USF Libraries
USF Digital Collections

Automatic vehicle location for measurement of corridor level-of-service


Material Information

Automatic vehicle location for measurement of corridor level-of-service statewide feasibility analysis, final report
Physical Description:
56 leaves : charts ; 28 cm.
Florida -- Dept. of Transportation
University of South Florida -- Center for Urban Transportation Research
University of South Florida, Center for Urban Transportation Research
Place of Publication:
Publication Date:


Subjects / Keywords:
Motor vehicles -- Automatic location systems -- Florida   ( lcsh )
Intelligent Vehicle Highway Systems -- Florida   ( lcsh )
Global Positioning System   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Includes bibliographical references (leaves 51-53).
Additional Physical Form:
Also available online.
Statement of Responsibility:
by the Center for Urban Transportation Research, College of Engineering, University of South Florida.
General Note:
Prepared for the Florida Dept. of Transportation.
General Note:
"December 1994."

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001928699
oclc - 32077882
usfldc doi - C01-00320
usfldc handle - c1.320
System ID:

This item is only available as the following downloads:

Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200277Ia 4500
controlfield tag 001 001928699
005 20080327154220.0
008 950302s1994 flud b 000 0 eng d
datafield ind1 7 ind2 024
subfield code a C01-00320
2 local
b .A98 1994
0 245
Automatic vehicle location for measurement of corridor level-of-service :
statewide feasibility analysis, final report /
by the Center for Urban Transportation Research, College of Engineering, University of South Florida.
[Tampa] :
University of South Florida, Center for Urban Transportation Research,
56 leaves :
charts ;
28 cm.
Prepared for the Florida Dept. of Transportation.
"December 1994."
Includes bibliographical references (leaves 51-53).
Also available online.
Motor vehicles
z Florida
x Automatic location systems.
Intelligent Vehicle Highway Systems
Global Positioning System.
1 710
Dept. of Transportation.
University of South Florida.
Center for Urban Transportation Research.
8 773
t Center for Urban Transportation Research Publications [USF].
4 856


----------' -' ' ' ...... ------.... ,.. , .. :\ ;_ /_,/ of ----. /I 'Corridor Level-:of-Ser:v_ice: / I \ I ; '.., / ptatewide I I ; ; '' \ ' ' ' ' --' ' ' ' ' 1 ' \ '-I ', 1 ' \ ' \ ' \ / ,' .. ----.. ', \ I r I t' ', \ \ 1 1 I<' '\ \ I I I \ \ I \ \ \ r r ' ' I ' ' ' ' \ ' \ \ ' ' \ ' ' \ ' ' ' ' ' \ ' ' \ \ \ \ \ ' ' ' ' ' ' / ' ' ' ' ...... .. --' ' ' ' ' ,, Prepared by the Center for Urban Transportation R es e a rch for the Rorida Department of Transportation


Automatic Vehicle location for Measurement of Corridor level-of-Service: Statewide Feasibility Analysis Final Report Prepared for the Florida Department of Tra.nsportation by the Center for Urban Transportation Research College of Engineering University of South Florida December 1994


I. Executive Summary II. Purpose Ill. The Miami Experiment IV. Other U.S. Applications TravTek ADVANCE TRANSCOM Houston AVI Illinois Tollway CAPITAL Summary Table of Contents V. Comparing Two Methods of Data Collection VI. Comparing Two .Methods of LOS Calculation Variability of Speeds LOS Calculation LOS Calculation on Other Roadway Segments LOS Calculation on Exceptional Roadways VII. Other Possible Applications in Transportation Operations Signal Modifications Incident Detection and Response Techniques Travel Time Contour Maps Detour Routing Speeds Adjustment Factors Link Speed Calibration for FSUTMS Dynamic Traveler Information Goods Movement Operational Changes in Transit Special Event Traffic Handling Impact of Weather on Travel Speeds Safe Travel Through Work Zones "Mayday" Alerts Along Isolated and Rural Areas Summary Page 1 Page 3 Page4 Page 10 Page 12 Page 14 Page 16 Page 17 Page 19 Page 19 Page 21 Page 24 Page 27 Page 28 Page 28 Page 31 Page 32 Page 35 Page 35 Page 36 Page 37 Page 37 Page 38 Page 38 Page 39 Page 40 Page 40 Page 41 Page 41 Page 41 Page 42 Page 42


VIII. Statewide AVL Feasibility Dead Reckoning and Map Matching Signpost Ground-Based Radio-Navigation LORAN-C Global Positioning Systems (GPS) Differential GPS (DGPS) Cellular Phones Vehicle Location Data Collected by Fleet Operators Summary IX. Conclusion Bibliography Appendix A: List of Contacts Appendix B: Vehicle Trips Used in LOS Calculation Page 43 Page 43 Page 43 Page 44 Page 45 Page 45 Page 46 Page 46 Page 47 Page 48 Page 49 Page 51 Page 54 Page 56


Table 1 Table 2 Tab l e 3 Table 4 Table 5 Tab l e 6 Figure 1 Figure 2 Figure 3 Figure 4 List of Tables Differences Between Observed and Calculated Travel Page 6 Speed : Confidence Interval Statistics Summary of U.S. Applications of AVL in Transportation Page 22 Planning Leve l s of Service for Basic Freeway Sections Page 30 Level -ofService Calculation on South Dixie Highway Page 33 Btw. 27th Ave and 17th Ave. Gene ralized Peak Hour Direction al Volumes for Florida's Page 34 Urbanized Areas: Level-of-Service Criteria AVL Positioning Techno l ogies Page 47 List of Figures Difference Between Observed and Calculated Travel Speed Method 1 Histogram and N ormal Distribut i on Difference Between Observed and Calculated Travel Speed Method 2 Histogram and Normal D i stribution Cost of Data Collection Using Con ven ti on a l and AVL Methods Sample Road Segment Dolphin Expressway Page 8 Page 9 Page 26 Page 29


I. EXECUTIVE SUMMARY Florida Department of Transportation has asked the Center for Urban Transportation Research (CUTR) to investigate the statewide feasibility of using automatic vehicle location (AVL) to measure roadway performance. This project builds on the findings from "The Miami Experiment," a previous operational test of ground-based AVL which CUTR conducted for the City of Miami. The findings of this research are: The Miami Experiment is the only IVHS project which currently uses data gathered by automatic vehicle location for transportation planning purposes. Using AVL to collect average travel speed can be less expensive than conventional travel-time studies if there are sufficient number of average speed measurements to amortize the fixed costs. Average travel speed data collected by the AVL can be used to calculate level of-service on roadway segments T he vehicle location and travel speed information collected by AVL systems is useful in several other transportation p lanning applications other than level-of service determination. Satellite-based global positioning systems (GPS) offers the best AVL technology option if FOOT wanted to repeat a data gathering experimen t similar to the one conducted in Miami in other areas of the state. Using vehicle location data gathered by Florida trucking companies whose vehicles are equipped with AVL transponders appears to be a viable option for gathering average travel speeds in the state's urban areas. Through the course of this study CUTR has identified additional areas of research which would be beneficial to FOOT to have investigated in the future: 1


Conduct a data gathering experiment similar t o the experiment cond u cted in the City of M iami in B ro ward or Pa lm Beach Coun ty using the groun d-based radio navi ga tion A VL t echno logy available there Driver recruitment will be s pecifically targeted at achieving cove rage of all cong es ted corridors in the study area in the peak direct ion during the peak period. Cond uct a data g a the ring expe riment s i mil a r t o the e xpe riment conducted in the City of M i am i, in a n urban or rural area in Florida outside of Dade Broward or Palm Beach counties using Sli!tellite-based global pos itioning system {GPS) techno l ogy to locate vehic l es Further investigate the suitability of vehicl e l oca tion data gathered by Florida trucking companies whose vehicles are equipped with AVL transpond e rs Details such as the number of equipped vehicles routes take n by eq ui pp ed vehic les and format of veh i cle l ocation data nee d to be specified. 2


II. PURPOSE On June 16 1 994 the Florid a Department of Trans p orta tion entered into a contract with the University of South Florida on behalf of the Center for Urban Trans portat i on Research to inves tiga t e the statewide feasibility of using automatic veh ic l e location to measure roadway performance. Th i s project builds on the find ings from a previous oper a tion a l test of ground based AVL which CUTR conducted for the City of M i ami. In March of 19 94 the City of Miam i contracted with the CUTR t o conduct a field operation al test of the use of g roun d-based AVL technology to measure vehicle operating speed s on sevente en of the City's transportat ion co rridors. CUTR se t up a data gathering experiment that used data compiled fr om A VL tra nspond ers in stalle d in the vehicles of 25 volunt eer drivers. The technology vendor of the AVL s ystem was AirTouch Teletrac w h ich supplied the equipment at m i nima l cost. The "Miami Experimen t ran from Apri l 25 to Au g ust 15 1994 reco rding over 4,40 0 vehicle trips This Final Report i s a compil a ti o n of tech nical memor anda performed for Tasks #1, #3 and #4 of this pro ject. Other instances of the use of AVL in transp ortation planning applications are explored, and differences between the cost of AVL and tradi tional methodol ogies of collecting data on vehicle oper ating spee ds are compared This Final Report compa res two metho ds of level -of-service (LOS) calculation on one samp le roadway : (1) the Florida Departm e n t of Transportation es t im a tion using traffic volume count s input in to the Generali zed Tables of FOOT's 1992 Florida Highway System Plan : Level-ofService Manual, and ( 2 ) average tr av e l s peeds based on the location data gathe red by the AVL system. Several other applica tio n s of AVL technology to gather information (besides average travel speeds) useful to transpo rtation professionals are identified Fin a lly this report assess es the various AVL technologies in terms of coverage cost and positional accuracy. Thi s assessm ent is performed in order to evalu ate t he feasibility of conducting a sim i lar op erat ional test of automatic vehicle location technology in othe r areas of the s tate. 3


Ill THE MIAMI EXPERIMENT I n March of 1994, th e Ci ty of Miami contracted with CUTR to set up a field operational test of the use of AVL to measure vehicle operating speeds on the city's seventeen transportation corridor s CUTR set up a data gathering experiment which used AVL units i nstalled i n the v ehic les of 25 volunteer drivers w ho traversed the City roa d network as part of their normal dai ly commute. The techno l ogy vendor of the AVL system was AirTouch Tele trac wh i ch donate d the equ i pment at m i nima l cost. The M iami Experiment" ran from Apri l 25 to August 15, 1 994 record ing over 4 400 vehicle trips. The AirTouch Teletrac AVL system locates vehicles using a positional technology called ground-based radio navigation, (a.k .a. 'terrestrial rad io-navigation and "signal tri late ration ") When u sing this type of positioning technology, the AVL vendor sets up a network of recei ving antennas throughout a metropo l itan area. Each probe vehic le is equipped with a device ( a transpond er' ') wh i ch bro-adcasts a radio-frequency ( RF) s i gnal to a ll nea rby antennas Bas ed on the time i t takes for the signal to t ravel from the transponder to the antenna the system can determ ine the distance between t he veh i cle and e ach antenna If the signal was received by three or more antennas, the vehicle's position can be uniquely determined Gro und-based radio-navigation uses a less sophisticated technology than the sa te llite based globa l positioning system (G PS ) and, consequently has l ess precis ion. Ground based radio-nav i gation AVL systems are among t he l east expensive AVL systems fo r the user However s i nce constructing th e necessary infrastructur e ( i e. rece iv ing antennae ) require s signi fican t financ i al investment on the part of the AVL vendor these systems are u sually only available in dense urban areas with large market potential AirTouch Teletrac has radio-navigation AVL systems operating in Los Angeles C h icago, Detroit, Da llas/Ft. Worth, Houston and south F lorida The south Florida AirTou ch Te l etrac system can locate a vehicle anywhere in Dade, Broward and Palm Beach count ies. The company g uaran t ees the accura cy of i ts A VL sys tem to the nearest 150 feet. Due t o south F l orida s rela ti vely flat terrain the accuracy of t he south Florida sys t em is generally w ithin the nearest 50 feet. AVL systems which use augmented g l oba l positioning systems technology are generally accu r ate to withi n 16 fe et. 4


The City of Miami was responsible for recruiting volunteer drivers. Many of the drivers were City of Miami or Dade County employees who live on the periphery of the city and commute daily to downtown, thus providing coverage of 5 of the 17 corridors in the peak direction during the peak period. Incentives to participate in the experiment included receiving a free vehicle breakdown and stolen vehicle recovery service for which many south Floridians pay a one-time $300 per vehicle start-up fee plus a monthly $10 service charge. Disincentives included a certain loss of privacy which left some potential volunteers unwilling to make the tradeoff CUTR, AirTouch Teletrac and the City assured volunteer drivers that their vehicle would be tracked only by its assigned number and that all vehicle location in formation would be used only for the stated purposes of the study. (How other IVHS projects handle the privacy issue will be addressed in Section IV. Other U.S. Applications.) AirTouch Tel e trac donated use of its AVL syste m to the City for minimal cost. CUTR paid AirTouch Teletrac $2,500 for insta llation and removal of the 25 trans ponders, for a workstation using FieetDirectorM software for 120 days. AirTouch charges its commercial customers $7,500 plus $250 per month for a similar equipment rental and services. The AirTouch Teletrac AVL system polled the vehicles for their locat ions every 30 seconds when the vehicles were on and every 5 minutes when the vehicles were off. The vehicle locations were recorded by the AirTouch Teietrac fleet management software FleetDirector"'' at a workstation located at the City of Miami offices. FleetDirector"' wrote the vehicle location data to a file for a 5-hour period in the morning and a 5-our period in the afternoon peak periods on weekdays, plus a period on Saturday. Every week City of Miami staff sent the latest copies of the files to CUTR in Tampa for analysis. CUTR researchers wrote two software programs to analyze the vehicle location data. One program, SPEED.EXE, averages travel speed for an entire trip, from the moment the vehicle ignition is turned on to the moment the vehicle ign itio n is turned off. The other, SEGMENT.EXE, correlates the geographic coord ina tes of road segments to the vehicle's recorded locations to derive average speeds for particular segments of the commute trip. 5


Both programs used two methods of calculating average travel speed Method 1 averages the speed values output by Fleei Oltectorr"'s over the entire trip Method 2 averages values for distance traveled between location readings. The FleetDirector,.... software has never before been used to track vehicles to calculate their speed. To determine if the system was producing valid data, CUTR researchers established a validation process to compare the observed values for average speed over a vehicle trip to values calculated by CUTR's data analysis software. City of Miami transportation planning staff, who were driving equipped vehicles, recorded the starting and ending times of their trips, plus the distance traveled as recorded by their car odometers. Comparing the data collected automatically and manually for the 30 validation runs conducting during the 113-day data gathering period, the mean difference was -1.07 mph for method 1 and +1.07 mph for method 2. Assuming that the differences between observed and calculated values follow a normal distribution pattern the 95% confidence intervals for the differences can be calculated, as shown in Table 1 Table 1 Difference Between Observed and Calculated Travel Speed Confidence Interval Statistics Statistic Difference: Difference: Method 1 Method 2 Mean -1.07 mph +1.07 mph Standard Deviation 3 .91 mph 2 .41 mph Standard Error 0.71 mph 0.44 mph 95% Confidence I n terval -1.07 mph 1.40 mph +1.07 mp h :1: 0.86 mph Figures 1 and 2 show a histogram of lhe differences for the 30 va l idat ion runs compared with a normal distribution (commonly called a "bell" curve.) Since Method 2 is the more accurate method with a smaller standard error its normal distribution pattern more closely resembles a bell shape. Since Method 2 is the more accurate, it is used in 6


calculation of average travel speeds in Section VI. Comparing Two Methods of LOS Calculation. Designing and implementing the data gathering experiment was Phase II of a three phased proj ect CUTR is performing for the City of Miami. In Phase Ill, it is anticipated that CUTR will use the data analysis software to compute average travel speeds for each of Miami's seventeen corridors in the peak direct i on during the peak period. The City Planning Div i sion then has the option o f using those computed average speeds to assist in its determination of corridor l evel -ofservice. 7


0.18 I 0.16 I 0.14 0 .12 > c 0.1 .. "' .. 0 .08 Lt.. 0 .06 0.04 0.02 0 Figure 1 Difference B etween Observed and Calculated Travel Speed Method 1 H istogra m and Normal "Be ll Curve IY.&'A.&11 l:i'M@'#A'%1 -BELL CURVE ---I -6.11 -5.39 .67 -3.95. -3.23 2.51 -1.79 -1.07 0 .35 0.37 1 .09 1.81 2.53 3 .26 3.97 Average Values


0.2 0.18 I 0.16 I 0.14 >-0.12 (.) c .. 0 1 :I a "' u. 0.08 0.06 0 .04 0.02 0 .... M .,., .... M "' M M N ' Figure 2 Difference Between Observed and Calculated Travel Speed Method 2 Histogram and Normal "Bell" Curve I ---------BELL CURVE _, I "' .... M "' "' .,., M .... "' .,., M .... -"' m M .... ... "' -

IV. OTHER U.S. APPLICATIONS The term "Intelligent Vehicle Highway Sysiems"1 is used to describe projects which apply advanced technologies to improve the efficiency and capacity of transportation systems. Passage of the l ntermodal Surface Transportation and Efficiency Act (ISTEA) of 1991, with its emphasis on IVHS, focused national attention on this emerging field. The ISTEA brought more than exposure to IVHS, authorizing $660 million over six years for research and operational test of IVHS technologies. During the Clinton Administration, Congress has appropriated $90 million for IVHS in addition to the ISTEA funding. States and regional authorities have followed the federal lead, also providing funding for numerous IVHS projects. There are many ways in which advanced technologies can be applied to transportation. The two types of IVHS projects discussed in this report are Advanced Traveler In formation Systems and Electronic Toll and Traffic Management. Advanced Traveler Information Systems (A TIS) projects use a variety of technologies to communicate real time up-to-the-minute traffic informat ion to travelers using a variety of modes. Electronic Toll and Traffic Management (ETTM) p rojects eriable drivers to pay tolls without stopping their cars at toll plazas, where payment is accomplished using wireless communications between a transponder ("tag") inside vehicle and an antenna installed .at the roadside. Automatic vehicle identificat ion (AVI) is the technology which makes ETC systems possible. Since passage of ISTEA in 1991, the U.S. Department of Transportation has funded over 15 A T IS -type projects. Twelve toll authorities in the U.S. have installed and eight agencies are planning to install ETC systems on their toll roads. While the term ETC describes the use of AVI technology for more efficient collection of tolls, without any action required by the driver or toll collector, ETC is the foundation of electronic toll and traffic management (ETTM) systems. ETTM uses AVI technology not only for toll collection, but also for more broad-based traffic management purposes, such as the projects described in this report 1 The term "Intelligent Veh icl e Highway Systems (IVHS) is slowly being replaced with "Intelligent Transportation Systems (ITS)". The tran sition is bei ng made so that the nam e encompasses all aspects of t ransportation, inc lu ding freight and transit, not just passenger vehicles and highways This report uses "IVHS" throughout the text, although "ITS may be substituted. 10


Another type of IVHS application is Automatic Vehicle Location (AVL). AVL is a means of continuously monitoring the location of vehicles in a road network. Typically vehicles are equipped with a transponder, the size of a video cassette tape which transmits a radio-frequency (RF) signal to a central location at regular intervals. AVL systems are being used by all kinds of customers in all kinds of applications around the world. Delivery companies use A VL to plan the most efficient dispatch of their fleet vehicles. Transit agencies use AVL in conjunction with information displays to inform passengers when the next bus will actually arrive, as opposed to when it is scheduled to arrive. Par atransit operators use AVL to log the distance traveled by Medicaid patients, and la ter use this data when applying for reimbursement from the state. Private citizens can even subscribe to an A VL service which will instantly dispatch a tow-truck to their car in the event of a breakdown or to recover their car if it is stolen. Vendors can set up AVL systems with relatively little investment in infrastructure, and consequently litt le need for federal or state financial support GPS World magazine, The IVHS Index, IVHS America's APTS Vendor Catalogue and the Federal Transit Administration's APTS State of the Art report list over two dozen vendors of A VL systems us ing a variety of different positioning technologies: dead-reckoning, map matching, LORAN-C. ground-based radio navigation and global positioning systems (GPS). Eleven transit agencies in North America have installed and fifteen agencies are planning to install AVL systems on their buses (For more detailed information on Florida transit agencies' use of AVL, see Section VII. Statewide Feasibility.) Of the travel information and electronic toll collection projects in the United States, the following are the only projects which use AVL to gather traffic information: TravTeKOrlando, FL ADVANCE Project Chicago, IL TRANSCOM Project "E-ZPass" New York City, NY Illinois Tollway "1-Pass" Chicago, IL Houston AVI -Houston, TX CAPITALWashington, DC 11


Each of these projects is profiled, providing information on the following issues; Who are the participants? What kind of positioning technology does the AVL system use? What kind of vehicles are being used as probes? How many probe vehicles/drivers are expected to participate? How are probe vehicles/dr i vers recruited? How are drivers concerns about privacy addressed? What kind of data is collected? What kind of analysis is performed on the data and what is its output? Who will use the data and for what purpose? What is the current status of the project? How much does the entire data collection system cost? A summary of these profiles appears in Table 2. TravTek When IVHS was first conceptualized in the l ate 1980's the most popular application of technology to transportation was route guidance, (al so called "in-vehicle navigation"). For a route guidance system, special computers are installed in vehicles. The computers contain map databases and have the ability to locate their own position. A driver inputs his desired destination, and the in-veh i cle unit gives him directions on how to get there. Development of in vehicle navigation units has progressed considerably i n recent years Oldsmobile will begin selling in-vehicle navigation units as an option in their Eighty Eight LSS performance sedan in California at the end of 1994 Dynamic route guidance takes the navigation task one step further by r eceiving information about current traffic conditions on the road network In giving directions to the driver, the in-vehicle uni t allows the driver to avoid areas and roads with heavy traffic congest i on TravTek, a one year operational test of IVHS technology conducted in Orlando, FL from March 1993 to March 1994, emp l oyed dynamic route guidance TravTek obtained its information on current traffic conditions from loop detectors embedded in Orlando-area roads, video surveillance cameras i nstalled along lnterstate-4 12


and other stationary sensors. This information was collected and processed at a Traffic Information Center (TIC) staffed by the City of Orlando traffic engineers. The American Automobile Association, Federal Highway Administration, Florida Department of Transpo rtation, General Motors Corporation and the City of Orlando made financial or in-kind contributions to the test. Each of the 100 vehicles TravTek was equipped with a computer touch-screen display, map database on CD-ROM, on-board compass and odometer and global positioning system (GPS) receiver. The positioning system used a combination of GPS, dead reckoning and map-matching to locate the vehicles. Volunteer drivers most of whom were tourists visiting Orlando rented the vehicles through Avis rental car agency at the Orlando Internationa l Airport. The American Automobi l e Association promoted the program to its members nationwide. Over 6,000 volunteer drivers had the opportunity to drive TravTek vehicles. Privacy concerns were not explicitly addressed as part of the project. Using electronic maps from both Etak, Inc. NavTech as a base, the TravTek project mainta ine d an estimated travel time fo r each link in the road network. The map database covered a 1,200-square-mile a r ea of west central Florida, maintaining travel times on over 1,400 links. Traffic management software estimated link travel times from a variety of dynamic sources: surveillance cameras i nstalled o n lnterstate-4 . incident reports logged by the traffic information center operators, the city's traffic signal control software, the city's road maintenance and construction schedules, and probe data from the TravTek vehicles themselves. If no dynamic information on a link's travel time was available, the system used a historical estimate based on the time of day and day of week. Each TravTek vehicle sent a report every minute to the TIC Each probe report contained the vehicle 10 number, latitude and longitude of the vehicle's position speed and direction, and the last three l inks traveled and their travel time 13


Software deve l oped by Farradyne Systems Inc fused data from these various sources into one travel time est i mate for each link. Probe vehicles accounted for only about 1 0% of the estimated travel times for links on the road network The city's traffic signa l control software accounted for 28% of the estimates. The system relied on historical infomnation for 56% of its estimates. The remaining 6% of est i mates were based on anedotal reports of incidents reported by TIC staff. The dynam i c estimates for link travel times were broadcast to the TravTek vehic les, in one-minute intervals, so that the navigation software cou l d use this data i n its calculation of the quickest route to each driver's destination The dynamic link travel times were not used for any other purpose Since the conclusion of the TravTek project in March 1993 software developers from Farradyne Systems have not trained Orlando city traffic engineers on use of the traffic management software, which still resides in the Orlando TIC Although data from the TravTek probe vehicles is no longer being collected, data from other sources could still be used to estimate vehicle trave l speeds on road segments. The City of Orlando is in negotiation with Farradyne on th i s matter.. TravTek was among the first large-scale field operational test of IVHS technologies. The total project cost was $8 million. ADVANCE Like TravTek, the ADVANCE project (Advanced Driver and Vehicle Adv i sory Nav i gation Concept) tests dynamic route guidance, but will rely much more heavily on data from vehicle probes The Federal Highway Administration Ill i nois Department of T r ansportation (lOOT), Northwestern university, University of Illinois at Chicago and Motorola are participat i ng in the project. F i nancial support for ADVANCE is provided by the FHWA lOOT and other pri vate commercial I nterests Each probe vehicle will be equipped with a Mobile Navigation Assistant (MNA). The MNA is composed of a computer, CD-ROM, driver interface, wheel speed sensors a compass and global position ing system (GPS) receiver Probe vehicles will report travel 14


time for road links as they traverse the test area. The data will be collected by on-board equipment then transmitted to a Traffic h'lfonnation Center (TIC). Information is also collected from loop detectors embedded in various parts of the roadway, plus anecdotal reports from drivers v i a cellular phone. The project test area will encompass a 36-square-mile area qf Schaumburg Illinois, an affluent suburb of Chicago, which happens to house Motorola's Corporate Headquarters. The project is intended to eventually include 5,000 private and commercial vehicles. Driver recruitment began in August 1994. The ADVANCE projects sent press releases to local, regional and community newspapers. ADVANCE staff held promotions of the volunteer program {including a demonstration of the route guidance techno l ogy) in local shopping malls. Direct mail advertisements were also distributed to area residents. The ADVANCE project set up a hotline to handle telephone inquiries Mondays through Saturdays Northwestern University developed criteria to select among willing volunteers; the criteria will ensure that adequate coverage of the test area is achieved and that the poo l of selected drivers is a representative sample of area drivers. Each vehicle will be assigned a unique identification number, and vehicles will be tracked using only that number to maintain drivers' privacy The basic data set will be transmitted via radio frequency signa ls, including vehicle number, l ocation and time of each reading. The MNA also records the driver's interaction with the system, so that statistics can be compiled on the options selected, route taken, and frequency of use. Argonne Nat i onal Laboratories will analyze th i s type of data. ADVANCE has received substantial funding as part of the annual U.S. Department of Transportation appropriations process, with over $10 million earmarked for the project to date. The estimated total project cost is $52 m i llion. At the conclusion of the test, lOOT will be left with state-of-the-art traffic surveillance equipment and Motorola will have developed an in-vehicle navigation unit which it can sell as an after-market product on passenger and commercial vehicles. 15


TRANSCOM Unlike the ADVANCE project or the Miami Experiment, TRANSCOM will use technology developed and implemented for another purpose electronic toll collection -to collect traffic information using probe vehicles. In electronic toll collection, vehicles are equipped with a transponder which communicates only when they pass near antennas installed at specific points along the roadway. This kind of system is called Automatic Vehicle Identification (A VI) to distinguish it from Automatic Vehicle Location (AVL) in which transponders constantly transmit their location to a central receiving point. Drivers who use the toll road regularly purchase or lease a transponder, plus make a deposit on a payment account of future tolls. Transponders can cost anywhere from $10 to $40. When the driver passes through a toll plaza, the AVI system automatically decrements the driver's pre-paid toll account balance. In this way, the driver pays his toll without having to stop at the toll gate. Toll authorities are turning to electronic toll collection as a way of reduc ing traffic congestion near toll plazas, and make traveling along those roads more appealing to drivers. Seven toll authorities in Pennsylvania, New York and New Jersey hav e agreed to install the same ETTM system for all of their roadways. (Two thirds of all toll revenue collected in the United States is collected in these three states.) The coalition calls itself the Inter Agency Group (lAG) and the electronic toll collection project "E-ZPass The lAG recently selected Mark IV Industries as the ETTM vendor. However, prior to Mark IV's selection, two member agencies Installed an ETTM system from another vendor Amtech for an initial feasibility test. The TRA NSCOM project will use the 40,000 drivers who have purchased ETTM transponders as probe vehicles to determine traffic conditions on the New York State Thruway and Garden State Parkway. When the system is operational, software developed by Farradyne, Systems, Inc. will randomly select a vehicle at a toll plaza equipped with an E-ZPass transponder. T he vehicle will be assigned a random tracking code and tracked using RF antennas along the two participating toll roads. Tracking code, location and time will then be recorde d and used for real -t ime incident detection. Farradyne has developed software to analyze vehicle location data, report on possible traffic incidents, and notify emergency response agencies. 16


TRANSCOM member agencies are currently installing roadside antennas. The system shou l d be operational by October of 1994 However, according to the TRANSCOM office, the c o mmittee has no plans to make the t r avel time .data available for transportation planning purposes. Like ADVANCE and TravTek, TRANSCOM has received substantial federal funding and earmark ing through the annual USDOT appropriations process. The estimated project cost for the ETTM portion of TRANSCOM is just over $2 million. Houston AV I The Texas Department of Transportation (TxDOT) is involved in two projects i nvolv ing volunteer drivers as probe vehicles in the traffic network. The first project entitled the "Cellular Telephone Demonstration Project used calls from selected drivers via cellular te l ephone to mon i tor traffic conditions on a Houston-area trave l corridor composed of parallel roads north of downtown Houston -the Hardy Toll Road, 1-45 and State Road 59. Two hundred probe vehicle dri vers were se l ected from a pool of drivers who use those routes as part of their normal daily commute. Vo l unteers received a ce ll ular telephone free of charge. In return, drivers were required to call a central telephone number and report on traffic conditions and incidents t h at affected t r affic flow As a volunteer dri ver passed a designated receiving station, he was required to call the TxDOT control center and report the station number and i dent i fication number of h i s vehic le. The operator would record the informat ion, plus the time and date of the call, ente r ing it into a travel time analysis database. Relaying the information and entering it into the database took 10 to 15 seconds per call Each probe vehicle niade between 10 and 12 reports per day. Software written by the Southwest Regional Transportation Center of Texas A&M Un i versity ana l yzed the database and gave a report of current travel conditions on the se l ected roadways. TxDOT used the current travel conditions to update its network of eight variable message signs and to instruct its incident response teams The information was also faxed to the broadcast studios of the three Houston-area traffic advisory servi c es. 17


The second Hous ton -area project, l i ke the TRAN SCOM and Illinois Tollway projects uses ETTM patrons as vehicle probes. TxDOT I s installing readers at 2 to 4 mile intervals on over 120 miles of freeways and 100 miles of reversible HOV lanes. Installation of readers will be conducted in three phases. Phases I and II cover 1-10, 1-45, SR 59 and SR 290, and are nearly complete Phase Ill will cover l-610 (the beltway) and the Sam Houston To llway Amtech Corporation of Dallas, TX was selected by Harris County Toll Authority as the ETC vendor The 32,000 drivers who have already purchased ETTM tags wi ll not be used because of privacy concerns TxDOT has issued 1,000 transponders to drivers who trave l 1 10, SR 290 and 1-45 as part of their normal daily commute TxDOT p lans to issue tags to an additional 3 200 volunteers A clause in the volunteers' contract sp ecifically states that the vehicle's travel time data will not be given to police for the pu rpose of issuing tickets to drivers The Metro Transit Authority has also installed transponders on buses wh i ch use exclus i ve h ighoccupancy vehicle (HOV) l anes i n these corridors As each vehicle passes a reader antenna informat i on o n the locat i on time and vehic l e identificati on number is collected at the roadside by the reader and tran smitted by rad i o to nearby field stations equ ipped with modems and telephone lines. The inf ormatio n is then sent by phone to the Central Control F acllity (CCF). Using software s imilar to what was used in the Cellular Telepho ne Demonstration Project, the trav e l time database i s analyzed to produce a summary of current trave l conditions. The information i s then distributed to v a ri ous u sers of real-time information In an article published in IT Journal two of the Houston A VI project managers allude to using the travel data to measure roadway performance, however no firm plans exist to date. "Since so many measures of eff ectiveness rely on the impact of speed and travel times, it is import ant to have the most reliab l e data avai labl e." stated Steve Levine of TxDOT and William McCasland of the Texa s Transportati on Institute The cost of Phase I i nstalla ti on was $2 million Th e cost of the other two i n stallation Phases are unknown 18


Illinois Tollway Along with the TRANSCOM committee and Harris County Toll Authority, the Illinois State Toll Highway Authority has plans to use patrons of an electronic toll collection system as traffic probe vehicles. In June 1994, the Illinois State Toll Highway Authority announced that it had selected ATIComm as the technology vendor and Science Applications International Corporation (SA IC) as the systems integrator for an ETTM system called "1-Pass" for over 200 lane miles of toll roads near Chicago. The Authority selected ATIComm after the vendor had successfully completed a test of the technology on the North-South Tollway invo lv ing 2,500 patrons. The contract requires installation of ETTM equipment on the entire North-South Tollway and on the central portion of the Tri-State Tollway, and involves the distribution over 10,000 electronic toll collection tags Following their interest in IVHS with the ADVANCE pro ject, the Illinois DOT has reached an agreement with the Illinois State Tollway Authority that will allow lOOT to use motorists' 1-Pass tags as traffic pr obes lOOT is negotiating with SAIC to design the hardware and software to determine average vehicle travel speeds on the tollway, based on data obtained by the ETTM readers at the toll p la zas SAIC's contract for the entire electronic toll collection system is worth about $12 million. The cost of the veh icle trave l speed data collection portion of SAIC's work is unknown. CAPITAL Instead of ground-based radio-frequency signals, global positioning systems, or automatic vehicle identification, the CAPITAL project will test the use of ordinary cellular telephones as AVL transponders. Engineering Research Associates (ERA), a subsidiary of E-Systems based-in Tampa, is developing a technology called E-CA PS which essentially turns any cellular telephone switched on to receive calls into an A VL transponder, and any vehicle with such a phone in it into a traffic probe. Bell Atlantic Mobile Systems is the vendor of cellular telephone services which has donated the use of its network for the test. The Maryland State H igh way Administration and the Virginia 19


Department of Transportation (VDOT) will make use of the traffic data gathered from probe vehicles. The test will gather traffic data on portions of three interstate highways in Virginia and Maryland outside of Washington, D.C.: 1-395, 1-495 (the beltway), and 1-270, plus some arterial streets which cross the interstates. The operational test is scheduled to begin in the summer of 1995 and last 90 days. Farradyne Systems is developing software to convert the vehicle l ocation data to real time traffic data on traffic speed and incid ent locations. F arradyne will maintain a database of traffic conditions and make available a graphic display of current traffic conditions to VDOT and the Maryland State Highway Administration via a phone line and modem connection. Unes representing the highways on an electronic map will change color to indicate average vehicle speeds. The maps will also highlight the locations of incidents the system has detected. Pri vacy concerns are an important issue, since all of the Washington, D.C.'s estimated 117,000 cellular telephone users are potential probe vehicles, but none have explicitly given their consent to participate in the test. ERA also provided documentation that their system will not violate the federal Telephone Disclosure and Disputes Resolution Act, which prohibits the use of scanners that intercept cellular telephone calls. The E-CA PS system only detects the locat ion of the cellular phone not the content of the calls. Vehicles are identified by number randomly assigned at the time the E-CAPS system locates the vehicle, not by the cellular phone's telephone number or serial number. In addition to providing a traffic infonmation database to transportation officials, Farradyne is also developing hardware and software to deliver the database over the cellular telephone communications network to special receivers installed in vehicles. Approx imately 10 vehicles belonging to a delivery service that has agreed to participate in the test will receive Farradyne's traffic informat io n via cellular telephone modems and laptop computers. The Maryland State Highway Administration is currently installing new software that will use data from multiple sources in-pavement loop detectors, overhead radar detectors 20


and the cellular telephone probes to mon i tor traffic flow i n real t i me Th e agency expects this system to be operational i n two to three years. Federal Highway Administration is contributing $5.5 million of the total estimated project cost of $7.1 million, which includes equipment Installation, software development, and actual day-to-day expenses of conducting the test. Summary Five federallyand regionally-funded projects in the United States projects are using vehicles as probes to collect information on traffic conditions. In each of these five projects, the data is used to del ive r real -time traffic informatio n to various users: state department of transportation, incident management and emergency response teams, radio and TV traffic Information broadcasters and even drivers with dynam i c route guidance units In thei r vehicles O n ly the Hous t on AVI project h as cons ider ed t h e poss ibil ity of u s ing the data co ll ected to measure level-of-serv i ce The M i am i . . Experiment is the only IVHS project in the Uni ted States to use data gathered from an AVL system for transportation planning purpose!!. 21


Ta ble 2 Summary of U S. Applications of AVL in Transportatio n Planning Project Participants AVL Number of Data Collected Purpose of Data Current Total Positioning Probe CoOection Status Project Technology Vehicles Cost TravTek MA GPS 100 surveillance video dynamic route Project was $8 million FHWA inci dent reports guida nGe completed Florida DOT vehicle location data March General Motors from probe vehicles 1993. City of Orlando ADVANCE FHWA GPS 5,000 traffic counts dynamic route Driver $52 million Illinois DOT anecdotal reports from guidance recruitment U of Illinois drlvers via cellular began Northwestern U phone August Motorola vehicle location data 1994. from probe vehicles TRANSCOM ToO authorities Signpost 40,000 vehiCle tocatlon data notify radio System $2 million of PennsylVania, (electronic from cars passing stations & should be New Yol1< and toll though toll plazas emergency operational New Jersey collection) response by 1994. vehicles Houston AVI Texas DOT Ca llers 200 anecdotal reports from variable Project was N/A Project #1 Harris County Identify own d riv ers via cellula r message signs completed Toll Authority location phone notify radio by March Texas A&M U stations 1994. ---------------------------------------22


Tab l e 2 (Continued) Summary of U .S. Applications of AVL in Transportation P l anning ---Project Participants AVL Number of Data Collected Purpose of Data Current T otal Positioning Probe Collection Status Project Technology Vehicles Cost Houston AVI Texas DOT Signpost 3 300 vehi cle location data varia ble I nstallation S2 million Project #2 Harris County {electronic from cars passing message signs of T ol Authority toll through toll plazas notify radio antennae Texas A&M U. collection) stations on 3 of out 5 roadWays is nearly complete Illinois Illinois DOT Signpost 10,000 vehicl e location data Not yet ETC $12 million Tollway Illinois State {electronic from cars passing determined vendors for entire Tollway Authority toll through toll plazas selected in ETC collection) June 1994 project CAPITAL E.Systems Celular 117 000 vehicle location data Not yet System $7 1 Maryland DOT phones from probe vehicles delermined should be mRiion Vwginia DOT (automatic operational Bell Atlantic location) in Sunvner Mobile Systems 1995. Miami City of Miami ground-25 vehicle location data test of Project was $25,000 Experiment CUTR based from probe v ehic les service comp leted AlrTouch radio calculation August Teletrac navigation 1994. 23


V. COMPARING TWO METHODS OF DATA COLLECTION The start-up costs of conventional travel-time studies are relatively low, while the day-to day cost of conducting the tests paying drivers, vehicle rental, gasoline, etc. -is relatively high. In an estimation of costs of different types of congestion management performance measures, JHK & Associates estimate that the start-up cost to be negligible and the operations cost to be $2 per vehicle mile of data collected. Start-up costs of the Miami Experiment included the $25,000 contract to CUTR to set up the test, supervise installation of the vehicles and write the necessary data analysis software. Although AirTouch Teletrac provided use of its system at a discount, the company usually charges its commercial customers $300 per vehicle start-up fee plus $10 per vehicle per 30 days rental. Assuming that vehicles make two daily commute trips on weekdays with an average distance of 12 miles, and assuming the data gathering experiment involved 25 drivers then either method would be collecting average speed data for 600 vehicle miles of travel per day. The cost per mile of data collection using AirTouch's AVL system and 25 drivers would be: $32,500+( $IO )(25vehic/es)(duration) = [ 30da;:s l (4.3weeks )(5days)(2commute -trips)( 12m1/es )( 25 vehicles)(chJration) 30days week day commute tnp = [($32,500+$8.33 Duration) ] (430* Duration) 24


This model also assumes that: no trips are recorded on weekends (Including weekend trips wou l d decrease the cost of data collection per vehicle mile even further ) drivers are volunteers and are not paid. Figure 3 shows the cost of data collection for conventional and AVL methods versus duration of the data gatheri n g experiment. The cost values for the AVL method was obtained using equation for the Miami Experiment described above For data gathering periods of longer than 45 days (27,000 vehicle miles), the AVL method is t he least expensive For data gathering periods of almost 1 year, the difference in cost is a l most two orders of magnitude. 25


6 -., .. '5 Q 6 -., -2 ., 4 u :c .. > .. Q. 3 c 0 ;; <.> .. '5 2 (.) "' 1;j Q ... 0 1 (.) 0 I 0 Figure 3 Cost of Data Collection Using Conventional and AVL Methods --AVL Method l l ltllllllllll l t lllllll 16 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 255 270 285 300 315 330 345 360 Duration of Data Gathering Period (Days) 0 9 18 27 36 45 54 63 72 81 90 99 108 117 126 135 144 153 162 171 180 189198 207 216 Vehicle Miles of Data Gathering (Thousands of Miles)


VI. COMPARING TWO METHODS OF LOS CALCULATION The original motivation for conduct i ng the Miami Experiment was to determine the feasibi l ity of using automatic vehicle location to measure average vehicle travel speeds, in order to calculate roadway level-of-service. A key finding in this experiment was the compar i son between two methods of calculating level of-service: standard FOOT planning methodology based on traffic volume, and average travel speeds collected by the AVL system. This section of the report compares the two methods of LOS calculation for one of Miami's seventeen transportation corrid.ors: Dolphin Expressway (SR 836). The two methods of LOS ca l culation are: the Florida Department of Transportation estimation using traffic volume counts and the Generalized Tables of FOOT's 1992 Florida Highway System Plan: Level of-Service Manual, which is an adaptation of the 1985 Highway Capacity Manual (Transportat i on Research Board Special Report 209 .). average travel speeds based on the l ocation data gathered by the AVL system. CUTR research staff wrote two software programs, SPEED.EXE arid SEGMENT EXE to analyze the vehicle location data gathered by the AVL system and report average speed. CUTR used SEGMENT.EXE to ca l culate the average speeds of vehicles along a 1.69mile segment of Dolph i n Expressway between 57th Avenue and 42nd Avenue. A map of this segment is shown in Figure 4. For a more detailed description of the SEGMENT.EXE software program, see the Final Project Report on the Miami Experiment, entitled Automatic Vehicle Location for Measurement of Corridor Level-of Service: The Miami Method. Dade County defines the daily peak period as the average of the two highest peak hours For purposes of this study, CUTR classified Miami's peak period as 4:00 pm 6 : 00 pm. Since the FOOT Generalized Level-of Service Tables determine level-of-service based on peak-hour directional volumes and speeds, CUTR averaged the speeds of westbound trips (outbound from the central business district) which occurred from 4:00 27


pm t o 6:00 pm The AVL system collected data on 29 peak period peak direct i on trips taken during the 113-day data gath er ing period from April25 to August 25 1994 A list o f those trips appears i n Appendix B. Variability of Speeds Exam i ning l ist of vehicle trips used in the AVL-based LOS calcu l ation it is I nteres ting to note the h igh degree of variab i lity in veh i cle speeds from day to day Thi s high variabi l ity differs from standard determ i nist i c transportation p l ann ing wh i ch use average speed values because those models assume that the average vehicle speed in a given hour on a giv e n roadway will be the same day after day. The list of average trave l speeds In Append ix B contains many values which may look l ike "ou tl ier" po i nts, I.e values much higher or much lower than the average CUTR has r etained these values i n ca l culation of an ave r age travel speed over the 113-day test pe ri od By inc luding o u tlie r" po i nts the average travel speed v a lu e ca n bette r re p resent realworld conditions o n that roadway For example if inc i dents frequentl y occur on a roadway which cause s i gnificant delays that roadway offers inferior service than roadway s with the same uncong ested spee d which are not prone to traffic accidents. Sample level-of-Service Calculation CUTR looked to the 1985 Highway Capacity Man!ia l published by the Tra n sporta ti on Research Board as Specia l Report 209 to dete rmi ne the level-of-serv i ce f o r Do l p h in Expressway based on average vehicle trave l speeds collected by the AVL system Since Dolphin Expressway i s a freeway, the average travel speed ranges for various grade levels-of-service, for free ways of variou s trave l design speeds can be found in Tab l e 3-1 "Levels of Service for Basic Freeway Sectio ns" which is reprodu ced in this report in Tab le 3 28


*W t. -,_ ... .. ,.. \ .... Figure 4 Sample Road Segment Dolphin Expressway NW6200St NWS4thSt .I I S.at9 Hwy 112 ----t&. NW7th St \V F laglet Sl I.J r z ::! f:l i!. ?I 1-85 z ::!


Table 3 Leve l s of Service for Basic Freeway Sections LOS Density OS OS OS OS OS OS OS OS OS (PC/MIILN) 70mph 70mph 70mph 60mph 60 mph 60 mph 50 mph SO mph 50 mph Speed" vic MSP. Speed" vic MSP Speed" vic MSF {mph) (PCPHPL) {mph) (PCPHPL) (mph) (PCPHPL) A "' 12 ;, 60 0 35 700 . . . B "'20 ;, 57 0 54 1 100 0.49 1,100 . c s30 :>:54 0 77 1,550 ;, 47 0.69 1,400 :0:: 43 0 .67 1 ,300 D ,; 42 :0::46 0 .93 1,850 ;, 42 0 84 1 700 :0:: 40 0.83 1 ,600 E :!067 ;, 30 1 .00 2.000 ;,30 1 00 2 000 ;, 28 1 .00 1 ,900 F >67 <30 <30 <28 --------------- Maximum service flow rate per lane under idea l conditions Average travel speed Highly variable unstable Note : All values of MSF rounded to nearest 50 peph. Reproduced from the Highway Capacity Manual, Transportation Research Board Special Report 209, 1985, Table 3-1, page 3. 30


According to the criteria contained in tt\is table, the level-of-servioe for this section of Dolphin Expressway based on average travel speeds collected by the automatic vehicle location system is LOS F The average measured speed was 15.92 miles per hour. FOOT Planning Methodology The most recent level-of service estimation computed by the F lorida Department of Transportation District 6 for this section of Dolphin Expressway is also LOS F T he FOOT calculation uses traffic volume counts applied to the Generalized Tables of FOOTs 1992 Florida Highway System Plan: Laval-of-Service Manual Comparing the Two It i s important to note that the AVL-based measurement reflects travel tha t occurred during the 113-day data gathering period between Apr il25 and August 15, 1994; while the FOOT estimation is based on data from 1991. Moreover, the AVL-based measurement is the average of the two peak hours while the FOOT estimation is for the peak (or worst) 15-minute period Furthermore any assessment of AVL s feasibility as a means of determ i ning LOS should cons i der the fact tha t A VL is a direct measurement of travel speed wh ile the FOOT methodology is an est i mate of speed based on traffic volumes This point is especially important when calculating LOS for a future year since the ability to forecast future traffic volumes Is widely established whereas predicting future travel speeds is not. LOS Calculation on Other Roadway Segments It s houl d be noted that the AVL-based level-of-service calcu l ation shown here is not limit ed to freeways and can also be performed on segments of frequently traveled arterial streets. CUTR performed an AVL-based calculation for a 1 17-mile segment of South Dixie Highway between 27th Avenue and 17th Avenue The average travel speed 31


for the 138 peak period peak direction trips made by the volunteer drivers was 18.21 mi les per hour, corresponding to a le vel-of-serv ice value of LOS D. In addition, CUTR attempted to perform an AVL-based level-of-service for other arterial streets. However, few of the 25 volunteer drivers who participated in the Miami Experiment traveled arterial streets in the peak direction during the peak period as part of their normal daily commute. For example, a compilation of all peak period, peak direction trips recorded on a 1.42-mile segment of US-1 between NW 79th Street and NW 54th Street revealed fewer than 6 trips, clearly not enough to produce a reliable value for average travel speed. Calculating AVL-based level-of-service on infrequently traveled arterial streets highlights the need for a more systematic process for driver recruitment if FOOT were to repea t an experiment similar to the one conducted in Miami. LOS Calculation on Exceptional Roadways The AVL-based method of LOS c;alculation on South D i xie Highway is particularly i mportant because this arterial has signalization characteristics far different from average conditions in th e rest of the state. The long green time for the afternoon outbound movement ori South Dixie Highway makes this arterial behave much like a freeway making it a popular commuter route among the volunteer drivers. The Florida Highway System Plan even singles out South Dixie Highway as an example of a roadway for which "actual data and computer models [i.e. ART PLAN] should be used In l ieu of the Generalized Tab les ."2 The particular segment of South Dixie Highway analyzed in this study has three traffic signals, one on either end-point [27th Avenue and 17th Avenue] and one at its intersection Vo(ith 22th Avenue as well. As stated above, the average travel speed for the 138 peak period, peak direction trips made by the volunteer drivers during the 1994 data gathering experiment was 18 .21 miles per hour, corresponding to a level-of-service value of LOS D. The minimum speed recorded was 6.61 mph and the maximum speed recorded was 34.47 mph. Florida Department of Transportation, Florida Highway System Plan: Level of Service Manual, (Tallahassee, FL: Florida Department of Transportation, 1992), page 4-4. 32


To obtain level-of-service from the average travel speed for South Dixie Highway, CUTR looked to the Generalized Level-of-Service Tables for Urbanized Areas published by FOOT in the Florida Highway System Plan: Level-of-Service Manual which is reproduced in this report in Table 5. In contrast, the Florida Department ofTransportation level-of-service estimate taken from the Generalized Tables for this roadway segment is LOS E. This estimate is based on 1991 traffic volume count data A recent ART _PLAN analysis of this segment using 1993 traffic volume count data resulted in a congested travel speed value of 28 miles per hour, which corresponds to a leve l -o f -serv ice value of LOS B. Table 4 compares the three methods of LOS calculation on this roadway segment and the resu lt ing level-of-service. Table 4 Level-of-Service Calculation on South Dixie Highway btw. 27th Ave. and 17th Ave. Ca l culation I n put Variables Peak Period Year Output Method Data Speed Collected (3eneralized volume/capacity worst 15-minute 1991 15 mph Tables ART_PLAN volume/capacity, peak hour 1993 28 mph green time/cycle length, etc. AVL average travel worst 2-hour 1994 18 mph speed (4:00-6:00pm) Output LOS LOSE LOS B LOS D This table illustrates the high variability of output LOS given the calculation method used for LOS determination 33


L OS A B c 0 E F Freeways (v/c) :;; 0.35 .. 0 54 s 0 77 .. 0.93 < 1.00 > 1 .00 Table 5 Ge n e raliz e d Pe a k H our Directiona l Volumes for Florida s Urbanized Area s Leve l -of-Service Crite ri a Uninterrupte d State Two-State Two-Stat e Two-Stat e Two-N o n-st a te Multilane W a y Arte r ials Way Arterial s Way Arte rials Way Arterials Arterials' C lass I Class II C l aaa Ill All (v/c) (av erage (a v e rag e t ravel (av e rag e tra v e l (intersection ( average trav e l speed) speed ) speed) vic) travel speed ) NIA ;;,:; 35 mph ;, 30 mph ;;,:; 2 5 mph .. 1 .00 -.. 0 .45 ;;,:; 28 m p h :!:: 24 m ph ;;,:; 19 mph :S 1.00 -.. 0 .60 ;;,:; 22 mph ;, 18 m p h ;, 13 mph .. 1.00 -s 0 76 ;;,:; 17 mph ;,: 14 mph .. 9mph s 1.00 -.. 1 00 .. 13 mph oo10mph ;;,:;7mph .. 1 .00 > 1 .00 < 13 m p h < 10 m ph < 7 mph > 1.00 same as state arte rials Othe r N o n-St ate Signaliz e d R oadWlya ( s lopped delay) ::!0 5 sec ::!0 1 5 sec "'25 sec ::!0 40 sec ::!0 60 sec > 60 sec Excerpted from Tab l e 3 1 olthe F lorida Highway Systems Plan: Level of Service M anua l FlOr ida Dep a rtm ent o f Transportati o n 1992 page 3 -4. 34


VII. OTHER POSSIBLE APPLICATIONS OF AVL IN TRANSPORTATION OPERATIONS The Miami AVL demonstration has proven the feasibility of utilizing land-based automatic vehicle location technology to accurately and automatically obtain average travel speeds. Travel speeds along corridors, through intersections, and even over specific roadway network links represent the fundamental indicator of traffic performance. From this basic indicator many other traffic characteristics can also be discerned. If average travel speeds (and travel times) can be collected more easily and on more of a regular basis, they can assist the transportation professional in numerous ways. For example. speeds taken before and after the implementation of traffic capacity improvements can assess the true effectiveness of the improvement. Further, simultaneous speed measurements along competing travel corridors captured through A VL technology can assist in more effective traffic operations planning by providing comparative performance travel speed data. Finally, travel speed data captured in real time by AVL-equipped vehicles "floating" in the traffic stream.can be utilized to provide the motoring public with more timely congestion information in selecting time, route, and even mode of tripmaking. Technical Memorandum 1 profiles operational tests of A VL technology currently being conducted in the United States, and describes these projects' use of AVL in transportation planning applications. The only transportation planning application of A VL currently being tested is the gathering of real-time congestion levels for traveler i nformatio n and dynamic route guidance. The various types of traffic performance measuring applications for AVL technology can also include the following. Signalization Modifications Traffic signal cycle le ngths, signal phasing, and signal timing plans are often modified to reflect changes in traffic patterns and demand. However, the ability to assess the effect of signal modifications is time-consuming and thus not routinely performed and 35


documented under actua l driv ing cond i tions Instead a trained visua l observat i on concludes w h ether the s i gnalization modifications have ad equately addressed the majo r p r oblem areas Further the effects of signal progress i on improvemtlnts along a corrido r of tra ffi c s i gna l s are typically simu l a ted or estimated not discerned i n the rea l env i ronment. AVL technology could monitor total travel time through a particular travel corridor or determine changes in vehicle approach delays (or dwell time) at individual intersections as a result of signalization modifications. This application would require a high frequency of polling vehicles' locations as well as a high degree of positional accuracy on the part of the AVL system. For example, the Air T ouch Teletrac system !Jsed in the M i am i Experiment has neither the polling frequency ( every 8 seconds) o r pos i tional accurac y { within 50 feet ) for th i s application Incident Detectio n and Response Techniques Incident management teams made up of emergency response agencies and t raffic engineers continually attemp t to recreate {from memory) the chronology of events following a traffic incident. This recreation of events is done to satisfy the two primary objectives of the teams: (1) to respond to future traffic incidents in the most timely manner, and {2) to restore normal traffic operations as soon as possible Ideally it is important to be able to detect an incident before it dramatically impacts the flow of tra ffi c o r at the very least, quickly restore traffic flow to norma l operations followi n g an incid e nt. AVL-equipped veh i cles continuously repo rtin g travel speeds can be compa r ed aga in s t hi storical speed data for simi lar periods to better document deteriorating co n dit i ons As conditions reach a certain threshold in speed reduction response teams could t h en be automatically alerted In the case when an inc i dent is not caught in time AVLe q u ipp ed vehicles passing through the area will be able to record running speeds. and eventually record these speeds in real -time. This speed data profile can then be compared against response team records to assess how quickly response procedures at that time restored normal operating speeds through the incident area. Changes in response procedures can be monitored over time and overall performance judged from actual data. 36


Travel Time Contour Maps Travel time contour maps serve as an illustrative technique to display the general quality of mobility within a metropol it an area. These maps provide a single snapshot of an area in tenns of travel times by geographic location and travel corridor. The lines, or contours, indicate the distances for various travel t ime increments. Typically, these travel time contour maps are shaped like stars, with the points of the star located the farthest out along the radial freeways and expressways of the area. This pattern signifies longer distances being able to be traversed because of the provisions for greater capacity and speed along these types of facilities. Travel time contour maps are usually compared over time to illustrate the effects of increasing congestion (i.e., shorter distances traversed over similar durations of time), and to establish market areas (or "zones of influence ") for businesses, hospitals, schools, etc. AVL-equipped vehicles can provide travel time data for the major travel corridors of an area on a more regular basis to monitor levels of congestion over time. Addit i onally as this travel time data is plotted in the form of travel time contour maps specific areas or corridors with the most rapid travel time deterioration can be visually ide nt i fied and priorities for capacity improvements better rationalized The effects of capacity improvements can also be compared and illustrated (by star points that stretch out further) with previous years' corridor travel times. Detour Routing The selection of the most feasible detour route during periods of roadway reconstruction and new construction is not always a simple traffic engineering task. The ultimate selection and designat ion of the detour route should be based primarily on overall minimum travel times, thereby minimizing inconvenience to drivers. Most often, however, detours choices are based on minimum distance, not minimum travel times. AVL technology can be utilized to analyze and compare alternative detour routes for best selection based on tot al travel times. If variable message signs can be temporarily located at strategic points along the roadside, motorists can be re-directed to alternate routes as needed based on real-time conditions. With AVL technology, static detour 37 I


routes within already congested urban areas can give way to more dynamic routing reflecting changes in congestion levels that typically occur during different times of the day. Adjacent roadway facility speeds supplied by AVL-equipped vehicles in the traffic stream integrated with variable message signs can supply the needed input to perform dynamic detour routing. Speed Adjustment Factors Similar to traffic counts that can be COI")Verted to other periods based on adjustment factors, travel speeds can likewise be adjusted to the desired period. It is well known that traffic counts that are needed for analysis can not always be collected at the desired hour of the day, day of the week, or week of the year. Therefore, based on data gathered at continuous count stations on similar or adjacent roadway facilities, adjustment factors are determined and applied to obtain the desired results. For example, based on a large historical traffic data base, a traffic count taken on the first Tuesday in August can be converted to a traffic count estimate for the second Tuesday i n January. A VL-equipped vehicles in the traffic stream that continuously collect travel speeds simultaneously on various roadways at different times of the year can provide the needed adjustment factors, or ratios, for conversion to the desired travel speeds. For example, modeling peak vs. off-peak or weekday vs. weekend conditions can better be simulated with adjusted facility speeds that represent these particular periods. Trends in historic speed data, by facility, can also be more accurately determined for modeling purposes. Link Speed Calibration for FSUTMS The Florida Standard Urban Transportation Model Structure (FSUTMS) has been developed to provide transportation planners with a well-tested easy-to-use transportation planning tool. However, many times ind ividua l link or roadway segment speeds used in FSUTMS are assumed to represent the average speed for a particular facility type or peak period, and may not reflect actual conditions being experienced. A VL technology not only has the capability to capture total travel time speeds, but can 38


also assist in obtaining i ndividua l link speeds. AVL provides an accurate and re l iable way of collecting average l ink speeds for various analys i s periods i e pea k off-peak mid -day, weekend, etc Time-distance stamps can be recorded automatically at predetermined node or intersection locations all with known longitude and latitude coordinates This capability of AVL techno l ogy can more representatively permit calibrat i on and validation of congested and free-flow link speeds for FSUTMS analysis AVL technology can a lso assist in identifying bottleneck a r eas in the roadway network, in order to more efficiently prioritize locations for capacity improvements Dynamic Trave l e r I nformat i o n All AVL-equipped vehicles can serve as probe vehicles to provide and ver ify real-time travel speeds and leve l s of congestion Thi s type of information can be captured in real time automatically transmitted to traffic management centers, and ultimately provided to travelers. Time l y i nformation regard ing congestion ca n be provided to travelers via roadside variable message signs and radio reports en route ; and updated radio te l evision, videotex or audiotex reports prior to tripmaking It has previously been shown that accurate and timely inf ormation provided t o t r avelers i n a desi r ab l e fash ion can affect tripmak ing habits by a l tering route se l ection depart ure ti me, and even mode of travel. For example, a survey of calle r s t o the Smar T ravele r telephone i nformation system in Boston found that 30% of its users "frequent l y changed their time route or mode of travel based on the r ea l -time traffic i nformation given out by the telepho n e serv ice; 96% of callers change the time r oute or mode of trave l occasionally, based on the SmarTraveler traffic information. (SmarT r aveler does not gather the bulk of its traffic information via AVL, but by surveillance cameras and other sources.) The end result of th i s AV L technology capability can assist agenc i es in trave l demand management and congestion management. 39


Goods Movement The preservation of efficient goods movement through or around a metropolitan area is a primary measure of mobility and economic vitality. Effective management of commercial fleet vehicle operations can reduce or eliminate delays in transport and ultimately minimize the cost of goods passed on to the consumer. AVL technology has already been utilized by many commercial vehicle operators, particularly time-sensitive delivery carriers, to mon itor fleet and individual package movements in real-time and redirect drivers as needed to avoid congestion and keep on schedule. AVL-equipped commercial vehicles themselves can also serve as barometers of mobility by recor ding the time it takes to traverse a metropolitan area during peak and off-peak periods, on particular routes and during particular periods. This information can be shared with, or even sold to, other commercial carriers. Additionally, the continuous monitoring of hazardous materials and hazardous waste is best served by AVL techno logy especially when an incident occurs invo lving a truck carrying hazardous materials. Knowing where ihe vehicle is at all times and what mateiials are being transported (through an automatic check of the driver's manifest) provides for the most efficient goods movement under these type of conditions. Operational Changes in Transit Many public transit vehicle operators utilize AVL technology, or point-to-point vehicle tracking (automatic vehicle identi fica tion) technology for fleet management and on-time performance monitoring. Transit agencies in Tampa, Miami, Fort Lauderdale and Palm Beach all either have AVL systems currently operating or have plans to procure and install an AVL system for their bus fleet. Operational changes are often made to imp rove the quality and reliability of transit service. These operational changes can include bus stop re-location/consolidation, re-routing, bus pull-outs, signal pre emption high-occupancy vehicle lanes, etc. Through AVL-equipped transit vehicles improvemen ts in travel speeds, or headways, can be monitored in real -time under actual operating conditions to assess the "before" and "after" impact of the various types of transit operational improve ments previously mentioned. 40


At the same time, transit vehicles are being monitored for imp roved service, non-transit vehicles equipped with AVL technology can be monitored for travel time improvements as a result of the tram;it operational improvements. For example, the construction of bus pull-outs and a consolidation of bus slops along a congested arterial roadway will certainly improve the flow of non-transit vehicles in the traffic stream by reducing vehicular conflicts. AVL has the capability to monitor and record the magnitude of non transit traffic flow improvement. Special-Event Traffic Handling The arriving and departing traffic patterns along corridors serving major sports arenas, concert halls, outdoor theaters, and other special event venues can be recorded in real time by AVL-equipped vehicles. Other one-time or annual events, such as sidewalk art festivals, fireworks displays, parades, etc., can also cause significant traffic congestion. Historical speed data profiles, gathered by AVL-equipped vehicles, can be analyzed to determine the best temporary facility operational improvements: signal timing modification, parking restrictions, turning restrictions and conversion of two-way streets to one-way travel. Full or partial automation of this procedure may also reduce or eliminate the need for traffic control personnel and manual override of the traffic control system. Impact of Weather on Travel Speeds Traffic congestion increases dramatically during periods of inclement weather because (1) people generally tend to drive more slowly and (2) the increased likelihood of accidents. Compilation of historical speed data by AVL technology, by facility, during periods of inclement weather can determine spot locations where variab l e speed limit signs or other temporary facil ity improvements may be necessary to prevent incidents. Safe Travel Speed Through Work Zones Excessive speeds (or inadequ ately posted speed limit signs) through construction work zones are the cause of many fatalities each year across the nation's highways. Particularly on long, continuous segments of reduced speed, drivers can become 41


impatient and attempt dangerous passing maneuvers. AVL-equipped vehicles passing through the "work zones" can monitor travel speeds to determine extent of violation and the need for imp roved enforcement. "Mayday" Alerts Along Isolated and Rural Areas A VL technology, because of its ability to continuously track vehicles, can be used as a distress call when destinations are not reac hed in the normal time expectations Similar to the flight plans pilots submit to the air traffic controller, known routes and expected travel times can predict likely arrival times with AVL technology. In fact, many AVL vendors include the added feature o f a "mayday" button which immediately i nd icat es vehicle location and signals a distress warning to the dispatcher or vehicle tracking center attendant. In September 1994, AirTouch Teletrac announced that it will be offering such a "mayday" AVL-system in conjunction with Avis Rental Car company in Miami. The system is specifically mar keted to address fears of anti-tourist crime. Even when a panic button is not included, AVL can monitor travel progression against expected travel times. For example, in south Florida, Alligator Alley is a 75-mile stretch of rural Interstate 75 through the Everglades National Park, its primary exits being a toll booth on each end. No significant intersections or interchanges exist along this stretch, so travel time f rom one end to the other can be accurately predicted. An AVL-equipped vehicle traveling along a remote corridor such as this can be continuously tracked, and s i gnificant delays in completing the journey can trigger an alert to the proper authorities. Summary The application of AVL technology, and innovative utilization of data obtained with AVL technology can significantly reduce traffic congestion, improve mobility and even enhance traveler safety. AVL technology allows for easy, cost-effective, and accurate performance monitoring of the transportation system ; and the measurement of traffic flow improvements following the implementation of capacity improvements. Real-time travel speed data collected through AVL technology can better assist the transportation professional in meeting the challenges of more efficient and effective transportation planning. 42


V III. STA T EWIDE AVL F EAS I BILITY The specific type of A VL system used in the Miami Experiment ground-based radio navigation is currently only available in six U .S. metropolitan areas: Los Angeles Chicago, Detroit, Dallas/Ft. Worth, Houston, and the greater Miami area In south Florida A i rTouch Teletrac's current coverage area includes only Dade, Broward and Palm Beach Counties. (Ai rTouch Teletrac has p l ans to expand coverage to Orlando but to no other areas in the state ) I f FOOT wanted to i mplement an AVL based system for gathering average veh i cle travel speeds in other pa rts of the state, another type of AVL positioning technology would have to be used until tri-lateration became available. This section conta i ns an assessment of the alternative types of AVL positioning technologies, weighing such factors as coverage, cost and positional accuracy. This assessment is summarized in Table 6. The purpose of this section Is to assist FOOT i n determ i n i ng the most appropriate technology for conducting the n ext phase of the Miami Experiment in other parts of the state Dead Reckoni n g and Map-Matching Dead-reckoning systems monitor a vehicle's internal compass and odometer and ca l cu l ate its position by measur ing its distance and direction from a central starting point w h ose posit i on is already known Because of the low accuracy of vehicle odometers dead-reckoning systems frequent l y get off t r ack, and are often corrected using a technique called map-matching Map-matching systems store a map of the vehicle s coverage area in a database and assume that when a vehic l e changes direction it must have turned from one road to another When a vehicle makes a turn, map-matching systems alter the vehicle s recorded location to the nearest possible po i nt at which the turn could have taken place. Because of the low degree of positional accuracy of both dead-reckoning plus map matching and dead-reckoning alone, most AVL systems use more advanced technology options. Signpost When vehic l es regularly travel a fixed route, such as transit buses, many fleet operators have found that signpost based positioning systems offer an affordable alternative to 43


more advanced AVL technologies Antennas are placed at locations throughout the vehicle's known route and record the time when the vehicle passes. A signpost-based AVL system can also be a valuable by-product of systems intended for other purposes. Electronic toll collection systems in Houston, Chicago and the New York metropolitan area use the automatic veh icle identification (AVI) tags to track vehicle speeds and measure congestion. Drawbacks of signposts-based systems inc lude their inability to track vehicles off their normal route, and the signposts' po tential to be subject to vandalism. Of the four transit agencies In F lorida Cijrrently using or planning to use A VL to track their bus fleet, three agencies [Hillsborough Regional Transit Authority in Tampa; Broward County Trans it in Ft. Lauderdale; and Jacksonville Transportation Authority in Jacksonville) use a signpost-based system available from Motorola. (The fourth, Metro Dade Transit Agency in Miami, is currently in the process of procuring a GPS-based system from Harris Corporation.) Hillsborough Area Regional Transit Agency (HARTline) has offered to make its bus locatio n data available to the Hillsborough County Congestion Management System, in order to measure levels of congestion as one of the CMS performance measures. One drawback to this approach is that HARTline buses can only be tracked on their predefined bus routes. In addition, because buses must stop freq uently to take on and let off passengers, they are difficult to use as probe vehicles to measure average vehicle travel speeds. Ground-Based Radio-Navigation In '1er restrial" or "ground-based" rad io-navigation, the AVL vendor sets up several receiving anten nas in a metropo l itan area. Each equipped vehicle broadcasts a radio frequency signal (usually in the 902-928 MHz range) to all nearby receiving antennas. From the time its takes for the signal to travel to the antenna, the distance of the vehicle to the antennas can be determ ined. If the vehicle's signal was received by three or more antennas the vehicle's position can be uniquely determined (i.e., multi-lateration) The AJrTouch Teletrac system used in the Miami Experiment employs ground based radio-navigation. The cost for such a system (without cost-sharing) is a start-up cost of $300 per vehicle, plus a monthly service fee of $10 per vehicle. F leet management 44


expenses, such as a computer workstation running the Flee!Directorno software are included in the per-ve hicle start-up fee. A irTouch Teletrac guarantees its system to be accurate to within 150 feet. AirTouch Teletrac staff state that, because of South Florida's relatively flat terrain, the system is accurate to within 50 feet LORAN-C LORAN C { Long-Ra nge Aid to Navigation ) uses low frequency rad i o waves to provide signa coverage The federa l government set up the communication system to a id the U.S Geological Survey in mapping. Instead of using multiple receivers to locate a signal transmitter, LORAN uses a sing le receiver to locate multiple transmitters LORA N-C often experiences radio frequency and electromagnetic interference Close proximity to overhead power lines and RF signal boosters stations in urban and industrial areas can cause significant error on the time difference calculations. I n addition, LORAN-C position systems can expe rience error due to poor signal reception in urban canyons Due to these drawbacks and unce rtainty about the governmenfs Mure plans for this system, a decrease in the number of commercial AVL systems us ing LORAN-C has occurred. Global Positioning Systems {GPS) Global Positioning Systems (GPS) use a network of 24 satellites in a geosynchronous orbit with the Earth. Antennas capable of receiving these satellite signals can determine the ir own location. GPS a n tennas receive the satelli t e signals free of charge ; however a license must be obtained The U.S. Department of Defense l aunched the sate ll ites in order to track objects of I nterest on the ground. The system was used to track tanks and even individual soldiers during the Persian Gulf War. GPS-based AVL systems are used in the federally-funded TravTe k and ADVANC E traffic management/traveler informat ion projects. CUTR obtained price and accuracy information from three GPS-based AVL system vendors. AutoTrac' s AVL system costs a one-time activa tio n fee of $7,000 per vehicle and is accurate t o wit hin 50 feet Highway Master's system costs a onetime start-up fee of $2,000 per vehicle and is accurate to within 30 feet Qua i Comm's system costs a one-time start-up fee of $4,500 per vehicle plus a monthly service charge of $170 per 45


vehicle per month. The QuaiComm system is only accurate to within 500 feet. (Qualcomm offsets the low positional accuracy with extensive fleet management software support.) For all three systems, fleet management expenses are included in the per vehicle start-up fee. Differential GPS (DGPS) One potential problem with using GPS for automatic vehicle location is that the Department of Defense intentionally degrades the accuracy of GPS positioning data used for non-defense purposes. Several companies have addressed this need for additional. accuracy by manufacturing systems that broadcast these correct ions to special receivers ("differential GPS receivers") using a variety of wireless transmission media. CUTR obta i ned price and accuracy information from two DGPS system vendors. A "premium" AVL system available from Differential Corrections, Inc: (DCI) costs $600 for one year's service, plus a one-time activation fee of $615 per vehicle. DCI's system is accurate to within 1 meter (3.3 feet). The Acc-Q-Point system costs $ 1,200 per receiver for one year's service, plus a one-time activation fee of $519 per vehicle. The Ace-a Point system is accurate to within 1 meter (3.3 feet). F or all three systems, fleet management expenses are included in the per-vehicle start-up fee. Cellular Phones A few AVL systems use the still-experimental positioning technology based on cellular phones. These systems use the passive RF signals em itted from cellular phones to detect the location of a vehicle, using special-purpose receiving antennas. The cellular phone technology is being tested in the CAPITAL project in the Washington, D.C. metropolitan area. Engineering Research Associates will use the system to track Washington, D.C.'s estimated 117,000 cellular telephone users during the 90-day test. The total project cost is $7.1 million. The positional accuracy of this type of AVL technology is currently not known, and is one factor which w ill be measured during the CAPITAL project 46


Table 6 AVL Positioning Technologies Technology Coverage Capital Cost Operating Cost Accuracy (per vehicle) (per vehicle per year) Dead-reckon i ng + Local Low Low Poor map-matching S ig npost Predefined High High Good routes Ground-based radioPredefined Low : Low: Good : naviga tion: metro Ai r Touch Te l etrac areas $300 $120 150ft (50ft i n SE Fla) LORAN-C North Med ium Medium Good America GPS: Globa l Very High: Low: Very Good : Auto-Trac $7000 N/A 50ft H ig hway Master $2000 NIA 30ft QuaiComm $4500 $170 500ft DGPS: G l oba l Medium: Med ium to H i gh: Excellent: DCI $615 $600 3 3 ft Acc.Q Po int $519 $1200 3.3 ft Source: Metro Magazine, May/June 1994 p. 44; and AVL vendors. Vehicle Location Data Collected by Fleet Operators In addition to assessing the various AVL positioning technologies CUTR also investigated the availability of data being collected by local and national fleet operators currently using AVL systems CUTR contacted AVL system vendors, ask i ng for names and contact information of their customers who travel through Florida QuaiComm, Inc., a manufacturer of GPS-based AVL systems gave the names of three of its Florida customers: Terr i Dicks Trucking of North Central Florida Arrnellenian T r ucking of Southeast Florida, and Commercial Carriers of Lakeland. All the fleet operators contacted said they were willing to share their vehicle location data with FOOT, given assurances that their drivers' privacy would be respected. 47


If FOOT were to use vehicle lo cation from Florida-based trucking companies, several details regarding the AVL system these companies use would have to be specified: How many equipped vehicles w ill be providing location data? What are the routes regula rly traveled by these vehicles? What i s the accuracy of the AVL system used? How often does the A VL system poll vehicles' for their location? What is the format of veh icle location data? How would vehicle location data be converted to average travel speeds by vehicle trip? By link or road segment? These questions should be answered in subsequent phases of this research Summary If Florida Department of Transportation were to conduct another technology evaluation similar to the Miami Experiment outside of Dade, Broward and Palm Beach counties, an AVL system based on global positioning systems (either with or without differential correction) offers the technology option which best suits FOOT's needs. The global coverage of GPS and DGPS enables FOOT to track a vehicle on any route. FOOT would need a technology which offered pos"ional accuracy at least as good as that used in the Miami Experiment, i.e., to within 50 feet. The accuracy of GPS and especially DGPS technologies enables FOOT to measure average travel speeds. Finally, GPS and DGPS automatic vehicle location system are available from multiple vendors, enabling FOOT to compare multiple bids. 48


IX. CONCLUSION Five federallyand regionally-funded projects in the United States are using vehicles as probes to collect information on traffic conditions. In each of these fiVe projects, the data is used to deliver real-time traffic information to various users: state department of transportation, incident manag ement and emergency response teams, radio and 1V traffic information broadcasters, and even drivers with dynamic route guidance units installed in their vehicles. Only the Houston AVI projects has considered the possibility of using the data collected to measure level-of-service The Miami Experiment is the only IVHS project in the United States to use data gathered from an AVL system for transportation planning purposes. AVL systems can be expensive than conventional methods of collecting travel in formation for long-term data gathering periods and more vehicle miles of travel. The Miami Experiment showed that automatic vehicle location is a viable method for measuring average vehicle speeds on Miami's roadways. T h is study showed that vehicle speeds measured by an AVL system can be used to calculate roadway level-of service, based on the LOS criteria out lined in the 1985 Highway Capacity Manual. On the sample roadway analyzed in this study, the official FOOT level -of-s ervi ce value is LOS F. Similarly, the level-of-service measurement based on average travel speeds collected by the AVL is also L OS F. These LOS values are based on measuring travel speed characteristics for different time periods. The AVL-based method reflect travel which occurred during the 113-day data gathering period between April 25 and August 15, 1994 ; whereas the official FOOT level-of-se rvice value is based on traffic vo lume data taken in 1991 Another significant finding in this research was the high day-to-day variability of average travel speeds during the same hour of the day on the same roadway segment. The vehicle locat ion and travel speed information collected by AVL systems is useful in several other transportation planning applications other than leve l-of-serv ice determination. However, only the transportation planning application of AVL currently being tested in other areas of the United States is the gathering of real-time congestion levels for traveler information and dynamic route guidance 49


F i nally, the satellite-based technology global positioning systems (GPS), whether augmented with accuracy enhancing differential GPS rece ive rs or not offers the best coverage statewide. Using vehicle location data gathered by Florida trucking companies, equipped w ith AVL transponders which use GPS positioning technology, appears to be a viable opt i on for gathering average trave l speeds i n the state's urban areas Severa l factors regard i ng these compan i es and their use of AVL need to be i nvestigated in subsequent phases of this research. 50


BIBLIOGRAPHY "Amtech to Install AVI for Surveillance on H oUston Freeways." Inside IVHS. (November 22, 1993): 6-7. AutoTrac brochure. Ba r rett, Lee. "How to Chose AVL Systems for Transit Bus Rail and Moto r Coach Metro Magazine. (May/June 1994) : 44-51. Blumentritt, Charles, Kevin N. Balke, P.E. and Edward J. Seymour, P.E., Ph .D. "TravTek System Architecture Evaluation: Preliminary Analysis of Selected Elements." Proceedings of the IEEE-lEE Vehicle Navigation and Information Syst ems Conference (VNIS '93). Ottawa, ON, October 12-15, 1993 Center for Urban Transportation Research Automatic Vehicle Location for Measurement of Corridor Level-of-Service: The Miami Method Tampa, FL: Center for Urban T ransportation Research. September 1994. "Driver Recruitment for the ADVANCE Demonstration Project." ADVANCE News (January 1994): 3 E ng i neering Research Associates brochure "ETIM Reviewed on Capitol Hill. ETTM Forum. (August 1994): 3 5. Florida Department of Transportation Florida Highway System Plan: Level of Service Manual. Tallahassee, FL: Florida Department of Transportation. May 4, 1992. G i bbons Glen, et. al. "Automat i c Vehicle L ocation: GPS Meets IVHS ." GPS World (April 1993) : 2226. "Hertz to Offer Navigation systems on Three Ford Models i n Eight Rental Markets." Inside /VHS (October 24, 1994): 1-6. Highway Master brochure. Illinois Department of Transportation and Federal Highway Administration. Advance News September 1992. "Illinois Tollway to Extend 1 -Pass wit h 1 -Pass with AT/Comm Tech." Inside IVHS (June 6, 1994): 1-2. 51


IVHS America Advanced Publ i c Transportation Systems Vendor Catalogue. (Wash i ngton D C. : IVHS America): 1992 "IVH S America's Board Approves Name Change." Inside /VHS (August 15 1994) : 1-3 JHK & Associates for the Hillsborough County Metropolitan Planning Organization's Congestion Management Systems (CMS) Steering Committee. Selection of Performance Measures. Orlando, FL: JHK & Associates. July 21, 1994 Levine, Steve Z. and Will iam McCasland "Monitoring Freeway Traffic Cond i tions with Automatic Vehicle Identification Systems ." ITE Journal (March 1994): 23-28. "Mark IV IVHS Division Wins lAG E ZPass Prize Inside IVHS. (March 28, 1994) : 1-4. "Oldsmobile Taps Zexel for Navigation System Inside IVHS. (January 17, 1994): 1 3. Pietrzyk Michael C., P.E. and Edward Mierzejewski, P .E. Electronic Toll and Traffic Management (ETTm) Systems: A Synthesis of Highway Practice. Washington, D.C.: Transportation Research Board 1993. QuaiComm brochure Shrestha, Ramesh. Fundamentals of GPS Technology and Its Techniques. Gainesville, FL: University of Florida. 1994. The SmarTraveler Operational Test: Early Findings Report. Cambridge MA: Smart Route Systems. May 4, 1993 Sm i th, Clark B., et. al. "A conceptual Overview of ADVANCE." Proceedings of the IVHS America Second Annual Meeting. Newport Beach, CA. May 17-20, 1994. "Test Taps Cellular Network To Measure Traffic Flow." Inside IVHS. (November 22 1993): 1-3. Transportation Research Board. Highway Capacity Manual: Special Report 209. Washington, D.C. : Transportation Research Board 1985. TravTek brochure Trimble Navigation brochure. Turner, Clark P., AICP. Transportation Corridors: Meeting the Challenge of Growth Management in Miami Miami, FL: City of Miami Department of Planning, Building and Zon i ng August 1989 52


U S Department of Transportation. Advanced Public Transportation Systems: Slate of the Art Update '94. Washington D C.: U S Department of Transportation, January 1994. U.S Department of Transportation. Department of Transportation's Intelligent Vehicle Highway Systems Projects. Washington, D.C.: U S Department of Transportation, March 1994. II Morrow brochure. 53


APPENDIX A List of Contacts IVHS Project Contacts TravTek City of Orlando 400 S Orange Ave., 8th Floor Orlando FL 32801 Tel : (407) 246-3255 Contact: Harry Campbell City Transportation Engineer ADVANCE Illino is Department of Transportation(! DOT) 120 West Chester Ct Schaumburg, IL 60195-3161 Tel: (708) 705-4800 Fax: (708) 705-4803 Contact : Joseph Ligas IVHS Program Manager TRANSCOM T ransportat ion Operations Coordinating Committee 111 Pivonia Ave, 6th FloorNewport Square Jersey City, NJ 0731 0 Tel: (201) 963-4033 Fax: (201) 963-7488 Contact: Pete Dwier Manager of Operations Center Houston AVI Texas Department of T ransportation 7721 Washington Ave Houston, TX 77251 Tel: (713) 956-4013 Fax : (713) 956-2784 Contact: Mark Conway Senior T raffle Management Engineer 54 Illinois Tollway One Authority Drive Downers Grove, IL 60515 Tel: (708) 241-6800 Fax: (708) 241-6109 Illinois State Toll Highway Authority Contact: Nick Demaris CAPITAL Maryland State Highway Administration 7941 Connelley Drive Hanover, MD 21076 Tel: (410) 787-5884 Contact: Glen Mclaughlin AVL Vendors Acc-Q-Point 2737 Campus Drive Irvine CA 92715 2925 California Street Torrance, CA 90503 Tel: (310) 618-7076 Fax: (310) 618-7001 Contact: Michael Dyment, Director of DGPS Communications AirTouch Teletrac 3330 N.W. 53th Street, Su it e 302 Ft. Lauderdale, FL 33309 Tel: (305) 484-1300 ext. 412 Fax: (305) 486-2799 Contact: Stephen Tine, Manager of Commercial Sales AutoTrac, Inc 9330 LBJ Freeway, Suite 900 Dallas, TX 75243 Tel: (214) 480-8145 Fax: (214) 907-2292 Contact: Pat Friend Director of Sales


Differential Corrections, Inc. 20045 Stevens Creek Blvd Cupertino, CA 95014 Tel: (408) 446-8350 Fax: (408) 446-8383 Contact: Bruce Noel, Product Manager Highway Master 16479 Dallas Parkway, Suite 780 Dallas, TX 75248 Tel: (214) 732-2500 Fax: (214) 250-0182 Contact: Ken Mitchell, Executive Vice President of Sales and Marketing QuaiComm, Inc. 10555 Sorrento Valley Blvd. San Diego, CA 92121-1617 Tel: (619) 587-1121 Fax: (619) 587-8276 Contact: Dan Hooper, Atlanta Office Trimble Navigation 645 North Mary Avenue P.O Box 3642 Sunnyvale, CA 94088-3642 Tel: (408) 481-8000 Fa x : (408) 730-2997 Contact: Charlie Vlcek, Eastern U.S. Sales Representative II Morrow (pronounced "Two Morrow") United Parcel Service of America, Inc P.O. Box 14135 2777 19th Street Salem, OR 97309 Tel: (503) 391-3684 Fax: (503) 581-7205 Contact: Karl Po ley Marketing Administration Manager 55 Florida Trucking Companies Terri Dicks Trucking, Inc. Route 3, Box Number 96 Lake City, FL 32025 Tel: (904) 752-1093 Armellenian Trucking, Inc. 3446 Southwest Arrnelleni Ave. Palm City, FL 34990 Tel: (407) 287-0575 Contact: Gary Sherak Commercial Carriers, Inc. 502 East Bridgers Ave. P.O. Box 678 Auburndale, FL 33823 Tel: (813) 967-1101 Contact: Micky Foutz Dade County Level-of-Service Analysis City of Miami Department of Planning, Building and Zoning 275 N.W. Second Street Miami, FL 33128 Tel: (305) 579-6086 Fax: (305) 358-1452 Contact: Clark Turner AICP, Transportation Planner Florida Department of Transportation District 6 602 South Miami Avenue Miami, FL 33130 Tel: (305) 377-5910 Fax: (305) 377-5967 Contact: David Henderson, Transportation Planner


Appendix B Vehicle Trips Used in LOS Calculation Segment Direction Day.ot.Week Date Vehicle Speed Dol phin West 04:26:03pm Fri 07/29/94 5076 20.01 Do l phin West 04:27:02pm Fri 08/05/94 1342 19.48 Dolphin West 04:28:13pm Tue 06/28/94 5076 20.97 Dolphin West 04:36:58pm Tue 08/09/94 5076 14.88 Do l phin West 04:37:57pm Man 07/11/94 5399 17.67 Do l phin West 04:40:27pm Mon 07/11/94 5399 33.46 Do l phin West 04:51 :33pm Wed 08/10/94 9071 13.24 Dolphin West 04:56:59pm Thu 06/02/94 6403 16.57 Dolphin West 05:01 :32pm Fri 08/05/94 5892 14.03 Dolphin West 05:03:29pm Man 08/01/94 9071 7. 1 2 Dolphin West 05:04:02pm Man 0 8/15/94 5076 13 .29 Dolphin West 05:08:55pm Thu 07/21/94 5076 11.55 Dolphin West 05: 1 0:03pm Fr i 08/05/94 5076 7.37 Dolphin West 05: 14:55pm Thu 07/21/94 5076 11.41 Dolphin West 05: 15:00pm Fri 08/05/94 5076 18.69 Dolphin West 05:26:29pm Mon 06/27/94 5016 14.27 Dolphin West 05:27:57pm Fri 06/03/94 5076 7.12 Dolphin West 05: 2 9:30pm Wed 08/17/94 5076 14.56 Dolphin West 05:31 :53pm Mon 07/04/94 4184 44.17 Do l phin West 05:31 : 58pm Man 08/15/94 1 342 12.73 Dolphin West 05:32 : 28pm Fri 06/03/94 5076 16.84 Do l phin West 05:32:39pm Fri 06/10/94 5076 7.06 Do l phin West 05:35:57pm Thu 06/02/94 5076 15.84 Dolphin West 05:36:36pm Ma n 07/25/94 5076 18 .45 Dolphin West 05:38: 1 3pm Fri 06/10/94 5076 12.05 Dol phin West 05:4B:OOpm Fr i 07/29/94 1389 8 77 Dolphi n West 05:50:43pm Fri 07/08/94 1778 16.41 Dolphin West 05:52:00pm Fri 07/29/94 1389 19.10 Dolphin West 05:53:44pm Wed 06/15/94 5076 14.70 56