Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5The findings presented in the study were based on a literature review [conducted to gather as much information as possible about Adaptive Traffic Control Systems (ATCS) operations and deployments from previous studies] and two electronic surveys: a shorter e-mail survey for vendors or developers of 10 major ATCSs and a website-based questionnaire for agen- cies that deploy ATCSs. The survey was originally distributed to 42 agencies that run ATCSs in North America (United States and Canada) and several dozen locations around the world. Numerous follow-up requests were made, by e-mail and telephone, to remind agencies that had not yet responded asking them to participate. Responses were obtained from 34 of the 42 agencies in North America using an ATCS, an 81% response rate. Also, 11 responses were received from agencies in other countries. Of the North American agencies, 42% were municipal entities, 20% were counties, 13% were state agencies, and 25% were other types of entities. This chap- ter will define and introduce ATCSs. A short history of ATCSs and their classifications will be provided. Major ATCSs in use throughout the world will be identified along with the agencies that operate these systems. BACKGROUND An ATCS adjusts, in real time, signal timing plans based on the current traffic conditions, demand, and system capacity. An ATCS is defined broadly, in the previous sentence, so as to include all major ATCSs that may vary significantly in their levels of responsiveness, algorithmic framework, and detection. However, ATCSs, as defined in this report, exclude any traffic-responsive pattern selection and purely actuated (free or coordinated) types of traffic control. An ATCS usually includes algorithms that adjust a signalâs split, offset, phase length, and phase sequences to minimize delays and reduce the number of stops. The system requires extensive surveil- lance, historically in the form of pavement loop detectors, and a communications infrastructure that allows for commu- nication with the central and/or local controllers. Emergence of ATCSs during the 1970s and early 1980s was largely attributable to a failure of traffic-responsive pattern- selection systems to efficiently respond to changes in traf- fic demand. In the early 1970s, there were a few attempts to develop ATCSs; however, there was no success with these early trials (Holroyd and Hillier 1971). At that time, traffic sig- nal practitioners believed that fluctuations in traffic flows could be addressed through the development of several timing plans covering various traffic-demand scenarios and a good selection process triggering replacement of these timing plans. However, several experiments around the globe, of which the most prominent was done by the FHWA in Washington, D.C., showed that traffic-responsive pattern-matching systems have serious operational problems (Fehon 2005). The experi- ments showed that traffic control based on traffic-responsive pattern selection is not efficient. By the time one pattern transi- tions to another, traffic demand may change and the newly introduced pattern may no longer reflect current traffic condi- tions. Furthermore, transitions themselves may represent a dis- ruption to traffic. The increasing frequency of pattern changes may improve matching between signal timings and traffic con- ditions, but the system may spend most of the time in transi- tioning, which may cause a continuous disruption to traffic. To solve this problem, traffic engineers in Australia and the United Kingdom responded by investigating adaptive control of signal timings, which resulted in the development of the two most widely used ATCSs: the Sydney Coordinated Adap- tive Traffic System (SCATS) (Lowrie 1982) and the Split Cycle Offset Optimization Technique (SCOOT) (Hunt et al. 1981). Development of these systems was quickly followed by a series of other new ATCSs. However, some of these new ATCSs abandoned conventional signal timing structures con- strained by cycle lengths and offsets and, instead, offered new approaches that were mostly based on various techniques of mathematical programming: OPAC (Optimization Policies for Adaptive Control) (Gartner 1982) and PRODYN (Program- ming Dynamic) (Henry 1983). At that time, OPAC, PRODYN, and SPOT (System for Priority and Optimisation of Traffic) (Donati et al. 1984) were largely concerned with the opera- tions of single intersections. Soon thereafter, UTOPIA (Urban Traffic Optimisation by Integrated Automation) was com- bined with SPOT to account to changes at the network level (Mauro and DiTaranto 1990). Although most of these developments were taking place in Europe, at approximately the same time the FHWA initialized development and deployment of ATCSs in the United States. The Adaptive Control Software (ACS) program included a research project called Real-Time Traffic-Adaptive Signal Control Systems (RT-TRACS) that had gone through several stages to the point where there were several adaptive systems on trial in U.S. cities (Fehon 2005). Although the program initially sponsored development of five prototype strategies, CHAPTER ONE INTRODUCTION
only two of those were successfully tested and implemented in the field: a modified version of OPAC and RHODES (Real- Time Hierarchical Optimized Distributed and Effective Sys- tem) (Head et al. 1992; Mirchandani and Head 2001). It is interesting to note that one of the first fully operating North American ATCSs, the Los Angeles Department of Trans- portation (LA DOT) ATCS, was not part of the RT-TRACS project, but was developed independently in 2003 by the city of Los Angeles. Although early tests showed significant benefits of deploy- ing OPAC and RHODES over fixed-time and actuated-traffic control, these two systems were not widely accepted in the United States. It appears that the major reasons for the lim- ited deployments of these systems (as well as of all other major ATCSs) are the complexity of their logics, extensive detection requirements, necessary hardware upgrades, and the need to acquire new knowledge; in shortâincreased costs of operations and maintenance. To respond to these issues the FHWA launched the development of another ATCS whose major role was to be more simplistic, user-friendly, compat- ible with existing infrastructure (detection and hardware), and, overall less expensive to operate and maintain. The system is called ACS Lite and, although it has been tested in the field at four locations throughout the United States, is undergoing further enhancement (Shelby et al. 2008). During this same period, SCOOT and SCATS were going through their own challenges with installations in the United States. Although their deployments in Europe, Australia, and Asia have steadily increased over the years they have strug- gled to increase their deployments in the United States. It appears that the major problems of early SCOOT and SCATS deployments in the United States were related to hardware and software, which were not fully customized for the U.S. market. Early problems with National Electrical Manufac- turers Association (NEMA)-incompatible traffic controller hardware caused problems with some SCATS deployments in spite of the evident operational benefits. Some SCOOT deploy- ments faced similar problems: suboptimal (for SCOOT oper- ations) detection placements and somewhat user-unfriendly Open Virtual Memory System (VMS) interface negatively impacted SCOOT operations at some deployment sites. Across the ocean, continental Europe struggled for a long time to keep up with the development of ATCSs in the United Kingdom, Australia, and America. French systems, such as CRONOS (Boillot et al. 1992) and PRODYN, which were early ATCS leaders in continental Europe, were not widely deployed in France or elsewhere. UTOPIA/SPOT appeared to work well in the networks of Italian cities for many public transit operations, but the first SPOT deployment abroad, in an environment with mostly vehicular operations, was not successful (Pesti et al. 1999). Development and application of German ATCSs, where SITRAFFIC MOTION (Kruse 1998) and BALANCE (Friedrich et al. 1995) represent the most notable systems, suffered for years before conditions were met 6 for their more extensive deployments. German systems were facing a series of local institutional barriers and were seen by the professional community, for a long time, only as scien- tific research tools (Mueck 2005). It was not until the late 1990s that their benefits were recognized by traffic signal practitioners. Two major characteristics make German ATCSs distinctive: they attempt to address optimization of traffic sig- nals based on network-wide changes in traffic demand by tak- ing into consideration the estimated originâdestination flows in the network, and their logics are adjusted to work with German industry standards for local traffic controllers and public transit priority. To summarize, ATCSs have been used since the early 1980s. Although there are at least 25 ATCSs deployed in the United States, these systems may not be well understood by many traffic signal practitioners in the country. Their opera- tional benefits have been demonstrated in several cases, but some professionals argue that the systems are no better than good time-of-day (TOD) actuated-coordinated plans. Other issues with ATCSs include detector maintenance and com- munications problems and overall that these systems are con- sidered expensive and complicated (Crenshaw 2000; Hicks and Carter 2000). Previous surveys on ATCS implementa- tions provided some of the underlying sources of agenciesâ reluctance to widely deploy these systems. One of the major purposes of this study is to provide insight into all these issues from the perspective of an ATCS user to explain why ATCSs have not been utilized more, especially in the United States. STUDY GOALS AND OBJECTIVES The goal of this study was to summarize the state of practice in deploying ATCSs in North America, with an overview of ATCS deployments around the world. In this study, a broader definition of an ATCS was adopted to include all systems that adjust their signal timings in real time based on changes in cur- rent traffic conditions (excludes actuated and traffic-responsive pattern-selection systems). This study adopts an ATCS defini- tion that includes all systems defined as traffic-responsive and traffic-adaptive control under the third generation of traf- fic signal control systems (Gordon and Tighe 2005). The goal was achieved through the following objectives: ⢠Describing operational characteristics of major ATCS deployments; ⢠Identifying and describing widely deployed ATCSs, including a description of their working principles and operational requirements; ⢠Identifying operational advantages and disadvantages of deploying ATCSs, along with the problems with imple- mentation and lessons learned; ⢠Identifying institutional problems at agencies that deploy ATCSs, along with documenting their experiences; and ⢠Investigating implementation costs and benefits per- ceived by ATCS users.
7STUDY METHODOLOGY The study methodology consisted of three tasks. The first focused on the selection of ATCSs, which are typically deployed in the United States (and worldwide) and iden- tification of ATCS agencies that need to be interviewed. The second task was to conduct a literature review and gather as much information as possible about ATCS oper- ations and deployments from previous studies. Finally, two electronic surveys were conducted: a shorter e-mail survey for ATCS vendors and a longer website-based survey for ATCS users. Selection of Adaptive Traffic Control Systems and Adaptive Traffic Control Systemsâ Deployment Agencies More than 20 different ATCSs have been developed during the last 30 years. However, only about a dozen of them have been applied in the real world and have more than one field imple- mentation. In this study, it was decided to focus effort only on those systems that are implemented in the field. There were a few international systems that were not possible to describe in detail because their developers/vendors did not express inter- est in participating in the study. As a result, five U.S. and five international systems were investigated. Seven of those 10 sys- tems are deployed in the United States, whereas the other three systems are currently deployed in Europe. The ATCSs con- sidered in this study are ACS Lite, BALANCE, InSync, LA ATCS, MOTION, OPAC, RHODES, SCATS, SCOOT, and UTOPIA. Selection of the agencies that deploy ATCSs was straight- forward because the intention was to interview representa- tives from all public agencies in the United States that operate ATCSs. However, identification of these agencies was a somewhat difficult process because there is no single source containing such information. Therefore, identification of the ATCS agencies was based on the literature review and com- munications with traffic signal professionals, among which the studyâs panel members provided important assistance. Literature Review A comprehensive literature review on ATCSs included the use of print and online resources such as Transporta- tion Research Information Services (TRIS), Transportation Research Records, and ASCE and Elsevier websites. Among the documents reviewed, some provided general descriptions of ATCS deployments and potential operational challenges. Others were about evaluations of specific ATCS implemen- tations. In addition, few publications contained information about the classification of ATCSs and conventional traffic signal systems. There were many academic studies in which ATCS logics were evaluated in a microsimulation environ- ment. Several documents from engineering conferences sum- marized the current status of ATCS deployments, with a future outlook of such systems. Analyzing these documents provided insight into ATCSs, created sound knowledge of existing implementation issues, and established a platform for evalu- ation of the survey data. The study itself does not refer to all of the documents gathered through the literature review. For the purposes of future research on ATCS, they are categorized and cited in Appendix C. Surveys The survey conducted under this study had two components. The first was a request sent by e-mail to all major ATCS devel- opers or vendors to provide accurate and up-to-date descrip- tions of their systems. The ATCS vendors were requested to provide descriptions of the adaptive logic, hardware and soft- ware requirements, system architecture, detection require- ments, and other special features of their systems. Most of the ATCS developers and vendors responded by identifying key studies that best describe their systems. Some ATCS vendors and users provided specific descriptions that they wanted to be part of this study. These descriptions followed the requested format but were sometimes broader than the scope of this study and therefore were edited. The second component of the survey was a questionnaire that included quantitative and qualitative questions and was delivered through a web-based survey tool. A link for the ques- tionnaire was sent to all ATCS users identified in the previ- ous task and responses were collected over 3 to 4 months. The questionnaire was designed to gather as much information as possible on major North American ATCS deployments. It had several sections (e.g., system requirements, operations, training, and costs), each of which contained multiple ques- tions. The questionnaire was designed to include both multiple- choice questions and open-ended questions. Multiple-choice questions were used when there was some certainty that the suggested answers adequately represented the range of likely answers. The option to add an answer was provided frequently. Open-ended questions were used when there was uncertainty as to the anticipated answer. The final version of the question- naire (slightly different than the one offered on the web, owing to technical modifications that the web version requested) is provided in Appendix A. AGENCY PARTICIPATION The survey was originally distributed to 42 agencies that run ATCSs in North America (United States and Canada) and several dozen locations around the world. Numerous follow- up requests were made, by e-mail and telephone, to encourage those agencies that had not responded to participate in the sur- vey. Table 1 is a list of those agencies that responded and the type of ATCS that these agencies operate. In a few cases, respondents who were interviewed do not currently work for agencies with an ATCS. However, they were recognized as the best experts to answer questions about the ATCSs even
after they left their respective agencies. Also, in few instances, individuals from agencies that do not currently run ATCSs were interviewed. Some of these agencies went through an ATCS procurement process but decided not to install a system. Other agencies had only probationary deployments of their ATCSs, which were removed after the trial periods. Finally, some agencies that shut down their ATCSs were also inter- viewed. Overall, the response rate from North American ATCS users was slightly more than 80%. 8 ANALYSIS APPROACH The analysis approach had two major components. First, when- ever ATCS users responded to a multiple-choice question, results were reported in the text without (or with minimal) further manipulation. For the questions considered to be very important, the descriptions of the results were accompanied by corresponding charts or tables. In the event of open-ended questions, answers were categorized before being presented. Agency System U.S. Deployments International Deployments City of Longview, TX W.E. Stilson Consulting Group, LLC, Columbus, OH City of Little Rock, AR California Department of Transportation â District 7, CA Culver City, CA Los Angeles Department of Transportation, CA City of Chesapeake, VA Town of Cary, NC Virginia Department of Transportation, VA Pinellas County, FL City of Tucson, AZ Washington State DOT, WA City of Chula Vista, CA City of Gresham, OR City of Menlo Park, CA City of Santa Rosa, CA City of Sunnyvale, CA Cobb County, GA Delaware Department of Transportation, DE Florida DOT District 4, FL Minnesota Department of Transportation, MN Pasco County, FL Road Commission for Oakland County, MI Utah Department of Transportation, UT City of Anaheim, CA City of Ann Arbor, MI Collier County, FL Orange County, FL Reedy Creek Improvement District, FL Short Elliott Hendrickson Inc., MN Econolite Canada Inc., Canada Dublin City Council, Ireland New Zealand Transport Agency, Auckland, NZ RTA, New South Wales, Sydney, Australia UOCT, Concepcion, Chile VicRoads, Victoria, Australia City of Blackpool Council, UK City of Red Deer, Canada City of Southampton, UK City of Toronto, Canada Derby City Council, UK Greater Manchester Urban Traffic Control Unit, Halifax Regional Municipality, Canada Hampshire County Council, UK I Mo TS Siemans Ltd., Beijing, China ACS Lite ACS Lite InSync LA ATCS LA ATCS LA ATCS OPAC OPAC OPAC OPAC, RHODES RHODES RHODES SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCATS SCOOT SCOOT SCOOT SCOOT SCOOT SCOOT RHODES SCATS SCATS SCATS SCATS SCATS SCOOT SCOOT SCOOT SCOOT SCOOTS SCOOTS SCOOTS SCOOTS SCOOTS TABLE 1 AGENCIES PARTICIPATING IN SURVEY
9For very important open-ended questions, whose answers represented lessons learned on a certain subject, most of the responses were summarized in tabular format. Along with the data extracted from surveys, existing literature was used as material in the report. REPORT ORGANIZATION This report consists of eight chapters. The first chapter defines an ATCS, provides a short history of the ATCS, and covers scope, objectives, and study methodology. Chapter two deals with the operational background of the interviewed agen- cies and environments in which their ATCSs work. Chap- ter three summarizes some of the working principles of major ATCSs deployed worldwide as well as their hardware and software requirements and operational benefits. Chapter four covers institutional aspects of ATCS implementations, with a major emphasis on the training, operations, and mainte- nance of ATCSs. Chapter five covers system requirements, which are necessary for proper operations of ATCSs. The chapter describes ATCS needs for traffic detection, hardware, software, and communications and how those needs are per- ceived by ATCS users. Chapter six covers costs and benefits from ATCS implementations. Chapter seven provides lessons learned from various ATCS usersâ perspectives. Finally, chap- ter eight summarizes the information presented in previous chapters and offers conclusions that may help agencies inter- ested in deployments. Separate lists of references and acronyms precede three appendices. Appendix A presents working prin- ciples of ten major ATCSs widely deployed in the United States and around the world. This appendix is based primarily on input from ATCS developers and vendors, and a comprehen- sive ATCS literature review. Appendix B contains the survey questionnaire. Appendix C provides an ATCS bibliography, categorized based on ATCSs described in the study.
