersection delay, travel time and level of service can be estimated quite readily. For traffic engineers, these performance measures to provide indicators for identifying problems, thereby, helping decision makers improve the level of service from individual intersections to the entire road network (HCM,2000) When demand exceeds capacity at an approach to a signalized intersection at the start of an effective green period, a queue forms. Because of the arrival of vehicles during the red phases, some vehicles might not clear the intersection during the given green phase. Back of queue refers to the number of vehicles queued at an approach to a signalized intersection due to the arrival patterns of vehicles and to vehicles unable to clear the intersection during a given green phase (i.e., overflow). Most queuing theories are related to undersaturated conditions. To predict the characteristics of a queuing system mathematically, it is necessary to specify the following system characteristics and parameters (HCM, 2000): • Arrival pattern characteristics, including the average rate of arrival and the statistical distribution of time between arrivals; • Service facility characteristics, including service-time average rates and the distribution and number of customers that can be served simultaneously or the number of channels available; and • Queue discipline characteristics, such as the means of selecting which customer is next. 30 Chapter Two Literature Review In oversaturated queues, the arrival rate is higher than the service rate; in undersaturated queues, the arrival rate is less than the service rate. The length of an undersaturated queue can vary but will reach a steady state with the arrival of vehicles. By contrast, the length of an oversaturated queue will never reach a steady state, but will increase with the arrival of vehicles (HCM, 2000). An undersaturated queue at a signalized intersection is shown in Figure (2-8). The figure assumes queuing on one approach with two signal phases. In each cycle, the arrival demand is less than the capacity of the approach; where no vehicles wait longer than one cycle; and there is no overflow from one cycle to the next. Figure (2-8) (a) specifies the arrival rate, v, in vehicles per hour and is constant for the study period. The service rate, s, has two states: zero when the signal is effectively red and up to saturation flow rate when the signal is effectively green. Note that, the service rate is equal to the saturation flow rate only when there is a queue. Figure (2-8) (b) diagrams cumulative vehicles over time. The horizontal line, v, in Figure (2-8) (a) becomes a solid sloping line in Figure (2-8) (b), with the slope equal to the flow rate. Thus, the arrival rate goes through the origin and slopes up to the right with a slope equal to the arrival rate. Transferring the service rate from Figure (2-8) (a) to Figure (2-8) (b) creates a different graph. During the red period, the service rate is zero, so the service is shown as a horizontal line in the lower diagram. At the start of the green period, a queue is present, and the service rate is equal to the saturation flow rate. This forms a series of triangles, with the cumulative arrival line as the top side of each triangle and the cumulative service line form the other two sides. Each triangle represents one cycle length and can be analyzed to calculate the time duration of the queue. It starts at the beginning of the red period and continues 31 Chapter Two Literature Review until the queue dissipates. Its value varies between the effective red time and the cycle length, and it is computed using Equation (2-18) (HCM, 2000). vtQ s (tQ r ) or tQ sr s v where: Eq. (2 − 18) 𝑡𝑄 = time duration of queue (s), 𝑣 = mean arrival rate (veh/h), 𝑠 = mean service rate (veh/h), and 𝑟 = effective red time (s). Figure (2-8): Queuing Diagram for Signalized Intersection. 32 Chapter Two Literature Review The queue length is represented by the vertical distance through the triangle. At the beginning of red, the queue length is zero and increases to its maximum value at the end of the red period. Then, the queue length decreases until the arrival line intersects the service line, when the queue length equals zero. Three queue lengths can be derived using the relationship shown in Figure (2-8): the maximum queue length, the average queue length while queue is present and the average queue length. These are shown in Equations (2-19), (2-20), and (2-21), respectively (HCM, 2000). 𝑄𝑀 = 𝑄𝑄 = 𝑄= 𝑣𝑟 Eq. (2 − 19) 3600 𝑣𝑟 Eq. (2 − 20) 7200 𝑄𝑀 𝑡𝑄 Eq. (2 − 21) 2𝐶 where 𝑄𝑀 = maximum queue length (veh), 𝑄𝑄 = average queue length while queue is present (veh), Q = average queue length (veh), 𝑣 = mean arrival rate (veh/h), r = effective red time (s), 𝐶 = cycle length (s), and 𝑡𝑄 = time duration of queue (s). 33 Chapter Two Literature Review 2.3.1.7.1 Existing Queue Models: There are three types existing models of queue length: 1. Catling (1977) adapted existing equations of classical queuing theory to oversaturated traffic conditions and developed a comprehensive queue and number of stops estimates that can represent the effects of intersections in a time-dependent traffic assignment model. The time-dependent assignment model requires division of a peak period into time intervals that are chosen so that the mean arrival rate remains approximately constant within an interval. It is; therefore, obviously important to determine the sensitivity of estimated queue length on oversaturated traffic signal approaches to the method of dividing the peak period. Consider an approach to a traffic signal which has cycle time C. When the length of queue is increasing with time increasing, Catling's formula for computing the mean queue length is described by Equation (2-22). where: 𝐺 (𝑡) = [(𝛽 2 + 2. 𝑥 2. 𝑐 2. 𝑡 2 . 𝛼. 𝑧)1/2 − 𝛽]/𝛼 G(t)= the mean queue length at time t (vehicles), α= 2(c 𝑡 𝑧), β= c t [c 𝑡 (1𝑥) 2 𝑧 x], 𝑐 = s g /C, capacity (vehicles/hour), 𝑧 = 0.55, g = effective green time (seconds), 𝑞 = average arrival rate of traffic (vehicles/hour), 𝑠 = saturation flow rate (vehicle/hour), 𝑥 = qc/gs, the degree of saturation. 34 Eq (2 − 22) Chapter Two Literature Review In fact, when x < 1, the steady-state is represented by a limiting value of G(t) as t . This limiting value is denoted by G1(x) and can be expressed by Equation (2-23) : 𝐺1(𝑥 ) = lim 𝐺 (𝑡) = . 𝐶𝑥 2 Eq (2 − 23) 1−𝑥 These expressions can be applied only for an interval with stationary mean arrival rate and starting with zero queue length. If there is a different demand during subsequent time intervals, the formulas cannot be applied since the queue, formed at the end of the first interval, will cause extra delay to the following arrivals. 2.The Canadian Capacity Guide for Signalized Intersections, published in 1995, contains evaluation criteria for estimating queue lengths. These criteria recognize all three queue definitions: Uniform Average Queue Accumulation, Uniform Maximum Queue Reach and Uniform Conservative Approach, as illustrated in Figure (2-9). The Uniform Maximum Queue Reach is represented as a liberal approach to determine the required storage space, and the Uniform Conservative Approach is represented as a conservative approach. These two parameters converge at high levels of saturation. 35 Chapter Two Literature Review Figure (2-9): Graphical Representation of the queue accumulation and discharge process (Canadian Capacity Guide, 1995). The CCG model deals with the random and overflow effects in a unique manner. Most models treat the probability of spillover in a mathematically rigorous manner, as shown in Equation (2-24) (Canadian Capacity Guide, 1995). 1 𝑃 (𝑛𝑜 𝑠𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟) = (𝑠𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟) 𝑃 Eq (2 − 24) However, the CCG model formulates this probability using Equation (2-25). 𝑃 (𝑛𝑜 𝑠𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟) = [1 − 𝑃 (𝑠𝑝𝑖𝑙𝑙𝑜𝑣𝑒𝑟)]2 36 Eq (2 − 25) Chapter Two Literature Review This formulation, which expresses the probability of no spillover on either of two consecutive cycles, is an appropriate surrogate for the explicit treatment of the random and overflow phenomena. The method of the CCG model also includes a separate procedure for treating oversaturated conditions. The equations of queue length for undersaturated conditions and oversaturated conditions in the CCG model are described by Equations (2-26) and (2-27) respectively (Canadian Capacity Guide, 1995). 𝑄1𝐶𝐶𝐺 = where: 𝑄2𝐶𝐶𝐺 = 𝑞𝐶 3600 [𝑡𝑒 (𝑞−𝑐)] 60 + 𝑄1𝐶𝐶𝐺 Eq. (2 − 26) Eq. (2 − 27) 𝑄1𝐶𝐶𝐺 = estimates of average queue reach (vehicles), 𝑞 = arrival flow rate (vehicle/hour), 𝐶 = cycle length (seconds), Q2CCG = maximum queue reach during the congestion period (vehicles), 𝑡𝑒 = evaluation time (minutes), c = capacity (vehicles/hour). 3. HCM 2000 Queue Model: The queue length definition used in this model is the back of queue. A relationship for the back of queue is developed, as described in the next sections. The back of queue is the number of vehicles that are queued depending on arrival patterns of vehicles and vehicles that do not clear the intersection during a given green phase (overflow). The model predicts the average back of queue. 37 Chapter Two Literature Review The average back-of-queue measure is the basis to calculate percentile back of queue. Equation (2-27) shows average back-of-queue characteristics at signalized intersections. 𝑄 = 𝑄1 + 𝑄2 Eq. (2 − 27) where: 𝑄 = maximum distance in vehicles over which queue extends from stop line on average signal cycle (veh), 𝑄1 = first-term queued vehicles (veh), and 𝑄2 = second-term queued vehicles (veh). The first term, 𝑄1, is the average back of queue, determined first by assuming a uniform arrival pattern and then adjusting for the effects of progression for a given lane group. The first term is calculated using Equation (2-29). 𝑄1 = 𝑃𝐹2 𝑉𝐿 𝐶 3600 𝑔 [1 − ] 𝐶 𝑔 1 − [𝑀𝐼𝑁 (1, 𝑋𝐿 ) ] 𝐶 where: 𝑄1 = first-term queued vehicles (veh), PF2 = adjustment factor for effects of progression, 𝑉𝐿 = lane group flow rate per lane (veh/h), C = cycle length (s), 𝑔 = effective green time (s), and XL = ratio of flow rate to capacity (vL/cL ratio). 38 Eq. (2 − 29) Chapter Two Literature Review 𝑄1 represents the number of vehicles that arrive during the red phases and during the green phase until the queue has dissipated. The second term,𝑄2, is an incremental term associated with randomness of flow and overflow queues that may result because of temporary failures, which can occur even when demand is below capacity. This value can be an approximate cycle overflow queue when there is no initial queue at the start of the analysis period. Initial queue at the start of the analysis period is also accounted for in the second term, 𝑄2. Equation (2-30) is used to compute the second term of the average back of queue. 𝑄2 = 0.25𝑐𝐿 𝑇 [(𝑥𝐿 − 1) + √(𝑥𝐿 − 1)2 + 8𝐾𝐵 𝑋𝐿 16𝐾𝐵 𝑄𝑏𝐿 + ] (𝑐𝐿 𝑇)2 𝑐𝐿 𝑇 Eq. (2 − 30) where: 𝑄2= second term of queued vehicles, estimate for average overflow queue (veh); 𝑐𝐿 = lane group capacity per lane (veh/h); T = length of analysis period (h); 𝑥𝐿 = vL/cL ratio; 𝐾𝐵 = second-term adjustment factor related to early arrivals; 𝑄𝑏𝐿 = initial queue at start of analysis period (veh); and C = cycle length (s). 39 Chapter Two Literature Review 2.3.2 Two-Way Stop- Controlled Intersection (TWSC intersection) Unsignalized intersections are the most common intersection type. Although their capacities may be lower than other intersection types, they do play an important part in the control of traffic in a network. A poorly operating unsignalized intersection may affect a signalized network or the operation of an Intelligent Transportation System (Rod J. Troutbeck; Werner Brilon). There are three type of unsignalized intersection (TWSC, AWSC, and roundabout).TWSC intersections, in which stop signs are used to assign the right-of-way, are the most prevalent type of intersections within the United States and abroad. At TWSC intersections, the stop-controlled approaches are referred to as the minor-street approaches and can be either public streets or private driveways. The intersection approaches that are not controlled by stop signs are referred to as the major-street approaches. A three-leg intersection is considered to be a standard type of TWSC intersection as long as the single minor-street approach (i.e., the stem of the T-configuration) is controlled by a stop sign (HCM, 1998). 2.3.2.1 Delay at Two-Way Stop- Controlled Intersection The delay experienced by a motorist is made up of a number of factors that relate to control, geometrics, traffic, and incidents. Total delay is the difference between the travel time actually experienced and the reference travel time that would result during base conditions, in the absence of incident, control, traffic, or geometric delay. Average control delay for any particular minor movement is a function of the capacity of the approach and the degree of saturation. The analytical model used to estimate control delay (Equation 2-31) assumes that, the demand is less than capacity for the period of analysis. If the degree of saturation is greater than about 0.9, average control delay is significantly 40 Chapter Two Literature Review affected by the length of the analysis period. In most cases, the recommended analysis period is 15 min. If demand exceeds capacity during a 15-min period, the delay results calculated by the procedure may not be accurate. In this case, the period of analysis should be lengthened to include the period of oversaturation (HCM, 2000). 3600 𝑣 [ ][ 𝑥 ] 2 3600 𝑣𝑥 𝑣𝑥 𝑐𝑚.𝑥 𝑐𝑚.𝑥 √ 𝑑= + 900𝑇 [ −1+ [ − 1] + ] + 5 Eq. (2 − 31) 𝑐𝑚.𝑥 𝑐𝑚.𝑥 𝑐𝑚.𝑥 450𝑇 where: d = control delay (s/veh), vx= flow rate for movement x (veh/h), cm,x= capacity of movement x (veh/h), and T = analysis time period (h) (T = 0.25 for a 15-min period). A constant value of 5.0 seconds is used to reflect delay during deceleration. 2.4 Calibration of The Traffic Models Calibration is the process of comparing model parameters with real-world data to ensure that the model realistically represents the traffic environment. The objective is to minimize the discrepancy between model results and measurements or observations (TRB, 2005). Microscopic simulation model must be calibrated first against field data or against other validated analytical models to ensure accuracy. This crucial step of calibration is unfortunately often neglected. Sargeant and Christie (2002) compared observed values of intersection queue lengths with the simulation results of SYNCHRO/SimTraffic 5.0 model. 41 Chapter Two Literature Review Adjustments were made to vehicle/driver behavior to reduce the differences between observed and modeled queues. R-squared analyses were performed on the maximum and average queues to obtain estimates of the percentage variation explained by the simulation model. As both of these values indicate that, over 90 percent of the variation has been explained, it can be stated that the SYNCHRO/SimTraffic model is good. Al-Jaman (2007) successfully calibrated and validated the SimTraffic model for the signalized intersections of Riyadh, Saudi Arabia. The calibrated parameters are travel speed, turning speed, headway factor (a measure of saturation flow rate), and driver type. The study discovered that, the intersection approaches in Riyadh seem to have high saturation flow rates, which supports the results from other studies indicating that, driving behavior in Riyadh is aggressive. 2.5 Local Traffic Improvement Studies The overviews of traffic improvement studies use computer program tools and experimental studies. The type of network improvements might range from simple striping or parking changes to complete versions of geometry and signalization. This section provides a summary of the local evaluation comparisons, and experimental studies of the computer programs in some of Iraqi cities. 2.5.1 Mosul City Sofia (1998) studied the driver behavior and traffic flow in Mosul city. He used TRANSYT-7F program in performance analysis, evaluation, and coordination of seventeen signalized intersections. These intersections form two sub-networks, one in central business district (CBD) area, and the other in suburban area. 42 Chapter Two Literature Review Two proposals were made in his study to the CBD network and four proposals for the suburban network, after the development of statistical models for the estimation of passenger car unit (PCU), saturation flow, and delay. 2.5.2 Baghdad City Baghdad city is the capital of Iraq, where the governmental buildings, ministries, colleges, shopping centers, and so on are located in different sectors in the city. Each sector has its own traffic problems, and congestions, so several studies have been adopted to alleviate these problems. Both Al-Karkh and AlRusafa sides were involved in these studies. Each study, subnetwork, and software used is summarized in Table (2-3). Table (2-3) Some Traffic Improvement Studies in Baghdad City Sub-networks Researcher Area Studied Software Used Al-Hummadawee (2002) Al-Adhamiya district (from Bab-Al-Muadam to Ras AlHawash) Al-Adhamiya sector HCS-2000,and TRANSYT-7F Warid (2004) Al-Karrada network HCS-2000,and TRANSYT-7F Abdul-Ghani (2004) AL-Rusafa side (AL-Rashid, AL-khulafa, and AL-Kifah Street) CYCOPT, TRANSYT-7F, and HCS-2000 Al-Hemdany (2003) 43 TRANSYT-7F METHODOLOGY AND DATA COLLECTION Chapter Three Methodology and Data Collection Chapter Three Methodology and Data Collection 3.1 General The availability of highway transportation has provided several advantages that contribute to high standards of living. However, several problems related to the highway mode of transportation, exist. These problems include highway related parking difficulties, congestion, and delay. To reduce the negative impact of highways, it is necessary to adequately collect information that describes the extent of the traffic problems and identifies their locations. This chapter presents description of the study area (CBD) in Sulaymaniyah city. Brief description of the methods used for data collection and abstraction is also presented, with description of the computer programs used for traffic data abstraction and analysis. 3.2 Study Area Description The study area has three intersections within CBD area in Sulaymaniyah city, as shown in Figure (3-1). A brief description of selected intersections is presented below: Palace Intersection (1): It is four leg signalized intersection. Mawlawi approach is one way for exit only. Other approaches consist of 2-3 lanes of through, left and right movements. The Mamostayan3 approach is for right turn only. The geometric layout is illustrated in Figure (3-2). 44 Chapter Three Methodology and Data Collection Sara Intersection (2): It is five leg signalized intersection, three of which are one way; Mawlawi approach consists of two lanes of left turn only. Sabunkaran consists of three lanes of through and right turn movement and Kawa approach consists of three lanes of through and left turn movement. Bekas approach is two lanes for exit only and Pirmered approach is two ways which consists of lanes for U-turn movement only. The geometric layout is illustrated in Figure (3-3). Great Mosque Intersection (3): It is five leg unsignalized intersection. The Court approach consists of two lanes of right turn movement. Bekas approach consists of two lanes of through movement. Other approaches are for exit movement only. The geometric layout is illustrated in Figure (3-4). Figure (3-1): The Study Area in Sulaymaniyah City. 45 Chapter Three Methodology and Data Collection Figure Sar (3-3): Figure (3-2): Palace Intersection Layout. a Int ers ecti on La you t. Figure (3-3): Sara Intersection Layout. 46 Chapter Three Methodology and Data Collection Figure (3-4): Great Mosque Intersection Layout. 3.3 Data Collection The data collection phase aims at assembling all data required to model the traffic flow condition at the study area. The required data are collected in January, February/2012at morning and evening peaks. All required traffic data are collected during good weather, because adverse weather condition may cause variation in the normal traffic flow pattern. Figure (3-5) shows the Procedure for traffic and geometric data collection. The methods of data collection are described in the following paragraphs. 47 Chapter Three Methodology and Data Collection From Video Records &Field Observations Peak Hour Selection Study Area selection Traffic Data Collection Field Queue Length Actual Phase Length and Sequence From Video Records Traffic Signal control Actual Cycle Length and Timing Traffic & Geometric Data Survey Traffic Volume Survey Locations Identification From Manual Method Speed Data From Manual Method Intersection layout and Approach Dimensions & Satellite Image Geometric Data collection From Personal Observations and Field Measurements Figure (3-5): Traffic and Data Collection Procedure. 48 No. of Lanes for each approach Chapter Three Methodology and Data Collection 3.4 Methods of Data Collection 3.4.1 Manual Methods The following data are collected manually from field. 1. Queue length: two observers have collected Queue length data during the peak hours. The observation is made at some approaches since it is not possible to cover the full queue lengths due to obstruction of line of sight of video camera. 2. Geometric characteristics: Lane widths are measured at the stop line using measuring tape to obtain the width. All approach slopes for all intersections are obtained using the level instrument. Other data are collected by observation (e.g., number of lane per approach, lane designation and so on). 3. Spot speed: Spot speed data are also collected manually by two observers using a stop watch, marker to mark the pavement, and tape or measuring wheel for measuring section distance. 3.4.2 Video Recording Method The video recording method allows for large number of events to be recorded at the same time. It also has the advantage that any incident which might affect the observed data will also be recorded. The incidents can be reviewed at a later stage to resolve any apparent abnormalities in the data. All other traffic data, which are not collected manually, are collected by video camera based technique, which overcomes many of the difficulties in traffic data collection at field. 49 Chapter Three Methodology and Data Collection The video camera is mounted at the roof of a building located nearby the intersection and is focused to cover each leg of the intersection. Care is taken to cover full queue formed on the approach and the camera should be hidden from the drivers to avoid its effect on their behavior. The digital video camera, which is used in data collection, is SONY HANDYCAM, digital 8,990x digital zoom. 3.5 Geometric Data Geometric data, like intersection spacing and right of way, have been collected by using GIS tools in map measurements depending on the available satellite images with resolution of 0.6m, updated to 2007 and available at Sulaymaniyah municipality. Field measurements are conducted for approaches width, number of lanes per approach at each intersection, lane width, and splitter island width at all approaches. Field survey is used to obtain the geometric features that cannot be drawn from satellite image due to unavailability of updated one, and the main geometric features and layouts of the selected intersections are shown in Table (3-1). 3.6 Traffic Data The traffic data have been collected, as described below: 3.6.1 Traffic Volume Data The traffic volume data are collected for seven days by video recording technique for each intersection from 3:00 to 5:00 pm after several personal observations, and pilot survey have been made in the study area. Also, personal interviews are made with interested people like traffic policemen at the study area and different road users in order to determine the peak hour period for traffic data collection. The network average traffic volumes (veh/hr) are shown in Figure (3-6) for typical weekdays. 50 Chapter Three Methodology and Data Collection Predicted traffic volume is defined as "the current (existing) traffic volume multiplied by the ratio of future traffic volume to current traffic volume" (SORB, 2005). This ratio is named "Traffic Forecast Factor (TFF)" and can be calculated from Equation (3-1). where: 𝑇𝐹𝐹 = (1 + 𝐴𝑅 )𝑛 𝐸𝑞. (3 − 1) 𝑇𝐹𝐹: Traffic forecast factor (composed growth factor) 𝐴𝑅: Annual rate of traffic increase (%). 𝑛: Traffic analysis period (year). The annual growth rate of traffic volume increases in Sulaymaniyah city which is estimated to be (3.4%) of growth collected from the census department of Sulaymaniyah city for the previous 5 years (Traffic directorate of Sulaymaniyah governorate, 2013). 51 Chapter Three Methodology and Data Collection Table (3-1): Geometric Features for the Selected Intersections. Lane Approach Lane Name Name Group Palace Salim 1 LT+UT 3.2 1 Two way Salim 2 TH 3.2 1 Two way Salim 3 RT 4.5 1 Two way Baban 4 LT 2.4 1 Two way Baban 5 TH 2.4 1 Two way Baban 6 RT 2.4 1 Two way Mamostayan3 7 RT 3.7 2 Two way Mawlawi 1 LT 4.1 2 One way Kawa 2 TH 3.2 2 One way Kawa 3 LT 4.5 1 One way Sabunkaran 4 TH 4.8 1 One way Sabunkaran 5 RT 5.0 1 One way Piramerd 6 LT+UT 2.35 2 Two way Piramerd 7 RT 4.2 2 Two way Bekas 1 TH 12.2 2 One way Court 2 RT 10.5 1 One way Sara Great Mosque 52 Movement Width (m) No. of One Way Intersection Lanes or Two Way Chapter Three Methodology and Data Collection Figure (3-6): Hourly Traffic Volume for Average Typical Weekdays. 53 Chapter Three Methodology and Data Collection 3.7 Speed Data Speed measurement is carried out by video recording technique by measuring time taken by vehicles to pass trap length. The travel time can be converted to the speed, as illustrated in Figure (3-7), with a sample of speed calculation in Appendix(A).The average spot speed can be computed from the following Equation (Pignataro, 1973): 𝑆 = 3.6 × 𝐿 𝐸𝑞. (3 − 2) 𝛴 𝑡𝑖 /𝑛 where: 𝑆 : the average measured spot speed of vehicle (km/hr), 𝐿 : the segment length(m), 𝑡𝑖 : the time required for vehicle (𝑖) to transverse the section (sec), and 𝑛: sample size. The recommended trap length for calculating spot speed is shown in Table (3-2). This method involves the use of transverse pavement markings, which are placed at each end of the 25 m course at the midblock location and the observer starts and stops the watch as the vehicle passes the markings. 54 Chapter Three Methodology and Data Collection Figure (3-7): Average Spot Speed for Each Link in the Study Area. 55 Chapter Three Methodology and Data Collection Table (3-2): Recommended Trap Length (Kennedy, N. kell J.H. and Hombuger W.S.). Recommended Average speed of traffic stream (km/hr) Trap Length (m) Below 40 25 Between 40 and 64 50 Over 64 75 3.8 Saturation Flow Data Saturation flow rate is defined as the departure flow rate per lane at which vehicles can pass through a signalized intersection. Average saturation flow rate has been calculated– as recommended by HCM(2000) by Equation (3-3): 𝑠= 3600 𝐸𝑞. (3 − 3) ℎˉ where: 𝑠 = saturation flow rate (veh/h), and ℎˉ = saturation headway (s). 3.9 Traffic Signal Data For signalized intersection (Palace and Sara), the cycle length, phase length, green time and red time are measured from video films, as shown in Table (3-3).While, the value of lost time is taken as the default value (2.5 sec) recommended by SYNCHRO software program. 56 Chapter Three Methodology and Data Collection Table (3-3): Phase order, Phase and Cycle Length for Signalized Intersection. Intersection Phase length (sec) Palace Cycle length (sec) 300 178 118 Phase length (sec) Sara 196 96 96 3.10 Minimum Green Time for Pedestrian Pedestrian is one of the important elements which requires the attention by the traffic engineer, particularly in urban and CBD locations. Moreover, pedestrians also represent an element of sharp conflict with vehicular traffic which results in high accident rates and traffic delay. Therefore, minimum green time for pedestrian must be calculated and provided for each phase within each intersection. This study adopted the Highway Capacity Manual 2000 to estimate the minimum green time for all phases, as it can be found in the following Equations (TRB, 2005): 𝐺𝑝 = 3.2 + 𝐺𝑝 = 3.2 + where: 𝐿𝑝 𝑁𝑝𝑒𝑑 𝑆𝑝 + (0.81 𝑆𝑝 + ( 0.27 𝑁𝑝𝑒𝑑) 𝐿𝑝 𝑊𝐸 ) for 𝑊𝐸 > 3.0 m for 𝑊𝐸 ≤ 3.0 m 57 𝐸𝑞 (3 − 4) Chapter Three Methodology and Data Collection 𝐺𝑝: minimum green time (sec), 𝐿𝑝 : crosswalk length (m), 𝑆𝑝 : average speed of pedestrians (m/sec), 𝑊𝐸 : effective cross walk width (m), 3.2 : pedestrian start-up time (sec), and 𝑁𝑝𝑒𝑑 : number of pedestrians crossing during an interval (p) in sec. 3.11 Data Abstraction Data abstraction is based on sessions of 15 minute periods of recorded data. The selection of a 15 minute period is based on these considerations: 1. To ensure that, the sample is sufficient to provide meaningful results. 2. The traffic flow should be effectively constant during the period without incidents. 3. Minimization of observer fatigue and subsequent errors which might result from continuously watching video recordings. The video camera takes continuous picture of the traffic and the pictures are recorded on a digital camera. The video recording approach has a number of advantages: 1. It is unobtrusive and requires little labor power. 2. It produces permanent, complete record of the traffic scene. 3. Recording may be reanalyzed at any stage. 4. It provides account of each traffic event and maneuvers. The disadvantage of video technique is the large amount of time effort needed for data abstraction. The data extraction process is mainly achieved with the aid of a computer program named EVENT, which was developed by Al-Neami (2000). Many studies, used 58 this program (Sofia, 1998), (Al- Chapter Three Methodology and Data Collection Azzawi, 2003), (Khalaf, 2003), (Razzaq, 2005), (Al-Zubaidi, 2006), (Al- Dulaimi, 2006(, stated that, EVENT program is a necessary and complementary tool for traffic studies, which uses the video recording technique. It is worth mentioning that, the accuracy of the recorded data using EVENT is 0.01 second. It may be important here to describe the procedure of how this program works. The recorded film is replayed on a large screen TV monitor to extract the desired information. Different types of data are retrieved. At the same time, EVENT program runs- it is better that the screen of the computer on which EVENT program is operated, be near the TV screen in order to see them at the same time and each function bottom (Fi) is allocated for movement maneuver. See Appendix C for operation illustration. EVENT program output is accumulative time for successive vehicles departure, using Microsoft Excel program to enable the calculation of the following traffic parameters (Talabany, 2000): 1. The time headway between successive arriving vehicles; 2. The time headway between successive departing vehicles in queue. 3. The travel time for successive departing vehicles in queue. 4. Saturation flow data calculated by taking the reciprocal of the average saturation headway. 5. The frequency distribution of the observed departure and arrival headway. 6. The average observed delay of an approach. 7. Duration of observed data session. 8. Vehicle arrival and departure flow. Sample of EVENT program output is shown in Appendix C . 59 Chapter Three Methodology and Data Collection 3.12 Queue Length The observation were made from replaying video films on each lane for all approaches to collect the queue length during peak hour periods. The observed data for the average queue length during red phase, are as shown in Table (3-4). Table (3-4): Average Observed Queue Length. Palace Intersection Average Q.L. Sara Intersection Average Q.L. Sailm (TH) 12(veh) Mawlawi (LT) 16(veh) Salim (LT+UT) 13(veh) Kawa (TH) 15(veh) Baban (LT) 25(veh) Kawa (LT) 13(veh) Baban (TH) 23(veh) Sabunkaran Piramerd (TH) (LT+UT) 20(veh) 13(veh) 3.13 Data Processing The abstracted data is then processed using Microsoft Excel program and output data files from EVENT program. These programs are used to measure the main traffic parameters that are illustrated in Table (3-5). Table (3-5): Summary of Abstracted Data. Major Category Data Type • Through and turning traffic Traffic volume data volume counts • Headways and saturation flow rate Driver performance characteristics • Gaps and follow-up time data • Cycle length, green time , red time Signal timing control • Phase length and sequence • Delay Data used in comparison with • Queue length simulated results 60 Chapter Three Methodology and Data Collection The output file from EVENT program is opened with the aid of EXCEL program. This program has an option which splits text data to columns. This method separates the recording time in column and the associate character in a next column. This arrangement makes dealing with the data easier by sorting them into digits and characters. For traffic volume data calculation, Pivot table is used to calculate the number of each character presented in the selected column. Other time-related parameters (saturation headway, phase and cycle length, and delay times) are calculated by arranging EVENT output file in a set of columns according to their characters. Then, the subtraction of one digit from another is done to find the value (time) for the subjected parameters. Samples of each parameter calculation are given in Appendix (C). 3.14 Delay Measurement Method suggested by HCM 2000 is based on direct observation of vehicles-in-queue counts at the intersection. This method does not directly measure delay during deceleration and during part of acceleration, which are very difficult to measure without sophisticated tracking instrument. However, this method has been seen to yield a reasonable estimate of control delay. The method includes an adjustment for error which may occur when this type of sampling technique is used, as well as an acceleration/deceleration delay correction factor. A sample worksheet, used for recording retrieved data, is included in Appendix (B). The survey period should begin at the start of the red phase of the approach, ideally when there is no cycle failure (no overflow queue) from the 61 Chapter Three Methodology and Data Collection previous green period. Recorded films are replayed to retrieve data for vehicle delay time. The procedure adopted to retrieve data is summarized below. • The moment signal turns to red, cassette is paused and VCP timer is set to zero. The overflow queue has been excluded from queue counts. This is due to the need for consistency with the analytical delay equation, which is based on delay to vehicles that arrive during the survey period. This time period may differ from analysis period which is typically considered as 15 minutes as per HCM(2000), because all the vehicles that join the queue within this analysis period should be included in queue count until they cross the stop line. • Cassette is played and the number of vehicles in queue is recorded at regular interval of 10 to 20 seconds (As per HCM 2000). The regular interval should not be an integral divisor of the cycle length. Meanwhile, it is necessary to keep track of end of standing queue by observing the last vehicle in those stops because of signal. This includes vehicles arriving when the signal is actually green, but stopped because vehicles in front have not yet started moving. The vehicles in queue counts often include some vehicles that have regained speed, but have not yet exited the intersection. • Count of vehicles-in-queue is done at regular interval for analysis period of about 15 minutes. End of the survey period must be clearly defined, since the last arriving vehicle (s) that stop in the period must be clearly defined and counted until they exit the intersection, per the next step. Stopping vehicles that arrive after the end of the analysis period are not included in the final vehicle-in-queue counts. 62 Chapter Three Methodology and Data Collection • Volume counts of total vehicles (𝑉𝑡𝑜𝑡 ) arriving during the survey period, and total vehicles arriving during the survey period that stop one or more times. Vehicles stopping multiple times are counted only once as a • stopping vehicle (𝑉𝑠𝑡𝑜𝑝 ) as per HCM(2000) delay measurement guideline. The average time-in-queue per vehicle arriving in the survey period is estimated as (HCM,2000) : where: Delay time-in-queue per vehicle, 𝑑𝑣𝑞 = (𝐼 ∗ ∑𝑉𝑖𝑞 𝑉𝑡𝑜𝑡 ) * 0.9 𝐸𝑞. (3 − 5) 𝐼 = interval between vehicle-in-queue counts, (s) 𝑉𝑖𝑞 = sum of vehicle-in-queue counts, (veh) 𝑉𝑡𝑜𝑡 = total number of vehicles arriving during the survey period, (veh) 0.9 = an empirical adjustment factor accounts for the errors that may occur when this type of sampling technique is used to derive actual delay values, which normally results in an overestimate of delay (As per HCM 2000). • Next, the fraction of vehicles stopping and the average number of vehicles stopping in a queue in each cycle are computed. 𝐹𝑉𝑆 = Fraction of vehicles stopping = 𝑉𝑠𝑡𝑜𝑝 𝑉𝑡𝑜𝑡 𝐸𝑞. (3 − 6) • Correction factor given by HCM is selected based on average free flow speed (that is measured at the upstream of the selected approaches) and average number of vehicles stopping per queue in each cycle. The values of correction factor are given Table (3-6). 63 Chapter Three Methodology and Data Collection • The fraction of vehicles stopping is multiplied by correction factor and the product is added to the time-in-queue value to obtain the final estimate of control delay. Acceleration and deceleration delay: 𝑑𝑎𝑑 = 𝐹𝑉𝑆 * 𝐶𝐹 𝐸𝑞. (3 − 7) Control Delay/vehicle, 𝑑 = 𝑑𝑉𝑞 + 𝑑𝑎𝑑 𝐸𝑞. (3 − 8) Control delay includes initial deceleration delay, queue move-up time, stopped delay, and final acceleration delay. Table (3-6) Acceleration/Deceleration Delay Correction Factor (CF) – second (HCM, 2000). Free-Flow ≤7 vehicles 8-19 vehicles 20-30 vehicles ≤ 60km/h +5 +2 +1 >60-71 km/h +7 +4 +2 >71 km/h +9 7 +5 Speed 3.15 SYNCHRO / SimTraffic SYNCHRO V.8 is a complete software package for modeling and optimizing traffic signal timings. This software includes: SYNCHRO : a macroscopic analysis and optimization program; SimTraffic: a powerful, easy-to-use traffic simulation software application. It is designed to model networks of signalized and unsignalized intersections, including roundabouts. The primary purpose of SimTraffic is to check and fine tune traffic signal operations. 64 Chapter Three Methodology and Data Collection 3D Viewer, a three-dimensional view of SimTraffic simulations; SimTraffic CI, an application that interacts with a controller interface (CI) device connected to a controller to simulate the operation of the controller with simulated traffic. The key features of SYNCHRO include: • Capacity Analysis. SYNCHRO provides a complete implementation of the 2010 Highway Capacity Manual, Chapter 18. • Coordination: SYNCHRO allows quickly generating optimum timing plans to minimize delays. • Actuated Signals: SYNCHRO is the only interactive software package to model actuated signals. SYNCHRO can model skipping and gap-out behavior and apply this information to delay modeling • Time-Space Diagram: SYNCHRO has colorful, informative TimeSpace Diagrams. Splits and offsets can be changed directly on the diagram. • Integration with SimTraffic, CORSIM and HCS: SYNCHRO features are preprocessors to these software analysis packages. Data are entered once with easy-to-use SYNCHRO, and then analyses are performed with these software packages. SimTraffic V.8 is microscopic stochastic simulation model used in conjunction with SYNCHRO. The following points briefly provide an overview of SimTraffic: • It incorporates vehicle and driver performance characteristics developed by FHWA. 65 Chapter Three Methodology and Data Collection • It is capable of simulating different types of intersection controls, various street geometries including lane drops and turning pockets, and a wide range of traffic flow conditions. • It uses “car following” logic for simulation. • It can be calibrated for local driver/traffic behavior by headway factor as input in SYNCHRO, and vehicle\driver behavior adjustment in SimTraffic. • It is not capable of modeling ramp metering, bus stops, bus routes, bus and carpool lanes, light rail, on-street parking and short-term events. 66 DATA PRESENTATION AND ANALYSIS Chapter Four Data Presentation and Analysis Chapter Four Data Presentation and Analysis 4.1 General This chapter presents the analysis of the collected data, calibration of the output results of the software used for data analysis, the development of delay and queue statistical models for signalized intersections to evaluate traffic performance and existing geometric design. Improvement of geometric features is also proposed and the results are evaluated using traffic software. 4.2 Evaluation of Existing Traffic Flow To investigate the existing conditions of traffic operation schemes and geometry of studied intersections, simulation of the actual movement is required to evaluate traffic performance. SYNCHRO Plus 8 is applied to simulate the existing traffic flow for three intersections at the selected network. It is a complete software package for modeling, optimizing, managing and simulating traffic systems. The initial release of this software occurred in June 2011. SYNCHRO models streets and intersections as links and nodes. These links and nodes are created in the map view. Every intersection to be analyzed in the study area is represented by a node. These links and nodes are created by using GIS tools in map measurements depending on the available satellite images with resolution of 0.6m, updated to 2007 and available at Sulaymaniyah municipality, like intersection spacing. See Figure (4-1). 67 Chapter Four Data Presentation and Analysis Figure (4-1): Network Layout as Printout by SYNCHRO. The data like geometric characteristics of approaches (number of lanes per approach at each intersection, lane width and splitter island width) and traffic characteristics (traffic volume, cycle length, link speed, phase sequence, pedestrian volume and grade) are required inputs to the program. All values are entered in easy-to-use forms. Calculations and intermediate results are shown on the same forms. Figure (4-2) shows process of entering lane setting (See Appendix D for more details). 68 Chapter Four Data Presentation and Analysis Figure (4-2): Lane Settings by SYNCHRO. Table (4-1) indicates the output results of simulation runs which include the degree of saturation, total delay and level of service for each intersection in the study area depending on existing traffic conditions. 69 Chapter Four Data Presentation and Analysis Volume (veh/hr) Movement Street Average Delay for (sec/veh) Level of Service 1069 0.95 71.5 TH 809 1701 0.95 71.7 RT )permitted) 656 1265 0.56 1.8 A LT 680 1057 1.27 204 F TH 440 1642 0.95 116.3 RT )permitted) 260 1173 0.24 0.5 RT )permitted) 680 2070 0.34 0.5 0.5 A A LT 614 824 0.68 41.1 41.1 D D LT 179 713 0.49 40.1 Sabunkaran Piramerd Great Mosque* * E E 135.7 71.9 52.9 D F F D 34.4 TH 312 2052 0.32 30.2 TH 130 1080 0.56 42.9 E A C C 31.9 Baban Salim 805 Mamostayan3 v/c LT+UT Kawa Palace* Sara* Saturation Flow Rate (veh/hr) Mawlawi Intersection Table (4-1): Total Delay and Level of Service for Isolated Intersections Produced by SYNCHRO for Existing Condition. C D 22.3 C RT )permitted) 224 803 0.34 1.2 A LT+UT 368 1293 0.44 36.7 D RT )permitted) 100 772 0.17 0.5 27.5 C A Max. v/c Average Delay for (sec/veh) Level of Service 1.90 158.8 F *Signalized Intersection **Unsignalized Intersection (Two Way Stop Control) 70 Chapter Four Data Presentation and Analysis The results in Table (4-1) show that, most of the intersections suffer from congestion with high total delay values. All the output results are categorized by medium to high total delays, thus the level of service for most of the intersections is (E, F) according to the LOS categories of the software based on HCM method. 4.3 SYNCHRO Model Calibration Calibration is the process by which the individual components of the model are refined and adjusted so that the model accurately represents field measured or observed traffic conditions (Turley, 2007). With regards to calibration, traffic models contain numerous variables to define and replicate traffic control operations, traffic flow characteristics and driver behavior. The model contains default values for each variable, but also allows a range of userapplied values for each variable. The measures of effectiveness estimated by the models such as queue length and delay are used for comparison with local traffic conditions in order to ensure that the model accurately represents the measured traffic parameters. The delay time estimated by models is mostly used by traffic engineers to evaluate intersection performance. For this study, the delay has been selected to be compared with existing traffic conditions and the values of the selected parameters have been manipulated until an acceptable convergence between the observed and simulated delay values has been reached. Unfortunately, the user manual of SYNCHRO does not provide substantial guidance on how the user should modify these parameters for different types of conditions. Therefore, the user has a great responsibility for ensuring that the appropriate changes made are based on field measured data and not on engineering judgment. 71 Chapter Four Data Presentation and Analysis The main parameter at signalized intersection that depends on driver behavior is the saturation headway. This parameter affects mainly the saturation flow rate. Saturation flow rate is the maximum departure (queue discharge) flow rate achieved during the green period at traffic signals. SYNCHRO saturation flow estimation is affected by many factors including lane width, grade, turning radius, parking, buses, pedestrian interference, percentage of heavy vehicles, shared lanes, opposed turns and short lanes. Furthermore, saturation flows estimated by SYNCHRO are dependent on signal timings. The default value of the basic saturation flow rate used by SYNCHRO model is 1900 pcuphgpl. This is applied at the signalized intersections in the study area. The delay time produced by SYNCHRO for each approach is compared with the average delay times measured from the field. The results of the comparison between measured and estimated delays are shown in Figure (4-3). This Figure explains that, for high values of v/c , SYNCHRO highly over estimates the delay time. The approaches field delay are regressed against the predicted ones producing the results shown in Table (4-2). The R² value indicates that, SYNCHRO explains about 77 % of the variability in the delay times. A comparison is made by using the paired samples t-test. The paired samples t-test is made by using SPSS V.20 program. The paired t-test results, as in Table (4-2), show that, the mean of delay differences is 14.7 seconds with a P-value of 0.281, so the hypothesis that the deviations are equal to zero is accepted at 95% confidence. In order to reduce standard error mean of delay, the model`s parameters should be adjusted. The calibration process is to adjust the basic saturation flow rate to be 2200 pcuphgpl in lane groups with high values of v/c, as recommended by 72 Chapter Four Data Presentation and Analysis HCM. The results show 93.3 % variability in delay times, See Figure (4-4). The paired t-test results show that, the mean of delay differences of 5.322 seconds is lower than the difference obtained by using the default value of saturation flow, with a P-value of 0.455. So, the hypothesis is that there is no significant difference accepted at 95% confidence. Table (4-2) Comparison of Actual and SYNCHRO Delay Prediction for Signalized Intersections. Default Values of Saturation flow rate (1900 veh/hr) Values after adjustment of Saturation flow rate (2200 veh/hr) Field delay 0.501* SYNCHRO Delay 0.808*SYNCHRO Delay R2 0.772 0.933 Adj.R2 0.744 0.924 Sig. 0.001 0.000 Paired Sample t-test Mean Difference in Delay (sec) 14.713 5.322 t 1.157 0.784 p 0.281 0.455 All p-value greater than 0.05, so that there is no statistically significant difference 73 Chapter Four Data Presentation and Analysis Figure (4-3): SYNCHRO Average Delay Values versus Field Delay Values for All Approach. Figure (4-4): SYNCHRO Delay Values versus Field Delay Values after Calibrating the Basic saturation flow rate for All Approach. 74 Chapter Four Data Presentation and Analysis 4.4 Development of Statistical Models for Delay and Queue Length The traffic data, which represent many variables are extracted from video records to be used in the process of building statistical models. These data include; traffic and geometric characteristics. The next paragraphs describe the statistical techniques used for the models development are required for the prediction of delay and queue length. 4.4.1 Process of Models Building The following steps, which are recommended by many researchers (Keller and Warrack, 2000), are followed in this study; a. Identifying the dependent variables. b. Listing potential predictors. c. Gathering the required observations for the potential models. d. Using statistical software to estimate the models. e. Determining whether the required conditions are satisfied. f. Using the engineering judgment and the statistical output to select the best models. 4.4.2 Identification of Dependent and Predictor Variables The variables, which are used to simulate traffic and geometric characteristics, are as follows: 1. dependent variables: g = Effective green time for movement or lane group (sec), c = Capacity of lane group (veh/h), C = Cycle length (sec), v = Vehicular flow rate (veh/h), w =Width of lane group at stop line (m). 75 Chapter Four Data Presentation and Analysis Ws = Total width of lane groups departing to the same exit roadway, in the same phase at stop line (m). We = Total width of exit roadway of intersection for traffic departing straight forward (m), 2. potential predictors: q = Average number of vehicles in queue (veh), d = Delay for lane group (sec/veh). These data are shown in Appendix E. 4.4.3 Data Analysis The data analysis includes the following steps: 4.4.3.1 Selecting Sample Size To determine the required sample size, (Kennedy and Neville, 1986) presented the following equation to calculate percent of error according to sample size: E=VT/ (n) 0.5 …………………Equation (5-1) Where: E=Error of mean, V=Coefficient of variation (c.o.v), t=t-statistics, and n= Sample size Therefore, it is suggested to use the whole data for the delay model building with a (0.026133) percentage of error, and the queue model building with a (0.050) percentage of error. 4.4.3.2 Scatter Plots Scatter plot is carried out between the dependent and independent variables for the requirements of the modeling process. The nature of relation 76 Chapter Four Data Presentation and Analysis between dependent and independent variables can be expected and the best relations are selected depending on resulted plots. 4.4.3.3 Outliers If one or more of observations is different significantly from all others, it is called “outlier“. The cause of a faulty observation may be a mistake(Kennedy and Neville, 1975). Outliers and influential observations are checked by using SPSS software (V.20).The results of this test can be found in Table (4-3).It can be seen that there is no outliers because there are no extreme values. Table (4-3) (a): Case Processing Summary for Outliers for Field Delay (SPSS, V.20). Cases Field delay Valid % Missing % Total % 100.0% 0.0% 100.0% Table (4-3) (b): Case Processing Summary for Outliers for Field Queue of Left Turn Movement (SPSS, V20). Cases Field queue Valid % Missing % Total % 100.0% 0.0% 100.0% 77 Chapter Four Data Presentation and Analysis Table (4-3) (c): Case Processing Summary for Outliers for Field Delay of Left Turn Movement (SPSS, V.20). Cases Field queue Valid % Missing % Total % 100.0% 0.0% 100.0% 4.4.3.4 Testing of Normality There are several methods of assessing whether data are normally distributed or not. The Kolmogorov-Smirnov test (K-S) and Shapiro-Wilk (SW) test are designed to test normality by comparing the data to a normal distribution with the same mean and standard deviation of the sample. If the test is not significant, then the data are normal, so any value above 0.05 indicates normality. If the test is significant (less than 0.05), then the data are non-normal. Thus, results of tests in Table (4-4) show that all results above 0.05 follow normal distribution. 78 Chapter Four Data Presentation and Analysis Table (4-4) (a): K-S test and (S-W) test Results for Field Delay (SPSS, V.20). Kolmogorov-Smirnov Field delay Shapiro-Wilk Statistic Sig. Statistic Sig. 0.199 0.122 0.928 0.258 Table (4-4) (b): K-S test and (S-W) test Results for Field Queue of Left Turn Movement of All Approaches (SPSS, V20). Kolmogorov-Smirnov Field Queue Shapiro-Wilk Statistic Sig. Statistic Sig. 0.117 0.200 0.967 0.804 Table (4-4) (c): K-S test and (S-W) test Results for Field Queue of Through Movement of All Approaches (SPSS, V20). Kolmogorov- Shapiro-Wilk Smirnov Field Queue Statistic Sig. Statistic Sig. 0.228 0.151 0.895 0.194 79 Chapter Four Data Presentation and Analysis 4.4.3.5 Multicollinearity Multicollinearity (collinearity and intercorrelation) is a statistical procedure to find the correlation between independent variables. The adverse effect of multicollinearity is that the estimated regression coefficients (b1, b2, etc.) tend to have large sampling variability. The basic methodology followed in this procedure is that, predictor variables are eliminated one – by – one based on multicollinearity analysis. With multicollinearity control, the process is repeated eliminating predictor variables remaining. At that point, interactions among the variables are considered; again the process is iterated discarding terms based on significance. A correlation matrix is produced to determine the correlation coefficients for the variables. In many cases, the correlation matrix will not be able to identify whether multicollinearity is a serious problem because there are many ways for variables to be related (Keller and Warrack, 2000). SPSS software (Ver.20) is employed for the development of the models. A confidence level of 95 percent is adopted, thus a significant level (α = 0.05) is employed. By using SPSS software, the correlation coefficients between all of the variables are calculated and the correlation matrix is setup. This matrix can be seen in Tables (4-5), (4-6) and (4-7) for delay and queue models. Then, the independent variables having the highest correlation coefficient with the designated dependent variables are selected and calculated, and the regression equation is formulated. 80 Chapter Four Data Presentation and Analysis Table (4-5): Correlation Coefficient Matrix for Delay Model. Field delay Cycle (sec) g/C v/c Field delay 1.0 Cycle (sec) 0.854 1.0 g/C -0.022 0.351 1.0 v/c 0.759 0.643 -0.016 1.0 We/Ws 0.659 0.697 0.171 0.719 We/Ws 1.0 Table (4-6): Correlation Coefficient Matrix for Queue Length of Left Turn Movement. Field Queue (sec) g/C v/c Width (m) Field Queue (sec) 1.000 g/C 0.05 1.000 v/c 0.870 -0.118 1.000 Width (m) -0.389 -0.127 -0.285 1.000 Cycle (sec) 0.829 0.351 0.666 -0.326 81 Cycle (sec) 1.000 Chapter Four Data Presentation and Analysis Table (4-7): Correlation Coefficient Matrix for Queue Length of Through Movement. Field Queue v/c g/C Width(m) Cycle(sec) Field Queue 1.000 v/c 0.831 1.000 g/C -0.144 -0.256 1.000 Width(m) -0.624 -0.520 -0.053 1.000 Cycle)sec) 0.854 0.664 0.292 -0.544 1.000 4.4.4 Regression Modeling Regression analysis is a statistical method that uses the relationships between two or more variable quantities to generate a model that may predict one variable from the others. The term multiple linear regression (MLR) is employed when a model is a function of more than one dependent variable. The objective behind (MLR) is to obtain adequate models, at a selected confidence level, using the variable data, while at the same time satisfying the basic assumptions of regression analysis, which are: • Severe multicollinearity does not exist among predictor variable • Influential observations or outliers do not exist in the data. • The distribution of error is normal. • The mean of error distribution is zero. The objective is accomplished by selecting the model, which provides the highest adjusted coefficient of determination (R²) and the lowest mean square error (MSE) (Keller and Warrack, 2000). 82 Chapter Four Data Presentation and Analysis 4.4.4.1 Stepwise Regression Procedure This procedure begins by computing the simple regression model for each independent variable. The independent variable with the largest f- statistic is chosen as the first entering variable. SPSS software uses the f- statistics and the standard is usually set at F = 3.8 which is chosen because the significant level is about 5%. The standard is called the f- to enter. If at least one variable exceeds the standard, the procedure continues, then the software will examine if the variables verify the condition of dependent variables remaining in the model or not. It then considers whether the model would be improved by adding a second independent variable. It examines all such models to determine which is the best and whether the f- statistic of the second variable (with the first variable already in the equation) is greater than fto enter (Ahmed, 2002). If two independent variables are highly correlated, only one of them will enter the equation. Once the first variable is included, the added explanatory power of the second variable will be minimal and its f-statistic will not be large enough to enter the model. In this way, multicollinearity is reduced. The procedure continues by deciding whether to add another independent variable to each step. The p-values of all variables are computed (at each step) and compared with the f- to be removed. If a variable f – statistic falls below this standard, it is removed from the equation. These steps are repeated until no more variables are added or removed. 4.4.4.2 Developed Models The summary of the stepwise regression and possible developed models is shown in the following articles for delay and queue length models. 83 Chapter Four Data Presentation and Analysis 4.4.4.2.1 Delay Model The summary of stepwise regression delay model can be seen in Table (4-8). Tables (4-8): Stepwise Regression Models Summary for Delay Model. Unstandardized Coefficients Model 1 Standardized Coefficients t Sig. R2 B Std. Error Beta Cycle 0.102 0.020 0.457 5.192 0.000 (1-We/Ws) 19.594 3.622 0.140 5.409 0.000 0.995 v/c 30.190 5.648 0.462 5.345 0.000 Adj.R2 SEE 0.993 4.175 The results of regression technique and selected developed delay model are shown in Equation (4-1) d = 0.102 C +30.19 v/c +19.59 (1-We/Ws) 𝐸𝑞. (4 − 1) The variables that have significant impact on total delay are explained in Table (4-9) for the linear regression model, with statistical characteristics. The minimum and maximum traffic parameters limits represent the range at which the model will be applicable. 84 Chapter Four Data Presentation and Analysis Table (4-9): Data Range and Statistical Characteristics for Delay Model. Minimum Maximum Mean Std. Deviation g/C 0.370 0.610 0.48253 0.068641 v/c 0.50 1.15 0.7395 0.230800 Cycle (sec) 97.0 300.0 215.067 70.1310 We/Ws 0.458 1.000 0.7298462 0.271017 4.4.4.2.2 Left Turn Queue Model The summary of stepwise regression model for queue length of left turn movements can be seen in Table (4-10). Table (4-10): Stepwise Regression Models Summary for Queue Length of Left Turn Movements. Unstandardized Standardized Coefficients Coefficients Model t B 2 Std. Error Sig. R2 Adj.R2 SEE Beta (Constant) -9.482 2.912 -3.256 0.007 v/c value 18.223 4.488 0.571 4.061 0.002 0.868 0.846 Cycle 0.051 0.016 0.449 3.195 0.008 3.1479 The results of regression technique and selected developed model are shown in Equation ( 4-2 ). q =0.051 C+18.223 v/c -9.48 85 𝐸𝑞. (4 − 2) Chapter Four Data Presentation and Analysis The variables that have significant impact on total delay are explained in Table (4-11) for the linear regression model, with statistical characteristics. The minimum and maximum traffic parameters limits represent the range at which the model will be applicable. Table (4-11): Data Range and Statistical Characteristics for Queue Length Model of Left Turn Movement. Variable Minimum Maximum Mean Std. Deviation v/c value 0.500 1.172 0.75427 0.251457 Cycle (sec) 97.0 300.0 215.067 70.1310 4.4.4.2.3 Through Queue Model The summary of stepwise regression model for queue length of through movements can be seen in Table (4-12). Tables (4-12): Stepwise Regression Models Summary for Queue of Through Turn Movements. Model Unstandardized Coefficients B (Constant) -11.618 3 Std. Error Standardized Coefficients t Sig. -2.755 0.025 R2 Adj.R2 SEE 0.853 0.816 4.10304 Beta 4.218 Cycle 0.071 0.024 0.540 2.978 0.018 v/c value 13.330 5.124 0.472 2.601 0.032 86 Chapter Four Data Presentation and Analysis The results of regression technique and selected developed models can be seen in Equation (4-3). q = 0.071 C+13.33 v/c -11.618 𝐸𝑞. (4 − 3) The variables that have significant impact on queue length of through movements are explained in Table (4-13) for the linear regression model, with statistical characteristics. The minimum and maximum traffic parameters limits represent the range at which the model will be applicable. Table (4-13): Data Range and Statistical Characteristics for Queue Length Model of Through Movement. Variable Minimum Maximum Mean Std. Deviation v/c value 0.15 1.30 0.7607 0.33875 Cycle (sec) 97 300 225.73 72.874 4.5 Models Validation Validation process is determining whether the selected model is appropriate for the given conditions and for the given task; it compares model prediction with measurements or observations (TRB, 2005). The objective of validation is to assess the adequacy of the proposed prediction models, and measure the error or accuracy of the prediction for the validation period. There are several methods used for models validation. One of these methods is to compare the model with another data set that is not included in model building. 87 Chapter Four Data Presentation and Analysis The data used for this purpose is one hour data abstracted from video recording films at different times for the same intersections in the network. The average field delay from one hour is regressed with the delay time predicted by the model. The regression results are shown in Figures (4-5), (4-6) and (4-7) and Table (4-14). It can be concluded from the models values of R 2 that, the predicted values from models can represent the actual field values of delay and queue length. Table (4-14): Regression Results for the Three Models. Model 1- Delay Model 2- Left Queue Model 3-ThroughQueue Model Model Fit R 2value 2 Adj. R -value Sig. 0.958 * Field Delay + 3.36 0.922 0.911 0.00 1.139 * Field Queue – 2.24 0.906 0.874 0.013 1.058 * Field Queue – 1.944 0.956 0.942 0.004 88 Chapter Four Data Presentation and Analysis Figure (4-5): Observed versus Predicted Delay. Figure (4-6): Observed versus Predicted Queue Length for the Model of Left Turn Movement. 89 Chapter Four Data Presentation and Analysis Figure (4-7) :Observed versus Predicted Queue Length for the Models of Through Movement. 4.6 Models Analysis The delay model depends on many variables such as cycle length(C), degree of saturation(v/c) and Total width of exit roadway of intersection for traffic departing straight on divided by total width of lane groups departing to the same exit roadway, in the same phase (We/Ws) while queue length depends on cycle length and degree of saturation. From the previous models of delay and queue length, it is obvious that, delay time and queue length increase as cycle time and v/c value increase. The increase in delay is attributed mainly to the increase in the effective red and hence in waiting time for vehicles before the start of green. The delay time has a proportional relationship with the degree of 90 Chapter Four Data Presentation and Analysis saturation (v/c) which is based on flow rate and proportion of cycle time effectively green (g/c).The minimum of g/c and large flow rate cause high v/c value. Delay time has an inverse relationship with (We/Ws). It is obvious that delay time increases when (We/Ws) is less than 1.0. The limit of (We/Ws) adopted in the model should be ≤ 1.0, as in Table (4-9) for lane group width not for the whole approach. 4.7 Comparison of Observed Delay with HCM 2000 Delay Values The 2000 HCM delay method suggested the given below equation for the purpose of calculation of vehicle delay. 𝑑 = 𝑑1 ∗ 𝑃𝐹 + 𝑑2 + 𝑑3 𝐸𝑞. (4 − 4) where: d = average overall delay per vehicle (sec/veh), d1 = uniform delay (sec/veh), d2 = incremental, or random delay (sec/veh), d3 = residual demand delay to account for over-saturation queues that may have existed before the analysis period (sec/veh), PF = adjustment factor for the effect of the quality of progression in coordinated systems. Figure (4-8) shows the calculated delay by HCM and field delay. The correlation coefficient obtained is R2=0.88 which suggests a good correlation. The regression relationship is as given below: Field delay =0.813 * HCM delay 91 for v/c ≤ 1.2 𝐸𝑞. (4 − 5) Chapter Four Data Presentation and Analysis The form of regression equation indicates that, the HCM equations provide higher estimate of delay values than the field delay. Table (4-15) lists the results of the ANOVA test, for the delay model showing that the regression model is statistically significant (p ≈ 0). Figure (4-8) HCM Delay versus Field Delay Table (4-15 ) ANOVA Test for the Regression Model. Sum of Squares Mean Square Regression 21773.123 21773.123 Residual 2802.473 311.386 Total 24575.595 92 F Sig. 69.923 .000 Chapter Four Data Presentation and Analysis 4.8 Improvements This section presents the improvement strategies applied to the traffic flow in the study area of Sulaymaniyah city, through the application of the calibrated software. SYNCHRO is used to evaluate the existing traffic operations. It is also used as a tool to determine the optimum traffic signal timings for a single intersection or for a series of coordinated or uncoordinated signalized intersections. This section divides the improvement strategies into two parts. The first part includes Signal Timing Optimization and Coordination. The improvement type is suggested and evaluated by SYNCHRO program. The second part includes the geometric improvement. 4.8.1 Signal Timing Optimization and Coordination Traffic signal control is a system for synchronizing the timing of any number of traffic signals in an area, with the aim of reducing stops and overall vehicle delay or maximizing throughput. One of the most common methods for increasing the efficiency of traffic operation is coordination of consecutive traffic signals. This is because, the higher increase in the traffic demand will result in higher increase in the need of coordination. Traffic signal coordination can reduce the traffic congestion in many areas. Substantial improvements in traffic flow and reduction in delay, fuel consumption and stops and uniform flow speed links could be achieved by coordination. Signal coordination is the process to synchronize the start of the “green light” along the major roadway, so that vehicles can travel through a group of signals with minimal or no stopping. There are three key timing parameters to optimize and are noticeable to the driver. These include the “cycle length”, 93 Chapter Four Data Presentation and Analysis intersection “offset,” or progression, and the individual traffic movement “green + yellow + red” phase (referred to as a movement “split”). SYNCHRO contains a number of optimization types. It optimizes cycle length, split times, offsets and phase sequence to minimize driver stops and delays. These types are applied to the network in the same order shown in Figure (4-9) (Husch and Albeck, 2006). Setup Intersection Timing Plan Optimize Splits and Cycle Length Optimize Network Cycle Length Optimize Offset and Phase Order Figure (4-9): Optimization Steps in SYNCHRO. The range of cycle length recommended by HCM is 60-120 seconds. Range of cycle length 100-120 seconds will be recommended to accommodate an extra 10% capacity. Long cycle lengths more than 120 seconds may have negative operation aspects including long queues, inefficient use of turning lanes, and blocking. Lower cycle lengths (30-60) may be better to reduce queues which in general increase capacity and provide smoother traffic operation (Sofia, 1998). The optimization results with the HCM recommended cycle range do not result in optimum cycle length, so the default range of SYNCHRO is used. 94 Chapter Four Data Presentation and Analysis To calculate the best cycle time and the optimal timing of the existing phase sequences of the observed intersections, SYNCHRO is used. The range of cycle times used in the optimization process is 50-200 seconds. This range is used for each observed isolated intersection. Table (4-16) provides a summary for the optimization results. This is together with the best cycle time for each intersection in the observed network. Table (4-16) shows that, the use of optimization functions for the isolated signalized intersections such as split and offset has good effect on the measure of effectiveness (MOE) for Palace intersections and has minimized the effect of (MOE) at Sara intersection. 95 Chapter Four Data Presentation and Analysis Sara palace sec sec/veh Total Fuel Consumption Total Uniform Stop LOS Total Delay Max. v/c at Intersection Cycle Length Improvement Type Intersection Name Table (4-16): Measure of Effectiveness for the Improved Intersections with the Optimum Cycle Length Produced by SYNCHRO. veh/hr % l/hr Base Condition 300 1.1 61.2 F 2417 0.55 350 Split 300 1.01 63.0 E 2453 0.56 359 Cycle length 120 1.05 45.6 D 2320 0.53 302 Network cycle length 110 1.05 42.46 D 2310 0.53 291 Phasing Sequence 110 1.05 42.46 D 2310 0.53 291 Base Condition 196 0.68 31.9 C 1106 0.57 128 Split 196 0.67 31.9 C 1102 0.57 130 Cycle length 65 0.92 29.7 C 1245 0.64 118 Network cycle length 65 0.92 29.7 C 1245 0.64 118 Phasing Sequence 65 0.92 29.7 C 1245 0.64 118 96 Chapter Four Data Presentation and Analysis The optimization results show that, although the cycle time increases to high values, the isolated intersections perform oversaturation and undersaturated conditions. So, it is necessary to improve the performance by coordination to reduce cycle length and delay. After the coordination process for signalized intersections in the network, the delay times are reduced by 6.9 % for Palace intersection and with no effect at Sara intersection. The fuel consumption is reduced from 302 l/hr to 291 l/hr, as shown in Table (4-16). Although the optimization and coordination runs improve the network measure of effectiveness, the degree of saturation does not reduce to a value less than 0.9 thus we should use another trial by improving the geometries of intersections. 4.8.2 Geometric Improvement The improvement proposals shall be as follows: 1. Re-marking of pavement and utilization of parking lanes. 2. Widening of approaches. 3. Change the intersection type. The measure of effectiveness produced for each strategy is illustrated in Table (4-17) for Signalized intersections analyzed by SYNCHRO. The first step of the geometric improvement is to apply pavement marking. From the measurement, it can be noticed that the number of lanes in many approaches of these intersections can be increased by decreasing the lane width (with keeping the width of lanes greater than a minimum value of 2.4 m according to HCM) which is actually observed to be used in field at intersection approaches. Increasing the number of lanes will increase the capacity of affected approaches, which leads to improve traffic performance measure of effectiveness for some intersections at the selected network. Therefore, in this 97 Chapter Four Data Presentation and Analysis step, consideration is given to pavement marking the approaching lanes so that re-arranging of queuing vehicles at stop line is attained. Palace Great Mosque Sara sec Base Condition Pavement Marking and utilization of parking lanes Base Condition Change intersection type (Un signalized to signalized) Pavement Marking and utilization of parking lanes sec/veh Total Fuel Consumption Total Uniform Stop LOS Total Delay Max. v/c at Intersection Cycle Length Improvement Type Intersection Name Table (4-17): Measure of Effectiveness for the Improved Intersections Produced by SYNCHRO (Geometric Improvement( . veh/hr % l/hr 65 0.92 29.7 C 1245 0.64 118 60 0.79 23.7 C 1246 0.64 99 --- 2.01 112.9 F --- --- --- 60 0.5 11.8 B 433 0.58 27 60 0.3 9.6 A 392 0.52 26 Base Condition 110 1.05 42.46 D 2310 0.53 291 Pavement Widening 60 14.0 B 2155 0.5 197 0.8 98 Chapter Four Data Presentation and Analysis The pavement marking is applied to Sara intersection to increase number of lanes with the same approach width. This method reduces the maximum v/c value from 0.92 to 0.79 and reduces delay from 29.7 seconds to 23.7 seconds. The fuel consumption reduces from 118 l/hr to 99 l/hr. The improvement is shown in Figure (4-10). The Great Mosque is unsignalized intersection (Two Way Stop Controlled). This intersection suffers from congestion at Court approach, so a suggestion is made by changing the intersection type to signalized intersection. Results obtained by SYNCHRO and illustrated in Table (4-17) show an increase in all measures of effectiveness and reduction in v/c to a maximum value of 0.5 for Court approach . This reduces delay time from 112.9 seconds to 11.8 seconds. Another trial is made by pavement marking and utilization of parking lanes, as shown in Figure (4-11). This reduces delay time from 11.8 seconds to 9.6 seconds and reduces the maximum v/c value to 0.3. Another trial is made for Salim approach and Baban approach at Palace intersection by widening due to the presence of wide sidewalk at this intersection section taking into consideration pedestrian volume, as shown in Figure (4-12). The obtained values of v/c are less than 0.8. Although the optimization and coordination is applied to Palace Intersection, but the degree of saturation does not reduced to value less than 0.9, and needs further modification to achieve the objected delay. Another trial is made by widening approaches due to presence of wide sidewalk at this section taking into consideration pedestrian volume. The pavement widening at Salim approach will increase the capacity. This reduces the maximum v/c value to 0.8, While, The pavement widening at Baban approach reduces the maximum v/c value to 0.77. Pavement widening at Palace intersection reduces the delay time from 42.46 seconds to 14 seconds and the fuel consumption from 291 l/hr to 197 99 Chapter Four Data Presentation and Analysis l/hr. Figure (4-12) shows the Palace intersection before and after geometric improvement. Figure (4-10): Improvement Proposal for Sara Intersection. Figure (4-11): Improvement Proposal for Great Mosque Intersection. 100 Chapter Four Data Presentation and Analysis Figure (4-12): Improvement Proposal for Palace Intersection. 4.9 Forecasted Network The growth factor of 3.4 % (i.e. traffic forecast factor TFF) is introduced for the application of forecasting traffic volume on the improvement proposals for intersections. The growth rate together with the analysis period is applied with SYNCHRO for the application of forecasting traffic volume on the improvement proposals. Table (4-18) shows the measure of effectiveness obtained from SYNCHRO optimization runs after the last improvement strategy for the target year 2017. 101 Chapter Four Data Presentation and Analysis Table (4-18):Measure of Effectiveness Produced by SYNCHRO after Last Improvement Strategy for the Base Year 2010, and Target Year 2017. Intersection Name Base Year 2012 Average Delay (sec/veh) Target Year 2017 LOS Average Delay (sec/veh) LOS Palace 14.0 B 17.5 B Sara 23.7 C 24.7 C Great mosque 9 A 10.4 B The LOS shown in Table (4-18) shows that the increase in traffic volume for the target year 2017 will not increase (B) or (C) which is acceptable. 102 delay so much and LOS remains CONCLUSIONS AND RECOMMENDATIONS Chapter Five Conclusions and Recommendations Chapter Five Conclusions and Recommendations 5.1 Conclusions Within the limits of the traffic and geometric features of the study area, the main conclusions that can be drawn from this study are summarized, as follows: 1. Delay time for signalized intersections produced by SYNCHRO has a no significant difference with the delay time measured at field with 95% confidence level, SYNCHRO has a good representation of field delay at low to medium values, but it is overestimated at high delay range. 2. Regression models, developed to estimate delay time and queue length, show good correlation with field values. Delay model can used to estimate delay time at any intersections knowing signal timing, traffic volume, ratio of total width of exit roadway of intersection for traffic departing straight forward to total width of lane groups departing to the same exit roadway, in the same phase at stop line (We/Ws). Queue model can be used to estimate queue length at any intersections knowing signal timing, traffic volume. These models may not be suitable for extreme range of input variable. 3. Delay estimate by HCM2000 model gives higher estimation of field delay. From the analysis, it is recommended that, the theoretical delay (due to random arrival and oversaturated condition) in HCM 2000 delay model can be reduced by (0.813) to better reflect field delay values. 4. All signalized intersections suffer from long cycle length that causes high delay time values. 5. Optimization and coordination process could improve the performance of the network. 103 Chapter Five Conclusions and Recommendations 6. Pavement remarking and lane designation slightly improve the intersections performance. 7. The delay increases when the width of exit roadway is less than the width at stop line, as in the case of bottleneck. 5.2 Recommendations Based on the study results, the following recommendations may be put forward: 1. Fixing permanent video recording camera on streets and intersections to measure traffic volume and speed of vehicle, which helps in the data collection process and analysis for future studies and for continuous monitoring of causes of delay and congestion. 2. Converting great mosque from unsignalized intersection (TWSC) to signalized intersection to improve intersection performance. 3. Locating traffic signals in all intersections and retiming the existing phase sequence, depending on existing traffic volumes on those intersections. 4. Marking and lightening all streets and intersections. 5.3 Future Studies 1. Few studies on traffic and planning have been conducted in Sulaymaniyah city; therefore, it is necessary to encourage more studies. This can be achieved by supporting researchers in terms of modern equipment supply that requires less manpower and provide data on different parameters of traffic engineering. 2. The possibility of constructing pedestrian overpass at streets and signalized intersections to reduce conflict with pedestrian, to provide safety, and decrease delay time. 104 Chapter Five Conclusions and Recommendations 3. Network optimization and coordination study in Sulaymaniyah for all close intersections in the city can be implemented. 105 Reference References AASHTO, American Association of State Highway and Transportation Officials, (2004): "A Policy on Geometric Design of Highways and Streets", 5th Edition, Washington D.C. Abd Ul-Ghani, M.M.(2004), "Evaluation and Improvement of Traffic Flow Patterns for a Selected CBD Area in Baghdad City", M.Sc. thesis, University of Technology, Baghdad, Iraq. Ahmed N.G., (2002): "The Development of Models for the prediction of Thermal Cracking in Flexible Pavements", Ph. D. Thesis, Civil Engineering Department, College of Engineering, University of Baghdad, Iraq. Akçelik R., (2009): "SIDRA Intersection User Guide", Akçelik& Associates Pty Ltd. Al-Azzawi, Z.A. (2003): "Evaluation of Signalized Intersections Capacity in Baghdad to Attain the Traffic Flow Performance", M.Sc. thesis, Al- Mustansirya University, Baghdad, Iraq. 106 Reference Al-Dulaimi M. R.,(2006): " Modeling of Pedestrian-Vehicle Conflict on Arterial Street Using Simulation Approach", M.Sc. thesis, AlMustansiriyah University, Baghdad, Iraq. Al-Hemdany H.M.(2003), "Network Traffic Performance within Adhamiyasector in Baghdad City", M.Sc. thesis, Al-Mustansirya University, Baghdad, Iraq. Al-Hummadawee B. H.(2002), "Evaluation of Performance for Selected Signalized Intersections in Baghdad City", M.Sc. thesis, University of Technology, Baghdad, Iraq. Al-Jaman, S.M. (2007):"Assessment and Calibration of SYNCHRO for Riyadh Traffic Conditions", M.Sc. thesis, King Saud University, Saudi Arabia. Al-Neami, A. H., "EVENT – A Computer Program for Abstraction of Traffic Data". Engineering Journal, University of Tikrit, vol.3, No.2, 2000 (Cited By Reference: Razzaq, 2005). Al-Zubaidi Y. K., (2006): "Determination of Passenger Car Unit Factors for Different Types of Vehicles in Baghdad City", M.Sc. Thesis, Al-Mustansirya University, Baghdad, Iraq. 107 Reference Burrow, I. (1989), "A Note on Traffic Delay Formulas", ITE Journal, Vol. 59, No. 10, pp: 29-32,UK. Catling, I. (1977), "A Time-Dependent Approach to Junction Delays", Traffic Engineering and Control, Vol. 18, No. 11, pp: 520523, 526. Fambro, D. B. and Rouphail, N. M. (1997), "Generalized Delay Model for Signalized Intersections and Arterial Streets", Transportation Research Record 1572, TRB, National Research Council, Washington, D. C., pp: 112-121, USA. FHWA Federal Highway Administration, (2009): "Manual on Uniform Traffic Control Devices for Streets and Highways". Garber,N.J., and Hoel, L.A.,(2009): "Traffic and Highway and Highway Engineering", Thomas Learning, Fourth Edition. Garber,N.J., and Hoel, L.A.,(2010): "Traffic Engineering", Thomas Learning, Fourth Edition. "Highway Capacity Manual", 4th Edition, Transportation Research Board, National Research Council, Washington D.C, 2000, USA. Husch, D., Albeck, J., (2006): "Traffic Signal-User Guide, Synchro Studio7". 108 Reference Keller, G., and Warrack, B. (2000): "Statistics for Management and Economies", 4th Edition, USA. (Cited By Reference: Ahmed N.G., 2002). Kennedy , J .B., and Neville, A.M.(1986): “Basic Statistical Methods for Engineers and Scientists" , 2nd Edition , Thomas Y., Crowell Company , In. U.S. (Cited By Reference: Ahmed N.G., 2002) Khalaf T. M., (2003): "Modified Predictive Models for Vehicle Delay at Selected Signalized Intersection in Baghdad City", Ph.D. thesis, Baghdad University, Baghdad, Iraq. Lee, J. and Chen, R. L. (1986), "Entering Headway at Signalized Intersections in Small Metropolitan Area", Transportation Research Record 1091, TRB, National Research Council, Washington, D.C., pp: 117-126, USA. Lin, Feng-Bor (1989), "Application of 1985 Highway Capacity Manual for Estimating Delay at Signalized Intersections ", Transportation Research Record 1225, TRB, National Research Council, Washington, D.C., pp: 18 -23, USA. McShane, W. R., Roess, P. R. (1990), "Traffic Engineering", Prentice Hall, Eaglewood Cliffs, New Jersey, USA. 109 Reference Md. Hadiuzzaman, (2008): "Development of saturation flow and delay models for signalized intersection in Dhaka city", M.Sc. thesis, Bangladesh University, Dhaka, Bangladesh. Miller, A.J. (1968), "Australian Road Capacity Guide", Australian Road Research Board (ARRB), Bulletin No. 4, Australia. Pickrell, S. and L. Neumann, (2000): "Linking Performance Measures with Decision Making", Presented at the 79th Annual Meeting of the Transportation Research Board, Washington, D.C, USA. Pignataro, L.T., (1973): "Traffic Engineering", Pentice –Hall Inc., Third Edition, USA. Polus A., Vlahos E., (2005): "Evaluation of Roundabouts versus Signalized and Unsignalized Intersections in Delaware", Department of Civil and Environmental Engineering, University of Delaware, Delaware, USA. R. Troutbeck, NCHRP Web Document 5(1997): "Capacity and Level of Service at Unsignalized Intersections", Final Report, Vol. 1—Two-Way Stop-Controlled Intersections. Razzaq N. A., (2005):"A Study of Pedestrian Behavior at Random Crossings in Baghdad City", M.Sc. thesis, Baghdad University, Baghdad, Iraq. 110 Reference Reilly, W. R. and Gardner,(1977): "Technique for Measurement of Delay at Intersections" ,Transportation Research Record 644, TRB, National Research Council, C. C. ,Washington, D.C., pp: 1-7, USA. Rogers, M., (2008): "Highway Engineering", Second Edition, Blackwell Publishing, Ireland. Sargeant S., Christie J., (2002): "Evaluation of a Series of Roundabouts Implemented on South Golden Road", Institute of Transportation Engineers, Annual Meeting, California,USA. Shamsul Hoqueand and Asif Imran, (2007): " Modification of Webster’s delay formula under non-lane based heterogeneous road traffic condition ", Journal of Civil Engineering (IEB). SIERPIŃSKI, (2007): "Proposition of Delay Model for Signalized Intersection with Queueing Theory Analytical Models Usage ", Transport Problems, Silesian University of Technology. Sofia, G.G. (1998):"Development of Traffic Flow Models at Networks in Mosul City", Ph.D. thesis, Baghdad University. SORB, State Organization of Roads and Bridges, (2005): "Highway Design Manual", Republic of Iraq, Ministry of Housing and Construction, Baghdad, Iraq. Teply, S. (1989), "Accuracy of Delay Surveys at Signalized Intersections", Transportation Research Record 1225, TRB, National 111 Reference Research Council, Washington, D.C., pp: 24-32, USA. (Cited by Md. Hadiuzzaman, 2008). "Traffic Directorate of Sulaymaniyah Governorate" ,( 2013) Department of Statistics, Vehicle Registering File. Trafficware, Ltd., (2011): "Traffic Signal-User Guide, Synchro Studio 8". TRB, Transportation Research Board, (2005):" Highway Capacity Manual", Washington D.C, USA. Turley C., (2007): "Calibration Procedure for a Microscopic Traffic Simulation Model", M.Sc. thesis, Brigham Young University, USA. Warid, A.K.(2004), "Evaluation and Improvement of Traffic Flow Patterns at Al-Karrada CBD Area in Baghdad City", M.Sc. thesis, university of Technology, Baghdad, Iraq. Webster, F.V. and Charles worth, G., (1958) "Some Factors Affecting the Capacity of Intersections Controlled by Traffic Signals ", 4th International Study Week in Traffic Engineering, Copenhagen, Denmark. (Cited by Md. Hadiuzzaman, 2008). Webster, F.V. and Cobbe, B.M., (1966) "Traffic Signals". Road Research Technical Paper No. 56, HMSO, London. (Cited by Md. Hadiuzzaman, 2008). 112 Reference Youn-Soo Kang,(2000): " Delay, Stop and Queue Estimation for Uniform and Random Traffic Arrivals at Fixed-Time Signalized Intersections ", Ph.D. thesis, Virginia University, USA. 113 Appendix (A) Appendix (A) Sample of Spot Speed Data Table (A-1) Speed Data for Mawlawi Street Vehicle No. Time (sec) Length (m) Spot Speed (kph) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 2.95 3.1 2.8 2.85 2.7 3.05 3.4 3 3.25 3.2 2.85 3.05 3.05 2.7 3.5 3.2 2.85 3.1 3.1 3.8 3.55 2.75 3.95 3.65 3.1 2.85 3 3.2 3.95 3.45 3.6 3.65 3.2 3.3 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 30.51 29.03 32.14 31.58 33.33 29.51 26.47 30.00 27.69 28.13 31.58 29.51 29.51 33.33 25.71 28.13 31.58 29.03 29.03 23.68 25.35 32.73 22.78 24.66 29.03 31.58 30.00 28.13 22.78 26.09 25.00 24.66 28.13 27.27 A-1 Appendix (A) 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 3.25 2.8 3.75 3.3 3.95 3.7 2.8 4.05 3.6 3.45 3 3 3.1 3.35 2.75 3.6 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 Average = 28.09 km/hr A-2 27.69 32.14 24.00 27.27 22.78 24.32 32.14 22.22 25.00 26.09 30.00 30.00 29.03 26.87 32.73 25.00 Appendix (B) Appendix (B) W ork Sheet for Field Measurement of Delay Intersection Control Delay W orksheet Analyst : Saad Mohsen Khalil Intersection: Palace / Baban St. Area Type : CBD Agency or Company : ( )ـــ Analysis Year : 2012 Date Performed : 10/1/2012 Analysis Time Period : 3:00 to 3:17 Input Initial Parameters Number of Lanes , L : 1 Total vehicle arriving, Vtot = 189 Free – flow speed, FFS ( km / hr ) : 38 km / hr Stopped vehicle count , V stop = 93 Survey count interval ( s ) : 15 sec Input Field Data Clock Time 3.00 Cycle Number Count interval 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 0 4 10 10 10 16 18 18 11 5 0 0 2 3 8 9 13 21 12 22 20 12 9 6 3 3 3 3 6 6 8 11 12 16 17 4 0 2 2 5 10 17 14 19 19 6 1 7 24 34 47 63 70 71 68 17 18 19 2 0 0 0 0 21 19 12 11 5 0 0 0 0 17 13 13 9 3 0 0 0 0 0 59 45 34 26 16 5 0 0 0 0 3.17 Total Total vehicles in queue, ƩViq =585 Computations No. of cycle surveyed, Nc =4 Time- in –queue per vehicle, dvq=(I*ƩViq/VTot)*0.9=41.78 sec Fraction of vehicle stopping, FVS=Vstop/Vtotal= 0.492 Accel/Decel delay,dad =FVS*CF = - 0.492 Accel/Decel correction factor, CF =-1 No. of vehicles stopping per each cycle = Vstop/NcXL = 23.25 Control delay = dad + dvq =41.288 sec B-1 Appendix (C) A- Appendix (C) Sample of EVENT Program Output This file has been opened with Microsoft Excel for counting traffic volume for each turning movement, and its composition for Intersection. Palace Intersection ( Left Turn of Salim Street ) Tuesday 3:00 10/1/2012 TO 3:15 pm Volume = 206 veh 154.6703343d 157.3626403d 161.0989073d 164.3955993d 167.4725343d 169.6153873d 171.4285743d 174.4505463d 176.9230803d 179.2307743d 181.9780273d 184.9450533d 188.2417603d 190.4395603d 193.2967073d 195.8241733d 198.1868133d 200.1648413d 201.8681343d 204.7252813d 207.9670263d 210.9340673d 214.0109863d 217.2527473d 219.0109863d 221.0439613d 224.5054933d 227.1428533d 231.1538393d 234.2857213d 238.1868133d 239.7252813d 242.3076933d 244.5604403d 246.9230803d C-1 Appendix (C) 249.0659333d 251.3186803d 255.4395603d 258.1867983d 259.8901063d 263.7362673d 265.9340523d 269.2857063d 272.6373603d 275.2747193d 278.3516543d 280.7142943d 283.5164793d 286.3736273d 289.3406683d 294.3955993d 298.6813053d 302.2527473d 439.7252813d 443.3516543d 445.7692263d 447.6923223d 450.1648253d 452.3626403d 455.6593323d 459.0109863d 462.3626403d 464.7802123d 468.2417603d 470.6044013d 473.6263733d 477.6373603d 479.7802123d 482.9120793d 485.9890143d 489.5054933d 492.5824283d 494.9450683d 497.7472533d 500.3846133d 502.7472533d 506.0439453d 509.6153873d 512.9121093d 515.1648563d 517.6923223d 520.6593633d 523.9011233d 527.5823973d C-2 Appendix (C) 532.9121093d 544.3956303d 547.8571173d 550.4945073d 552.9670413d 556.2088013d 558.6264043d 561.4285893d 563.5164793d 565.6593633d 568.2417603d 570.4395753d 573.2417603d 575.7692263d 577.8021853d 579.7802123d 583.0219733d 584.3956303d 586.7033083d 589.2857063d 592.0878913d 596.1538703d 598.6264043d 601.4835213d 603.6813353d 607.3076783d 609.7802123d 611.6483763d 614.2857063d 618.0769043d 620.2197883d 622.0329593d 624.2307743d 654.5604253d 727.9121093d 731.4285893d 733.6813353d 737.0878913d 740.0000003d 742.5274663d 744.3956303d 750.2197883d 752.0329593d 754.4505623d 756.6483763d 759.1758423d 761.5384523d 763.8461303d 765.8791503d C-3 Appendix (C) 769.2307743d 772.6923223d 775.8241583d 778.2417603d 781.2088013d 784.5604253d 787.2527473d 790.1648563d 794.0659183d 796.3735963d 797.1978153d 798.6813353d 801.5384523d 803.7911993d 805.9890143d 809.2857063d 811.7582403d 814.5604253d 816.9780273d 820.6593633d 822.9670413d 826.0439453d 830.1648563d 833.1868293d 835.8241583d 839.3956303d 842.3626103d 844.5604253d 847.0329593d 849.6153563d 852.3076783d 855.1648563d 858.0769043d 860.0000003d 862.0329593d 864.6153563d 866.7033083d 870.1648563d 871.8681033d 873.9560553d 876.8131713d 878.9560553d 881.1538703d 884.2857063d 886.9780273d 892.6923223d 893.5164793d 895.8241583d 898.0219733d C-4 Appendix (C) 901.4835213d 904.4505623d 906.7582403d 908.4615483d 910.4395753d 912.1978153d 914.2857063d 915.9890143d 918.4066163d 921.9230963d 924.6153563d 926.3735963d 928.1318973d 930.2747193d 932.8021853d 934.8351443d 936.7582403d 941.0439453d 942.8571173d 944.7802123d 948.5714113d 951.4835213d 1014.0109863d 1028.0219733d C-5 Appendix (C) B- Procedure Shows Event Program Operation. Event program Press any key to continue Press ESC button to start Operate video tape on TV Press Fi* buttons, when the front tire of the vehicles touches the stop line At end, press ESC button to stop the program A new window asks for the name of the file Write a name and press enter to close the window C-6 Appendix (D) Appendix (D) SYNCHRO Output SYNCHRO software provides different output results in graphical and in tabular form. The main output used in present study is presented. D-1 Appendix (D) D-2 Appendix (D) D-3 Appendix (D) D-4 Appendix (D) D-5 Appendix (D) D-6 Appendix (D) D-7 Appendix (D) D-8 Appendix (D) D-9 Appendix (D) Time Space Diagram for Palace Intersection D-10 Appendix (E) Appendix (E) Traffic and Geometric Condition of Signalized intersections Table (E-1) Traffic and Geometric Condition for Delay Model Cycle(sec) v/c g/C Field delay (sec) 270 1.15 0.390 67.263 1.000 3.50 657 300 1.06 0.610 69.108 1.000 3.20 805 266 1.02 0.565 51.070 1.000 3.20 776 300 0.95 0.370 62.482 1.000 3.50 691 266 0.95 0.420 50.482 1.000 3.50 541 270 0.79 0.585 47.622 1.000 3.20 606 97 0.73 0.480 40.870 0.740 4.20 495 225 0.69 0.440 50.383 0.458 4.80 222 97 0.58 0.470 36.650 0.458 4.80 179 196 0.56 0.486 47.340 0.458 4.20 379 97 0.55 0.470 38.559 0.458 2.85 368 196 0.54 0.486 49.852 0.458 2.85 360 225 0.53 0.540 46.750 0.458 4.20 357 196 0.50 0.486 43.390 0.458 4.80 161 225 0.50 0.440 49.370 0.458 2.85 337 E-1 We/Ws Width(m) Volume (veh/hr) Appendix (E) Table (E-2) Traffic and Geometric Condition for Queue of Left Turn Movements. Width (m) Cycle(sec) v/c g/C Field Queue Length(veh) 3.2 300 1.06 0.61 31.333 3.2 270 0.79 0.585 15.166 3.2 266 1.022 0.565 20.66 3.5 300 1.172 0.37 24.333 3.5 270 1.154 0.39 23.75 3.5 266 0.95 0.42 23.666 4.2 97 0.73 0.48 11.666 4.2 196 0.56 0.486 9.665 4.2 225 0.526 0.54 13.666 4.8 97 0.577 0.47 2.625 4.8 196 0.5 0.486 6.166 4.8 225 0.691 0.44 16.666 2.85 97 0.547 0.47 6.5 2.85 196 0.535 0.486 12.666 2.85 225 0.5 0.44 11.25 E-2 Appendix (E) Table (E-3) Traffic and Geometric Condition for Queue of Through Turn Movements. g/C v/c Field Queue Length(veh) Cycle (sec) Width (m) 0.59 1.06 28.33 300 3.2 0.61 0.80 12.75 270 3.2 0.59 0.87 19.25 266 3.2 0.39 1.14 26.67 300 3.5 0.39 1.30 21.25 270 3.5 0.39 0.90 20.67 266 3.5 0.49 0.47 6.00 97 3.2 0.50 0.51 4.40 196 3.2 0.49 0.53 15.67 225 3.2 0.49 0.65 1.88 97 4.8 0.50 0.15 3.00 196 4.8 E-3 المستخلص ( )% 10ليعكس بشكل افضل للقيم الحقلية .الموديل المقترح للتأخير يبين عالقة جيدة مع القيم الحقلية ضمن حدود هذه الدراسة. تم اقتراح تحسينات لجريان المرور ضمن منطقة الدراسة في مدينة السليمانية من خالل تطبيق برنامج ( )SYNCHRO/ SimTrafficالمعدل .هذه التحسينات تنقسم الى جزئين ،الجزء األول هو إعادة ضبط التوقيتات والتنسيق بين التقاطعات اما الجزء الثاني يتضمن تحسينات هندسية .هذه التحسينات لها تأثير جيد جدا على مقاييس الفعالية للشبكة ضمن منطقة الدراسة. .واخيرا تم إيجاد مقاييس الفعالية بواسطة برنامج ( )SYNCHRO/ SimTrafficبعد أخر التحسينات وللسنة المستقبلية .7102 المستخلص المستخلص تعاني مدينة السليمانية من االزدحام المروري ،تأخير المركبات على معظم الطرق والتقاطعات وخصوصا خالل ساعات الذروة وهذا يسبب الكثير من المشاكل المروية ويؤثر سلبا ً على اقتصاد البلد. زيادة عدد المركبات ومحدودية بناء طرق جديدة ستزيد من مشاكل الشبكة .عملية التنبؤ الصحيح لتخمين التأخير طوابير السيارات تصبح مهمه لالستخدام األمثل لشبكة الطرق الحالية بوساطة توفير أفضل عميلة تشغيل وسيطرة مرورية. الهدف من هذه الدراسة الحالية تقييم أداء حركة المرور للشبكة المختارة من خالل تقييم أداء التقاطعات وتطوير نماذج للتأخير وطوابير السيارات في التقاطعات في حركة المرور بالنسبة للظروف الحالية في مدينة السليمانية ويعقب ذلك اقتراحات لبعض مقترحات التحسين والتي تتضمن تغير خطة التوقيت لإلشارة الضوئية وتحسين هندسة التقاطع وتغير نوع التقاطع. لتحقيق الهدف المذكور اعاله ،تم جمع بيانات حركة المرور للشبكة الواقعة ضمن المنطقة التجارية المركزية في مدينة السليمانية بوساطة استخدام كاميرا فيديو لجمع البيانات الحقلية ألنها توفر سجل دائم للبيانات مع الحد األدنى للقوى العاملة ,هذه البيانات تستخرج من تحليالت الفيديو بواسطة برنامج ( )EVENTوتعالج بعد ذلك بواسطة أوراق ( )EXCELمعدة لهذا الغرض .تم استخراج السرعة اللحظية لكل أجزاء الشبكة بطريقة عالمات التبليط. تم استخدام برنامج ( )SYNCHROلتقييم وتحليل التقاطعات واقتراح افضل خطة توقيت لإلشارة الضوئية والتنسيق بين التقاطعات وتم تقييم االقتراحات لسنة الهدف 7102باستخدام البرنامج المذكور. النماذج اإلحصائية التي طورت لتخمين التأخير و طوابير السيارات بينت عالقة جيدة مع القيم الحقلية. نسبة العرض الكلي لمخرج التقاطع والمستخدم للمرور الى العرض الكلي للخطوط التي يتجه فيها حركة المرور للمخرج نفسه وفي نفس الطور) (We/Wsأظهرت تأثير على التأخير .وهذه النسبة يكون لها تأثير سلبي عندما تقل عن . 0 التأخير المخمن اعتمادا على ) (HCM 2000تم مقارنته مع التأخير الحقلي .بناءاً على النتائج التي تم الحصول عليها من التحليل فانه من األفضل تعديل نموذج ()HCM 2000 ليمثل التأخير الحقلي وقد لوحظ ان نموذج ) (HCM 2000يستمر بحساب قيمة اعلى للتأخير عندما تكون درجة التشبع اكبر من 0باالعتماد على نتائج التحليل يجب تخفيض التأخير النظري لل)(HCM 2000 بنسبة جمهورية العراق وزارة التعليم العالي و البحث العلمي ألجامعة المستنصرية كلية الهندسة قسم هندسة الطرق والنقل نمذجة طوابير السيارات والتأخير للمنطقة التجارية المركزية في مدينة السليمانية رسالة مقدمة إلى قسم هندسة الطرق و النقل /كلية الهندسة /الجامعة المستنصرية كجزء من متطلبات نيل درجة الماجستير في علوم هندسة الطرق و النقل من قبل: سعد محسن خليل )بكالوريوس /قسم هندسة الطرق و النقل ) 0202، بأشراف أ.م.د غاندي غانم صوفيا م.د علي جبار كاظم أيلول 0202م ذي القعدة 0323هـ