Click for next page ( 170


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 169
APPENDIX F DESCRIPTION OF DATA SETS USED FOR MODEL EVALUATION IN NCHRP PROJECT 3-48 The analytical models developed under NCHRP Project 348 are intended to predict signal timing and estimate delay at a signalized intersection with traffic-actuated operation. To verify and evaluate alternative models, several hypothetical data sets and one field data set were used In this study. For hypothetical data sets, the predicted signal timing and delay were compared with results from TRAF- NETSIM simulation. This appendix describes the intersection geometry, phase sequence, tragic volumes and the actuated control parameters for all data sets used for the mode! evaluation. The details and results of the evaluations are presented in separate sections ofthis report. lIYPOTEIETICAL DATA SETS To provide a general comparison oftraffic-actuated signal timing and delay prediction between HCM Chapter 9 Appendix II method and proposed analytical model, several hypothetical data sets based on four HCM Chanter 9 sample problems (SPI, SP3, SP4 and SP51 and seven hv~othetical examples fit ~ TT~^ AT'-~ TV'- A TT'-r TT'-~ 1 It__ ~ <~1, Am, Am, Am, Am, ~o and am/' were used. Chinese data sets were selected from a group of sample problem data sets prepared bv the Signalized Intersections Subcommittee ofthe Hi~hw.av in- - ~ ~ r--r~-~~ I ~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~ ~~ ~ ~ ~ ~ J Capacity Committee. Some of the traffic and operational data were modified to increase the range of conditions included In this analysis. NETSIM was used as the evaluation too! for both signal timing and delay models. The WHICH data base manager program was used for the data input process. The ACT348 program, developed under NCEIRP Project 348 to implement the proposed models and produce the signal timing and delay, can be executed directly Dom WHICH. The equivalent Input file for NETSIM can also be created auto- maticaDy Tom WHICH. For arty input file, NETSIM defaults were adopted when specific data were not available. Base Intersection Layouts In order to intensively evaluate the proposed analytical models, possible practical intersection layouts have been considered. There are a total often base intersection layouts (SPI, SP3, SP4, =l, ~2, ~3, HE4, =5, HE6 and HE7) used In this study which are presented as follows. In HCM sample calculations, the geometry of HCM sample 5 (SP5) is similar to HCM sample ~ (SPl), so only the intersection layout of S ~ is shown in Figure If-. This is a one-lane street for both north-southbound approaches and a two-lane street for the east-westbound approaches. The intersection for HCM sample 3 shown in Figure F-2 is a major CBD junction of two arterial streets. Both facilities have ~ four lanes, with exclusive lePr-turn lanes provided at the intersection on all four approaches. Figure F-3 shows the geometry of HCM sample 4. This is a six-lane street versus a four-lane street. Each approach has an exclusive ledc-turn lane. The hypothetical intersections of ~l, HE2, HE3, HE4, YES, HE6 and ~7 are shown in Figures F-4 through F-IO, respectively. Appendix F: Page 1

OCR for page 169
~ -1t1~' _ coo _ I, --.. ~ I . . . terra ~ ~ Figure F-~. Intersection layout for SP! I, .. . I, -,, 1 LE_I , . . . ~ - 1 '~ ~ 1 -: 1 ~ 1 1 ,..,,, 1 ~ ..... Figure F-3. Intersection layout for SP4 Figure F-5. Intersection layout for HE2 Appendix F: Page 2 Figure F-2. Intersection layout for SP3 ALLY Figure F-4. Intersection layout for DIE! - GEOMh=Y r`~: ~o ~ia I ma_ _GEOMh-~-xY 1 --- I91 411R 1 ~ Figure F-6. Intersection layout for EE3

OCR for page 169
K~C ~GEOMETRY ia 11 ' 11 ~'~ 11 1 B'-....-i,.,l, Elm. 1 2T1] l l -1 ~ 1 iC_ ~ ox -sent 1~ I_ craws Figure F-7. Intersection layout for HE4 GEOlilh-1 BY K._ ~ GEOMETRY 1~ ~! I~L1 1=~t it ~ 1~ IL Figure F-~. Intersection layout for HE5 Figure F-9. Intersection layout for HE6 Phase Sequence. Traffic Volume and Actuated Parameter Range GEOUE=Y Figure F-IO. Intersection layout for :HE7 Besides the intersection layout, phase sequence, traffic volume and actuated parameter range wall influence the signal timing and vehicle delay. The hypothetical data sets cover eight different cases of phase sequences. The following figures show the cases in the north-south direction: Figure F-l I: Case I: phase sequence for simple permitted turns Figure F-12: Case 2: phase sequence for leading green Figure F-] 3: Case 3: phase sequence for lagging green Figure F-14: Case 4: phase sequence for leading and lagging green Figure F-15: Case 5: phase sequence for left turn (LT) phase with leading green Figure F-16: Case 6: phase sequence for leading dual left turns Figure F-17: Case 7: phase sequence for lagging dual leD turns Figure F-~: Case S: phase sequence for leading and lagging with dual left turns. Case 4 and Case ~ are interchangeable. For example, when northbound led turn volumes are heavy and through traffic volumes are light, Case 4 may become Case S. Appendix F: Page 3

OCR for page 169
4''^ +',>1 x 4 x 1E~ 7 8 Figure F-11. Simple permitted turns ~k'^~\,'~ !'~ 1 .. ~ 3 4 ~ 1 ~ l X g Figure F-13. Lagging green >; Nt ~''^ ~- I, l . 1 Em ft 14'`+',>1 3 4 :1 ~ 7 ~1 ~ 8 Figure F-12. Leading green Ny ~ 4''^ +\,' 4; 3 ~ [I 7 8 Figure F-14. Leading and lagging green ~ >; 14''''','1 \ 11 ~ 7 8 Figure F-15. LT phase with leading green Appendix F: Page 4 \ ~ ~ 7 8 Figure F-16. Leading dual left turns

OCR for page 169
k~^ +~1 EN 1 ~ 7 8 Figure F-17. Lagging dual left turns ~ >t ~ >; I 4; ~ , 1 3 4 ! ~ 7 8 Figure F-18. Leading and Lagging with dual left turns Note that, for the shared lane scenarios, only permitted left turns were used. Compound left turn protection was intensively evaluated at approaches with an exclusive left turn lane. The traffic volumes and actuated parameter range used in the hypothetical data sets are shown as follows: Traffic volumes per lane: Minimum green time: - maximum green time: Detector length: Allowable gap: Ideal saturation flow rate: 50vph-350vph 50 vph - 800 vph 5 see- 15 see 10 see - 20 see 10 see- 30 see 30 see - 80 see 25 tic - 30 fit 2 see - 3.5 see 1800 vphgp! - 1900 vphgp} (left turn) (through plus right turn) (left tum) (through plus right turn) (led turn) (through plus right turn) (placed at the stopline) The summary of the characteristics of the sample data sets is shown in Table F-~. There are a total of 27 data sets, some of which have subordinate data sets. The file name with an extension ".WC~' indicates a WITCH input file. These two worksheets come Dom the output of the ACT3-48 pro gram. For the data set with subordinate data sets, only the first subordinate data set is shown because the others are similar to the first one. The phase sequences are determined by the control specifica- tion in WHICH. They fall into the category of eight different phase sequences as mentioned before. The specification and notation of the data set is addressed in the comment column in Table Fat. Data sets from I-21 were used in this report for evaluating the proposed signal timing prediction model. Data sets Tom ~ to 19 and 22 to 27 were used for comparing candidate delay models. Appendix F: Page 5

OCR for page 169
Table F-~. Summary of the characteristics of the sample data sets No.of ~File Nme ~ Type of ~ Case ofPhase ~ Comments Data Set ~(.wc l) | In. Layout ~Sequence | 1 | TEMP 1-8) | ~1 | Case ~| dentical approach Only through traffic Volume range Dom 100 vph to 800 vph ~ Inc.=100 vph) 2 TEST(1-~) SP4 Case 6 Exclusive LT lane NB and SB identical . EB and Ah7B identical Volume range from 200 vph to 1250 vph (Lnc.=150 vph) for NB and SB Volume range Lom 400 vph to 1 100 vph (Lnc.=100 vph) ~ for EB end WB 3 ~ ~] 1 ~HE1 ~Casel 4 | HEZ 1 ~HE2 ~Case 1 ~ 5 | HEZ 2 ~HE2 ~Case 1 ~ Only one lane for EB _ 6 ~ HE3 1 ~HE3 ~Case 7 ; HE4 1 ~HE4 ~Case 1- ~ HE4-2 HE4 Case 1 Identical volume for each l l l I approach 9 ~ HE; 1 ~HE5 ~Case 10 ~ SP} 1 ~SP1 ~Case 1 ~ Il ~ SP3 1 ~ SP3 T Case2 T 12 SP3-2 SP3 Case 3 13 ~ SP3 3 ~SP3 ~Case4 ~ 14 SP3-4 SP3 Cases ~ & 5 Case 1 for EB and WB . Case 5 for NB and SB 5 1 SP3 5 ~SP3 ~Cases ~ 6 SP3-6 SP3 Case 7 Appendix F: Page 6

OCR for page 169
Table F-~. Summary of the characteristics of the sample data sets (continued) 1 7 SP4-1 SP4 Case 5 . Exclusive LT lane for each ~ approach SP4 Cases 4 & 6 Case 4 for NB and SB Case 6 for EB and WB SP4 Case 4 HE6 Case 5 HE3 Cases 1 & 5 Case 1 for EB and WB Case 5 for NB and SB HE6 Case 6 NB and SB identical EB and WB identical Volume range Dom 225 vph to 825 vph ~ Inc.=75 vph) for NB and SB Volumes are fixed as a constant for EB and WB SP3 Case 2 Slight volume change for EB and WB . Fixed volumes for both NB and SB . SP3 Case 6 HE7 Case 1 NB and SB identical EB and WB identical Volume range from 90 vph to 810 vph ~ Inc.=90 vph) for NB and SB . Volume range Tom 44 vph to 396 vph ( Inc.=44 vph) for NB and SB HE6 Case 5 Based on field data SP4 Cases 4 & 5 . Exclusive LT lane Only one lane for NB and SB in data set 6 Case 4 for data sets 2, 4, 5 & 6 and case 5 for 1 & 3 18 19 20 SP4-2 SP4-3 MUSEUM(0 1-02) 21 22 23 24 25 26 27 H3IR TESTN(l-9) SP3-lEX(1-4) SP3-lEX(5-6) PERMIT(l-9) MUSEUM(1-5) HEAVY(1-6) L I I Appendix F: Page 7

OCR for page 169
WELD DATA To fi=her verify the analytical mode} for traffic-actuated signal timing prediction, a field study was performed at the intersection of Museum Road and North-south Drive on the University of Florida campus In Gainesville, Florida. In total, 32 hours of data were recorded over three weekdays. The study periods included the morning peak the May peak and the afternoon peak. The site was a four-legged intersection with one through lane and one exclusive 250-foot left turn bay on each approach. The intersection layout is shown in Figure F-19. At this intersection, standard dual-ring phasing was applied with protected plus permitted left turns. The idea saturation flow rate was assumed to be 1900 vehicles/hour. Pedestrian recall was set in the controller and the duration for WALK plus Flashing DON'T WALK was 22 seconds. The intergreen time (yellow plus all red) was 5 seconds. The allowable gap setting was 3.5 seconds for through movements and 2.5 seconds for left turns. Ah detectors were 25 feet in length placed at the stopline. The minimum green time was set at 16.5 seconds for through phases and 10 seconds for leR turns. The maximum Keen time was set at 45 seconds for through phases and 15 seconds for left turns on Museum Road. The corresponding maximum Keen time settings for through and leD turn phases for North-south Drive were 30 and 10 seconds, respectively. IS Orion SB t"'".2. ~2.-. \::: 1 ~ Cob 1 1- l~o \1 .-.N . . . IS Drive NB 111 -^ Cl: Figure F-19. Intersection configuration of Museum Road and North-south Drive on the campus of the University of Florida Appendix F: Page 8

OCR for page 169
EXTRACTION OF SIGNAL TIMING AND DELAY ESTIMATES FROM NETSIM SIMULATION NET SIM was used as the simulation too} for the analytical mode! evaluation. NETS]M is able to mode} eight-phase dual-ring controller explicitly, recognizing all of the phase-specific parameters. NET SIM produces very detailed tables displaying several performance measures, including intermedi- ate values. It does not, however, provide sufficient information on the operation of the controller itself in the standard output tables. To obtain this information, it was necessary to develop a post- processor to extract the operational data from the special file used to support the animated graphics feature of NETSIM. The actuated-controller data for each second of operation are recorded and stored in a text file that is given a file name extension of ".F45" by NETSIM. A postprocessor was developed to read the .F45 file and produce a summary of the operation. The postprocessor has been caned "NETCOP" for "NET Sew Controller Operation Postprocessor?' [2~. It can produce phase-specific information such as percent gapout, percent maxout, average cycle length, average phase time, adjusted cycle length and adjusted phase time. A sample of NETCOP output is presented In Figure F-20. NCHRP PROJECT 3-48: CAPACITY ANALYSIS OF TRAFFIC ACTUATED SIGNALS ANALYSIS OF TRAF-NETSIM SIMULATION RESULTS FOR NODE 5 PHASE RING 1 RING 2 DATA 1(WBL) 2(EBT) 3(NBL) 4(S8T) 5(EBL) 6(WBT) 7(SBL) 8(NBT) # Displayed66 4165662766 Pcnt Skipped0.00 37.881.520.0059.090.00 Pcnt Gapout100.00 92.68100.00100.00100.00100.00 Pcnt Maxout0.00 7.320.000.000.000.00 Pcnt Forced0.00 0.000.000.000.000.00 Average G/C0.47 0.150.380.470.080.45 G/C % Var16.70 84.7628.1716.79117.6422.04 Av Grn Secs25.53 7.8520.6425.474.3224.23 Secs % Var28.27 89.1326.3728.43123.0227.45 Adj Grn 25.47 7.8520.6425.414.3224.23 Adj Grn (D)25.47 7.8520.6425.414.3224.23 Occup on Grn0.29 0.050.190.270.020.21 Dwell Time Total Cycles Dwell Adopted See -- Cycle Length - Sec Av Var Adj Pcnt Variation Green G/C Controller 3565 66 4 54 21 53 48 54 Detector 3565 66 4 54 21 53 48 54 Figure F-20. Sample of NETCOP output Appendix F: Page 9

OCR for page 169
Note that two Ends of circle lengths are shown in the sample. The "average cycle length" is obtained by dividing the total number of seconds processed by the total number of cycles. NETCOP trims partial cycles at the beginning and end of the analysis to ensure that only whole cycles are represented. The "adjusted cycle length" is computed by subtracting the number of seconds of dweD time (i.e., the time during which no demand was registered) Dom the total number of seconds simulated before dividing by the number of cycles. The "adjusted phase time" is computed according to the adjusted cycle length. Since the adjusted phase time can represent the effective use of phase time, it is adopted by this study for phase time comparison. The last two lines In Figure F-20 show the awed time, average cycle length and adjusted cycle length. The dweD time in the first line is called the controller awed time. This is the time during which no demand was registered on any phase. The dwell time in the second line is caked the detector dweD time. This is the time during which no demand was registered on any detector. In this study, the controller dwell time was used and the phase times were computed accordingly. Delay estimates can be obtained from the NETSIM output file. In this study, total delay was used for the delay comparison. A program called NETSUM was developed to read the NETSIM output file arid summarize the measure of electiveness (MOE). A sample of NET SUM output is shown in Figure F-2 I. SUMMARY OF TRAF-NETSIM MOE ~ S BY MOVEMENT Simulati on Time was 3600 seconds Measure Northbound 1, 5) Southbound ( 2, 5) Westbound ( 4, 5) r R L ~R L T R L r Eastbound ( 3, 5) R Volune Pent. Stops Qume Delay Stopped Delay 60.2 278.0 739.0 96.0 78.9 70.7 52.4 61.4 40.8 39.7 170.0 280.0 83.5 85.7 56.0 63.0 42.9 53.8 41.3 51.1 648.0 78.1 38.8 77.4 91.5 48.0 55.3 35.0 47.1 34. 1 46.0 202.0 888.0 87.2 87.7 59.9 65.0 47. 1 46.7 46.1 45.1 89.1 64.4 56.0 54.4 248.0 88.6 90.7 64.9 68.0 49.1 47.7 51.3 49.4 Figure F-21. Sample of NETSUM output Appendix F: Page 10

OCR for page 169
APPENDS F REFERENCES Courage, K.G., Wizard of Helpful Intersection Control Hints (A Computer Software), Research Version, Transportation Research Center, University of Florida, Gainesville, April 1995. 2. Courage, K.G. and P-S Lin, NET SIM Controller Operation Postprocessor (A Computer Software), Transportation Research Center, University of Florida, Gainesville, June 1995. ACKNOWLEDGiVIENT All of the intersection layout graphics were produced by the Signalized and unsignalized Intersection Design and Research Aid (SIDRA), developed by ARRB Transport Resources, Ltd. AppendlixF: Page 1 1

OCR for page 169