Managing Selected Transportation Assets: Signals, Lighting, Signs, Pavement Markings, Culverts, and Sidewalks (2007)

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.

OVERVIEW Traffic signals help manage intersecting streams of automo- bile and truck traffic, pedestrians, cyclists, and other road and transit vehicles by assigning the right-of-way to individual streams in turn. They are placed where the volumes of traffic or crash histories justify their need, where crossings near schools require signal control, or signal installation is needed as part of a coordinated signal plan to ensure a smooth, pro- gressive flow of vehicle platoons. The Manual on Uniform Traffic Control Devices for Streets and Highways (MUTCD 2003) establishes standards and warrants for signal installa- tion and operation, as well as general guidance on responsi- bility for maintenance. The Code of Federal Regulations (23 CFR 655.603) recognizes the MUTCD as the national stan- dard for traffic control devices, including signals, on all pub- lic highways, streets, and bicycle trails in the United States. It further requires that any supplementary manuals or guidelines issued by other federal and state agencies shall substantially conform to the national MUTCD. Traffic control devices must also conform to standards issued or endorsed by the FHWA (MUTCD 2003). The study survey asked agencies to identify their key sources of technical guidance for management of traffic signals. The purpose of the question was to understand per- ceptions of what are the important drivers of engineering and management decisions regarding traffic signals, rather than to cite a complete list of legal and engineering authorities. Fig- ures 1 and 2 present these results for two key aspects of asset management: new construction and installation, and mainte- nance and rehabilitation, respectively. The importance of in- dividual agency guidelines as well as national standards such as the MUTCD is evident. AASHTO (A Policy on Geometric Design of Highways and Streets 2004; Guide for the Plan- ning, Design, and Operation of Pedestrian Facilities July 2004), TRB’s Highway Capacity Manual (HCM 2000), and the Institute of Transportation Engineers (Giblin et al. 1989) provide additional guidelines regarding signal and lamp char- acteristics; recommended geometric characteristics of inter- sections; methods to compute the capacity of signalized inter- sections; pedestrian signal timing, phasing, and warrants; and recommended preventive maintenance schedules. AASHTO has published Standard Specifications for Structural Supports for Highway Signs, Luminaires, and Traffic Signals (2001), and the Roadside Design Guide (2002) for locating and in- stalling signal supports in the roadside, particularly regarding 14 safety in cases where vehicles run off the road. Building on the AASHTO specifications, subsequent studies have devel- oped updated, more comprehensive information on several technical aspects of structural supports for signals (e.g., Dex- ter and Ricker 2002; Fouad et al. 2003). Individual agencies may address signalization as part of their intersection design guides (e.g., Florida Intersection Design Guide . . . 2002). The city of Edmonton noted corresponding Canadian guidance by TAC, the Canadian MUTCD, the International Municipal Sig- nal Association, the Canadian Standards Association, and the TAC Chief Engineers’ Council for new installations; and the International Municipal Signal Association and Chief Engineers’ Council for maintenance and rehabilitation. The Saskatchewan H&T mentioned that another source of guid- ance it complies with is the requirements of the municipality with which the provincial agency partners in managing its sig- nals in this urban area. The MUTCD cites a number of potential benefits of correct signal installation: the orderly and efficient movement of what would otherwise be potentially conflicting traffic streams, improved safety, and increased intersection capacity. From an asset management perspective, maintaining signals in a state of good repair can serve these and other transportation objec- tives. The agencies participating in the study survey ranked several factors in order of perceived importance, as shown in Table 2. Although safe, efficient traffic movement was at the top of the list, the survey results also confirmed the importance of very responsive maintenance policies in reducing the life- cycle costs of managing these traffic control assets. The par- ticular ranking shown in Table 2 received very strong majori- ties among the responses of agencies at all levels of gov- ernment from the United States and Canada. This importance of signal systems to effective transportation operations has been recognized in several quarters, including this comment by the General Accounting Office [now General Account- ability Office (GAO)] as cited in a report by the National Transportation Operations Coalition (NTOC): Properly designed, operated, and maintained traffic control sig- nal systems yield significant benefits along the corridors and road networks on which they are installed. They mitigate congestion and reduce accidents, fuel consumption, air pollutants, and travel times. These benefits are documented in numerous evaluations, provided to us (the GAO) by the Federal Highway Administra- tion (FHWA), states, cities and other sources that compared before-and-after results when signal systems were installed, expanded, or retimed (U.S. General Accounting Office, Mar. 1994, as reported in NTOC 2005a, p. 3). CHAPTER TWO TRAFFIC SIGNALS

15 Notwithstanding this consensus on the importance and the value of traffic signals, reviews of current practice by others have identified shortcomings. An earlier NCHRP synthesis study considered good-practice system engineer- ing techniques as applied to traffic signals, including the use of a structured analysis, identification of goals and problems to be addressed, project management approach, alternatives evaluation and project evaluation, specific topics within traffic signal systems engineering (e.g., need for signals, signal timing, signal coordination, and coordination of traf- fic control systems), communication system engineering, local intersection control (e.g., local actuation strategy, signal priority, railroad and emergency vehicle preemption, and transit signal priority), signal procurement, operations and maintenance, and training (Gordon 2003). Responses to the survey conducted in that study indicated that although certain systems engineering techniques are well known and widely applied (e.g., evaluation of need for signals, signal timing, emergency vehicle and railroad preemption, main- tenance, and training), several of the other available engineering methods are not widely or frequently used by practitioners, for the following reasons (Gordon 2003): • Practitioners are unfamiliar with the methods or lack a user-friendly format or tool for easy application. • Agency guidelines and standard specifications may limit designer choices among alternatives and represent preference or selection criteria different from those as- sumed in the available engineering methods. • Resource constraints and compatibility requirements with existing systems or equipment may further limit choices among design alternatives and favor simple, easy-to-maintain equipment. In 2005, NTOC provided a report card on the nation’s traffic signal systems (NTOC 2005a,b). The objectives of this exercise were to: • Determine the current state of signal system operation in six areas and create an awareness of signal status, • Strengthen the understanding of the congestion-reducing benefits of good traffic signal operation, • Build a case for more attention and additional investment in signal systems, and • Provide a benchmarking tool for agencies to assess their own performance. This report card was developed by a self-assessment incorporating responses from 378 state, county, and local agencies across the United States, representing agencies having signal system inventories ranging from fewer than 50 signals to more than 1,000. The self-assessment was or- ganized and prepared by AASHTO, APWA, Institute of Transportation Engineers, Intelligent Transportation Society Agency Guidelines Public Policy Natíl. Standards Statutes Other No Response 0 20 40 60 80 100 Percentage of Responses FIGURE 1 Technical management guidance for new construction and installation of signals. Agency Guidelines Public Policy Nat’l. Standards Statutes Other No Response 0 20 40 60 80 100 Percentage of Responses FIGURE 2 Technical management guidance for maintenance and rehabilitation of signals.

of America, and the FHWA. It considered signal operation in five areas (a sixth area received a small number of responses and was therefore not graded), and aggregated these results to produce an overall national rating. The report card grades assigned to the five areas and the overall national result were as follows: proactive management (F), coordinated systems (D−), individual intersections (C−), detection (F), mainte- nance (D+), and overall national results (D−). These low grades do not mean that signals across the country are failing to display green–yellow–red. Rather, they point to deficiencies in system operation and integration, a limited degree of proactive management, and the effect of resource constraints. Additional findings of the NTOC sur- vey are presented in later sections. A survey of 120 state and local agencies with traffic signal responsibility was recently conducted by the FHWA as part of its signal systems asset management review of state of the practice (“Signal Systems Asset Management . . .” n.d.). These results reinforce some of the findings of the NTOC survey. • Respondents were asked to ascribe high, medium, or low priority to a number of signal system operational im- provements. The improvements that received the greatest number of high-priority responses (i.e., by more than 40% of respondents) included adjusting and upgrading existing signals, integrating signals within one’s own jurisdiction, improving system capabilities, and establishing or up- grading a traffic management center. • Other system-related improvements, including signal- izing more intersections, coordinating with other juris- dictions, complying with Intelligent Transportation System architecture, and upgrading system software, received fewer high-priority responses; generally less than 30%. • Participants were also asked about their priorities for physical signal system repair. Only repair and replace- ment of equipment received a high priority from more than 50% of respondents. Upgrading communications, reducing responsive repair costs, and standardizing components were rated as a high priority by at least 30% of respondents. 16 Agencies participating in the FHWA signal system asset management survey also reported a range of estimated annual budget amounts for signal systems, considering funding from all sources (federal, state, and local) (“Signal Systems Asset Management . . .” n.d.). • Regarding annual construction budgets for new signal installations and upgrades of existing systems, agencies divided almost uniformly among four annual cost ranges: less than $0.5 million; $0.5 to $1.0 million; $1 to $2 million; and more than $2 million. These levels of expenditure were roughly correlated with the size of the signal system, ranging from small systems (fewer than 300 signals) to large systems (more than 1,000 signals). • Regarding annual maintenance budgets for preventive and emergency work, reporting agencies were again distributed almost uniformly among the four expendi- ture levels. However, there was little apparent corre- lation between the level of expenditure and system size. • Regarding annual operations budgets for items such as signal timing plans, almost half the responding agen- cies reported annual budgets of less than $0.5 million. The remaining agencies were divided almost uniformly among the other expenditure categories. Again, there was little correlation between annual expenditure and system size. MANAGEMENT PRACTICES Synthesis and AASHTO–FHWA Survey Findings Maintenance of traffic signals is often characterized by a sharing of responsibility among public and private organiza- tions, as indicated in Figure 3 based on the study survey. Although some DOTs, local agencies, and provincial ministries are solely responsible for both overall management as well as conduct of signal maintenance, many agencies en- list other groups in this work through outsourcing to private contractors (by all levels of government) and partnerships or intergovernmental agreements with other levels of govern- ment (e.g., state and provincial agencies in arrangements with counties or municipalities). When other levels of government Rank Factor 1 Public safety; accident and accident risk reduction 2 More efficient travel; maintain intended flow and operating speed; reduce travel time 3 Preservation of existing road infrastructure; reduced agency life-cycle costs 4 Comfort and convenience of the traveling public (motorists, pedestrians, cyclists) 5 Road aesthetics and appeal TABLE 2 PRIORITY OF TRANSPORTATION OBJECTIVES SERVED BY TRAFFIC SIGNALS

17 are involved, they typically exercise management responsi- bility for their work. By contrast, in the majority of cases where private firms maintain signals, they are not given man- agement responsibility. The Pennsylvania DOT (PennDOT) described its statewide signal management arrangement in which traffic signals, including those on state highways, are owned, oper- ated, and maintained by the municipality in which they are located. PennDOT is responsible for approving the installation, revision, or removal of traffic signals. PennDOT issues a traffic signal permit to the local municipality, which outlines the design and operation of each specific signalized intersection. The munici- pality is responsible for operating and maintaining the signal in accordance with the permit. PennDOT also maintains statewide standards, specifications, lists of approved materials, and main- tenance guidelines. — Pennsylvania DOT (PennDOT) Saskatchewan noted that its sharing of maintenance responsibility affects management practice and available information about these assets. Our department does not have an inventory or budget for traffic signals. The department usually [enters] into a cost-shared agreement with the municipality for installation. The urban mu- nicipality is normally responsible for maintenance after initial installation. — Saskatchewan H&T Other aspects of asset management practice are revealed through an agencies’ methods of budgeting for preservation, operation, and maintenance of traffic signals, and their approaches to preserving and maintaining signals (including re-timing) once in service. The options that were presented to surveyed agencies are listed in Tables 3 and 4, accompa- nied by abbreviated descriptions used to describe the survey responses in Figures 4 and 5, respectively. Because agencies could choose more than one response in each of these topics, and many did so, the percentages in Figures 4 and 5 do not sum to 100%. Regarding methods of budgeting, a large number of responding agencies at all levels of government chose the “Previous Budget Plus Adjustments” option and the “Staff Judgments, Political Priorities, and Citizen Demands” options as best describing their processes. These two were often selected in combination and sometimes in conjunction with one or more of the other options shown in Figure 4 as well. The “Other” methods indicated in Figure 4 specify that, for one agency, traffic signals are rated in its safety improvement 0 20 40 60 80 100 Own Agency Private (Outsourced) Other Gov’t. Unit Other Entities No Response Pe rc en ta ge o f R es po ns es Mgmt. Resp. No Mgmt. Resp. FIGURE 3 Responsibility for maintaining signals once in service. Budget Method Options—Full Description Abbreviated Description Budget reco mme ndations based on the cost to achieve a perform ance target for assets (i.e., target drives budget) Target Drives Budget Budget reco mme ndations ma xi mi ze the asset perform ance target that can be achieved for the available funding (i.e., budget drives target) Budget Drives Target Budget reco mme ndations based on addressing a percentage of the inventory each year Percent Inventory Annually Budget reco mme ndations based on previous year’s budget plus inflation and other adjust me nts Previous + Adjust me nts Budget reco mme ndations based on staff judgm ents, political priorities, and citizen de ma nds Judgm ent, Politics Budget reco mme ndations based on a percentage of the total anticipated budget Percent of Total Budget Other approach (describe briefly) Othe r No specific approach No Specific Approach TABLE 3 SURVEY DESCRIPTIONS OF BUDGET METHOD OPTIONS

program for purposes of budgeting. For two other agencies, this option was chosen because signals are maintained by local governments. The survey results in Figure 4 show that the number and target performance of assets are used to a degree in budget- ing, but are not the primary drivers of budget processes among survey respondents. Approaches based on Target [Asset Performance] Drives Budget, Budget [Asset Perfor- mance] Drives Target, and Percent of Inventory Budgeted Annually each were identified in less than 20% of responses. By contrast, methods based on Adjustments to the Previous Budget and those that involve Staff Professional Judgment, Political Priorities, and Citizen Demands each garnered roughly 40% of the responses (bearing in mind that agencies could select more than one approach). The general thrust of these results was complemented by a January 2000 AASHTO survey of roadway safety hardware reported by the FHWA (Hensing and Rowshan 2005). When asked whether asset inventory and asset condition were used as the basis of funding allocation, 11 of 39 states (28%) responded 18 affirmatively for signal inventory, and 15 of 39 (38%) for signal condition—again, less than a majority in each case. The corresponding results for supports for signs, signals, and lighting in this AASHTO survey were 6 of 39 states (15%) responded affirmatively that funding allocation is based on supports inventory, and 11 of 39 (28%) that allocation is based on supports condition. A related question in the January 2000 AASHTO survey (Hensing and Rowshan 2005) asked whether state DOTs have a separate budget line item for maintenance of signals. Twenty-one of 39 agencies (54%) responded affirmatively. The corresponding result for maintenance of sign, signal, and lighting supports was 8 of 39 agencies (21%) responding affirmatively. Although there was no corresponding question for budgeting of new signal installations, the survey did address tracking and updating of asset inventory. These additional responses are reported later in this chapter. Multiple selections were often made as well by agencies describing their approaches to preserving and maintaining 0 20 40 60 80 100 No Response No Specific Approach Other Pct. of Total Budget Judgment, Politics Previous + Adjustments Pct. Inventory Annually Budget Drives Target Target Drives Budget Percentage of Responses FIGURE 4 Annual budgeting method for signal preservation and operation. Approaches to Asset Maintenance and Preservation—Full Description Abbreviated Description Preventive maintenance carried out on a set schedule Preventive—Schedule Immediate—repairs carried out as soon as possible after damage or failure is reported Corrected Immediately Corrective—repairs prioritized and scheduled to meet performance targets subject to resource constraints Prioritized—Avail. Res. “Worst-first”—limited number of repairs each year, but backlog exists Worst First Deferred maintenance—little or no work performed annually Deferred Maintenance This agency does not maintain traffic signals No Maintenance Responsibility Other Other TABLE 4 SURVEY DESCRIPTIONS OF PRESERVATION AND MAINTENANCE OPTIONS

19 signals. Immediate correction of problems was the most prevalent response, as shown in Figure 5, reflecting the importance of signals to safety, good traffic movement, and the other transportation objectives discussed earlier. A pre- ventive approach and a priority approach subject to resource constraints were also selected by many agencies. The shared responsibilities for signal maintenance discussed earlier are also reflected in the responses. The FHWA survey of state and local agencies that was con- ducted as part of its state-of-practice review for signal system asset management asked respondents about their use of signal performance data for decision making (“Signal Systems Asset Management . . .” n.d.). At least half of the respondents reported using performance data for several kinds of decisions. The most prevalent uses (more than 70% of responses) were to identify needs for signal coordination and for improvement. These were followed by the identification of changes needed in traffic control, need for periodic signal timing, real-time signal timing, and planning for equipment replacement. Comments by several agencies, which are paraphrased here, provided additional details on other methods of signal management and why they describe signal maintenance and preservation often with multiple approaches: [“Other” approach]: We conduct audits of existing roadways on a 3- to 5-year cycle and make any signal-timing changes needed as part of these reviews. We also look at signal timing when requested. — Kansas DOT [Evolving approach]: While now following an Immediate ap- proach, we are now moving in the direction of preventive main- tenance. — New York State DOT (NYSDOT) [Multiple approaches]: Preventive Maintenance: limited, but occurring; Immediate: agency does address signal trouble calls; Worst first: capital reconstruction dollars are prioritized this way. — City of Portland, Oregon National Transportation Operators Coalition Report Card Findings The NTOC report card (NTOC 2005a) emphasized signal system management and operations, and identified a number of best practices that were used to benchmark the grading of responses: • Proactive management, including documentation of agency procedures and their communication to employ- ees; availability of technical personnel outside of normal business hours; communication with the public regarding problems such as signal outages, excessive delays, inci- dents and work zone closures, and other signal-related conditions affecting travel; coordination with outside groups such as special-event organizers, law enforcement agencies, and emergency service providers; easy access by the public to report/complaint centers; and internal policies and encouragement to agency staff to obtain relevant licenses, certificates, and degrees. • Coordinated signal management, including reviews of traffic signal timing every 3–5 years, or more fre- quently as needed; development and implementation of new citywide or corridor timing plans within one year of initial identification of need; use of effective data collection, analysis, and field testing procedures in de- veloping and implementing timing plans; development of plans for different traffic patterns and contingencies (e.g., special events, incidents, road work, and inclement weather); and coordination of signal timing with adjacent jurisdictions. • Signal operation at individual intersections, including a documented process identifying factors that will trigger 0 20 40 60 80 100 No Response Other No Maintenance Responsibility Deferred Maintenance Worst First Prioritized—Avail. Res. Corrected Immediately Preventive—Schedule Percentage of Responses FIGURE 5 Approach to maintaining and preserving signals.

reviews of timings; a documented, centrally accessible, current inventory of approved signal phasing and timing for each intersection; analyses of appropriate information supporting timing reviews, such as turning movement counts, pedestrian volumes, accident histories, complaint histories, field observations of clearance intervals, and checks on any geometric changes to the intersection since the last review; and quick implementation of timing plans once developed (within two working days). • Specialized operation of traffic signals, which addresses unique locations that require frequent study and adjust- ments, such as railroad crossings, light-rail corridors, reversible lane and ramp-metering locations, and loca- tions that experience incident response and emergency vehicle access. Best practices include an inventory of signals within 200 ft of grade crossings and signals operated by others; installation of signal preemptions at those grade crossings with and without adaptive controls in place; and regular measurement of the num- ber of train movements and speeds as well as vehicular traffic volumes and speeds, noting changes therein. • Detection systems, including an established process for gathering data on intersection traffic volumes and turn- ing movements; use of this information in computing signal timing; quality assurance procedures to check the accuracy of surveillance data; and basic quality checks such as physical inspection of detectors and communi- cations links. • Signal system maintenance, including adequate organiza- tional staffing (30–40 intersections per technician recom- mended); on-going funding commitment to signal system repair, upgrade, and replacement; inclusion of needed repair or replacement of signal system components dam- aged by road maintenance or utility work, as part of the project; training programs on signal maintenance, includ- ing latest equipment and procedures; regular inspections and assessments of signal control equipment condition and operation, and a semi-annual comprehensive assess- ment of all operating conditions; near-real-time monitor- ing and emergency response, including the computer and communications technology to provide reports of failure to maintenance personnel within 5 min of detection; use of a maintenance management system that supports pre- ventive maintenance policies and tracking of equipment performance histories to identify unreliable equipment; and establishment of agency policies, procedures, and cri- teria to prioritize among competing problems and define appropriate response times. The NTOC report card provided these best practices as a guide for agencies to improve their signal system management scores. Specifically, the NTOC recommended strengthened investment in traffic signal hardware, routine updates to sig- nal timing, and good maintenance to help reduce traffic delays, fuel consumption, and harmful vehicle emissions. Particular shortcomings in the several aspects of current 20 agency practice that were identified by the NTOC in responses to its report card questionnaire are as follows (NTOC 2005a): • Issues in proactive management—NTOC’s report ob- serves that its most noteworthy finding is the very poor grade attributed to proactive management: . . . 68 percent of respondents [either have] no documented management plan for their traffic signal operation or they are managing their signals on an ad hoc basis. Travelers use the transportation system 24 hours every day and traffic signals need to perform efficiently during that entire time; however, 71 percent of the agencies do not have staff resources committed for other than typical working hours, even if peak periods occur outside these hours. Even in the largest sig- nal systems (more than 450 signals), less than half (42 percent) reported good progress in this area. — NTOC 2005a, p. 10 • Issues in signal operations in coordinated systems— Although reviews of traffic signal timing are critical to optimal system operation and smooth traffic movement, 57% of report card respondents reported that they do not conduct these reviews routinely every 3 years or that their procedures in this area are ad hoc. Once the need for retiming is identified, 55% reported that they take more than 18 months to complete the task. Fewer than half of the reporting agencies coordinate signal timing with neighboring jurisdictions and only approximately one-quarter indicated that they adjust timings for re- vised traffic flows during special events. • Issues in signal operations at individual intersections— Routine reviews of signal timing at individual intersec- tions are not generally done, with 77% of respondents reporting “only ad hoc or no such process.” Survey results indicated that “little planning and organizational management of traffic signals updates are done. [Agen- cies’] resources are more likely to be allocated to deal with critical situations as they arise.” However, results also showed that when signal timing is ready to be addressed, more than 70% of the agencies stated that “they regularly update all aspects of the signal timing.” • Issues with detection systems—Effective signal timing relies on good information regarding traffic counts and movements. However, 33% of the report card respon- dents, and 23% of those responsible for large systems of more than 450 signals, reported “no regular process for collecting data to support traffic signal timing. Again, this is a likely indicator of staffing deficiencies.” • Issues in maintenance—Although good practice gener- ally recommends a maintenance technician for every 40 or fewer traffic signals, 29% of respondents reported ei- ther a level of 60 or more signals per technician, or that they have not considered their staffing level at all. This finding was taken as a further indication of reactive sig- nal maintenance in the face of resource constraints— that is, “putting out fires.”

21 • Overall issues—The low scores in the report card presented earlier are the result of resource constraints that inhibit more system-based actions and proactive management, and encourage agencies to resort to a “fire-fighting” mode. As a result, many agencies strive to meet only a basic level of service that provides safety and avoids liability. Although this solution may not be optimal from the perspective of vehicular and pedestrian traffic, in a technical sense the system is “working.” The low grades on the report card were “remarkably similar across the country and across jurisdictions,” suggesting that many signal systems “have the potential for greatly improved performance.” MEASURING ASSET PERFORMANCE Signal systems, encompassing support structures, the signal head and lamps, and electronic control and vehicle detection devices, may be characterized by many aspects of their performance. A number of options were listed in the study survey, categorized as follows: • Physical condition—structural condition, corrosion, inoperable or nonfunctioning components, use- or time- related degradation (e.g., dirt accumulation), and other factors identified by responding agencies. • Age of the system or asset. • Hours in service. • Operational performance—for example, proper signal timing. • System reliability—for example, number of failures in a certain time interval. • Performance or health index—a composite measure of condition or performance, the basis or computation of which agencies were asked to explain briefly. • Qualitative ratings of condition—for example, good–fair–poor, and the basis on which they are developed. • Asset value, in dollars. • Customer-related measures—that is, data from cus- tomer surveys and number or frequency of customer complaints. • Other factors identified by the agency. For several of these categories, agencies were also asked to specify the frequency with which these assessments are made: • More than once a year, • Annually, • Biennially, or • Less frequently than biennially. The information provided by agencies on performance measurement of traffic signals is summarized in Figure 6. Many agencies use both physical and qualitative measures of structural condition, the age of the signal system, operational performance, system reliability, and customer complaints. Other measures have varying degrees of use; none of the responding agencies reported using hours in service, a signal- related performance index, or customer surveys (even though one agency reported conducting such a survey). In addition to the measures shown, North Carolina monitors the condition of anchor bolt nuts on posts, and the city of Portland (Oregon) reports a qualitative measure based on age. As a group, the physical measures of condition are the most widely used among survey respondents. The frequencies with which these physical measures are assessed are shown in Figure 7. More than two-thirds of responding agencies monitor condition annually, and almost three-fourths at least biennially. The FHWA review of state of practice in signal system asset management asked respondents about the data they maintain on different signal system components. Results are shown in Table 5. Although signal heads and controllers were identified by the greatest number of agencies, the FHWA reported “significant variation” among agencies as to the type of data maintained. Note in Table 5 that only five cells represent component/type-of-information combina- tions that were reported by more than half of the respondents (“Signal Systems Asset Management . . .” n.d.). The rate of responses for items such as asset age, physical condition, and nonfunctioning components are similar in Table 5 and Figure 6. The FHWA survey also looked into the types of opera- tional system performance data that agencies track. More than 75% of responding agencies identified intersection crashes, intersection fatalities, and traffic volume or throughput. More than 50% added traffic speed and cus- tomer complaints. Other items such as queue lengths, stops, and signal downtime were tracked by 30% or less of the reporting agencies. Transit performance as a function of signals was not selected by any of the reporting agencies (Harrison et al. 2004). The methods used by responding agencies to assess signal condition and performance are reported in Figure 8. Visual inspections and customer complaints are by far the most com- mon methods used, with other options reported by no more than one in five respondents. In the latter cases, few agencies mentioned the specific technologies or devices they employ to gather signal condition data. Under “Other” methods, North Carolina listed ground rod resistance and Meggar tests (tests of resistance between the loop and ground) for induc- tive loops; the Oregon DOT noted the use of standard draw- ings from the time when signal poles were originally installed; and the city of Edmonton included ultrasonic nondestructive tests of signal supports.

22 More Than Once A Year Annually Biennially Less Freq Than Biennially FIGURE 7 Frequency of physical condition assessments of signals. PHYS: Structural Condition PHYS: Corrosion PHYS: Not Functioning PHYS: Use- or Time-Related PHYS: Other Asset Age Hours in Service Operational Performance System Reliability Performance or Health Index QUAL: Structural Condition QUAL: Corrosion QUAL: Not Functioning QUAL: Use- or Time-Related QUAL: Other Asset Value Customer Complaints Customer Surveys Other No Response 0 20 40 60 80 100 Percentage of Responses FIGURE 6 Measuring performance of traffic signals. PHYS = physical; QUAL = qualitative.

23 ASSET SERVICE LIFE Information on service life was obtained in the study survey for three major components of signal systems: the structural sup- ports—poles and mast arms, the controller system, and signal display items—the signal heads and lamps. For each of these components, agencies were given the opportunity to report ser- vice lives for different materials that are typically used or for other materials that they employ. Survey participants were also asked to list the main sources they use to estimate service life. Responding agencies rely on several methods to estimate these service lives, as shown in Figure 9. The activities in Figure 9 that can contribute to estimates of service lives include the following: • Development of predictive models or management in- formation systems to support management of these assets. • Development and use of life-cycle cost analyses to compare the performance and costs of alternative components. • Documented agency experience—for example, historical databases or other records of asset performance and service life. • Literature describing service-life experience by others. • Professional judgment of agency staff. • Manufacturer’s performance data. • Other sources of information identified by the re- spondents. Type of Information Signal Head s (%) Detectors (%) Controllers (%) Structures (%) Communication Equipm ent (%) General Characteristics (e.g., equipm ent m odels, functions, etc.) 46 46 62 35 50 Serial Nu mb ers of Com ponents 12 12 31 80 12 Maintenance Require me nts 12 15 27 80 15 Maintenance History, Costs 42 38 46 35 38 Repair or Failure History 38 31 50 35 38 Age or Condition 19 27 46 23 31 Source: FHWA “Signal Systems Asset Management...” n.d. Data show percentage of respondents to FHWA survey. TABLE 5 TYPE OF INFORMATION MAINTAINED FOR SIGNAL SYSTEMS No Response No Info. Collected Other Customer Complaints Customer Surveys Non-Destructive Testing Physical Measurement Photo, Video Visual Inspection 0 20 40 60 80 100 Percentage of Responses FIGURE 8 Data collection methods for signal condition and performance.

Among the 40% of reporting agencies that identified at least one method, the emphasis was on collective agency knowledge, whether embodied in documented experience (e.g., a database of observed historical service lives) or in the professional judgment of their staffs. Manufacturers’ data were also noted as an important source of information. The agencies’ estimates of service-life values are summa- rized in two ways. Table 6 presents statistical results in terms of the minimum and maximum values, and the three com- monly used measures of central tendency—mean, median, and mode—for every component and type of material reported. The number of responses on which these statistical parameters is based is also given. The second method is a display of histograms for those components and materials that represented relatively large numbers of data points in the survey. Figures 10 through 13 show service-life distributions for several types of signal supports, Figure 14 for traffic con- troller cabinets, and Figures 15 through 17 for signal display components. The labels on the horizontal axis in these figures give the upper values of each range of service-life data. For example, if these labels are 0, 5, 10, 15. . . , then the column labeled 5 shows the number of responses for estimated service life of zero to 5 years; the column labeled 10, the number of responses for estimated service life of more than 5 to 10 years; the column labeled 15, the number of responses for estimated service life of more than 10 to 15 years; and so forth. These graphics provide a clearer understanding of the shapes of the underlying distributions of estimated service life. It should be noted again that the data in Table 6 and Figures 10 through 17 may be derived in part from the professional judgment of agency personnel. A related question is how agencies determine where signal components are in their respective service lives. Knowing how much life is consumed, and how much remains, is necessary in 24 applying the service-life concept. Agencies included in the survey were presented with a number of ways to determine the current status of an asset regarding its service life and asked to rank each method by relevance to their agency. The results are shown in Table 7. Note that two instances of tie values occurred in this particular ranking process. The survey results in Table 7 reflect the importance reporting agencies assigned to quick maintenance response following asset damage or failure. This response is consistent with earlier findings regarding the importance of signals to customer safety and efficient movement, and the desirability of immediate correction of problems, particularly for signals at high-priority locations. An equally strong response, how- ever, was to indicate that many reporting agencies do not use or monitor service lives in their management of traffic signals. It should be noted, however, that several agencies were able to provide data on estimated service lives of signal components, even though service life is not used within their current management procedures. Although maintenance and rehabilitation are believed to extend service life, only 2 (of 31 total) responding agencies indicated that they take this into account, and only one pro- vided an explanation. The city of Portland conducts partial intersection reconstructions in lieu of complete intersection replacements as the result of budget constraints. Work includes replacing only those poles, span wires, and signal heads that are in bad condition. The expectation is that this strategy extends the life of the signal system at the intersec- tion for an additional 25 years. Although service life is one dimension of performance, other aspects of signal system operation are also critical. The TRB Millennium Paper on traffic signal systems discusses several issues related to the complicated operational No Response Do Not Use Service Life Other Manufacturer’s Data Professional Judgment Literature Agency Experience LCC Analyses Model Develop, MIS 0 20 40 60 80 100 Percentage of Responses FIGURE 9 Sources for determining service lives of signal systems. MIS = management information systems; LCC = life-cycle cost.

25 environment in which modern signals may be expected to operate—for example, with advanced features such as closed-loop systems coordination, preemption by emergency vehicles, transit vehicle priority, and handling of bicycle traf- fic—as well as the need for system component compatibility and integrated, interoperable systems (Bullock and Urbanik 2000). As a specific example, the Ohio DOT has considered the safety aspects of substituting light-emitting diode (LED) lamps for incandescent lamps, given the electrical character- istics of existing signal hardware that detect when a lamp has failed and respond by placing the signal in a flashing mode. The study has concluded that the LED lamps that were tested are compatible with modern hardware in existing incandes- cent systems (Gilfert and Gilfert 2002), although this topic may be researched further. The Ohio DOT report notes that whereas incandescent bulbs are replaced at 12-month or 18-month intervals, LED lamps should provide, based on current findings, service lives of at least 5 years. The structural performance of signal supports has also been a topic of recent interest. Research has resulted in updated guidelines and specifications for structural supports (Standard Specifications for Structural Supports . . . 2001, up- dated 2003; Fouad et al. 2003). Investigations have also been conducted of premature structural failures in signal supports (Chen et al. 2002). Component and Material No. of Responses Minimu m (Years) Maximu m (Years) Mean (Years) Median (Years) Mode (Years) Structural Com ponents Tubular steel ma st ar m 14 10 50 24.6 20 20 Tubular alum inum ma st ar m 7 2 0 3 5 24.3 20 20 Wood pole (and span wire ) 9 2 30 15.1 15 15 Concrete pole (and span wire ) 2 1 0 1 5 12.5 12.5 — Steel pole (and span wire ) 9 1 0 3 0 22.8 20 20 Galvanized pole and span ar m 1 — — >100 — — Controller System Components Perm anent loop detector 14 3 2 0 8.6 7.5 10 Non-invasive detector 12 5 2 0 10.4 10 10 Traffic controller 18 4 2 0 13.5 15 15 Traffic controller cabinet 17 10 30 17.5 15 20 Twisted copper interconnect cable 11 5 3 0 17.7 20 20 Fiber optic cable 7 2 0 3 0 23.6 20 20 Signal Display Components Incandescent la mp s 15 0.5 3 1.4 1 1 Light-em itting diode lam ps 18 5 1 0 7.2 6.5 5 Signal heads 15 7 3 0 18.8 20 15 Pedestrian displays 1 — — 1 5 — — Notes: —, value is undefined for the particular distribution. When distribution is based on only one data point, its value is shown in the Mean column TABLE 6 ESTIMATED SERVICE LIVES OF SIGNAL SYSTEM COMPONENTS

87 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 12 Estimated service life of wood pole and span wire supports for signals. 8 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 11 Estimated service life of tubular aluminum mast arms. 26 8 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 10 Estimated service life of tubular steel mast arms.

89 10 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 Estimated Service Life, Years 8 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 14 Estimated service life of traffic controller cabinets for signals. 27 8 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 13 Estimated service life of steel pole and span wire supports for signals. FIGURE 15 Estimated service life of incandescent lamps for traffic signals.

Rank Factor 1 Assets are repaired or replaced as soon as they fail without regard to service life 2 Service life is often determined more by functional obsolescence than by wear and tear 2 Compare current age of asset with the maximum age that defines service life 4 Monitor condition of the asset on a periodic schedule 5 Monitor condition of the asset occasionally 5 The agency does not use/does not monitor service life for this type of asset 7 Assets are replaced on a preventive maintenance schedule without regard to where they are in their service life 8 Apply deterioration models to estimate where the asset is on “the curve” 8 Compare service hours to date with the maximum number of service hours that defines service life TABLE 7 RANKING OF METHODS TO DETERMINE WHERE TRAFFIC SIGNAL ASSETS ARE IN THEIR SERVICE LIVES 8 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 Estimated Service Life, Years FIGURE 17 Estimated service life of traffic signal heads. 28 8 9 10 7 6 N o. o f R es po ns es 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 Estimated Service Life, Years FIGURE 16 Estimated service life of light-emitting diode lamps for traffic signals.

29 INFORMATION TECHNOLOGY SUPPORT Synthesis and AASHTO–FHWA Survey Findings As a practical matter, applying asset management to today’s transportation systems typically requires substantial informa- tion technology (IT) support. This support can be provided through a number of IT features for data processing, analytic modeling, decision support, communication of asset perfor- mance, and management accountability. Agencies participat- ing in the study survey identified their key IT capabilities as shown in Figure 18. Many (but not all) agencies have an inventory of signal assets accompanied by information on lo- cation. Photographs, dates and recommendations of inspec- tions, the age of assets, maintenance schedules, and tracking of public comments were reported by many agencies. One in five respondents noted that information on anticipated service life is available in their IT systems. No strong distinctions in the findings represented by Figure 18 were observed among levels of government. By comparison, responses to the January 2000 AASHTO survey reported by the FHWA (Hensing and Row- shan 2005) indicated that 27 of 39 agencies (69%) had an inventory of signals, and 25 of 39 (64%) updated their inven- tory by some method, in most cases by manual survey. In addition to these individual IT capabilities, agencies were asked to characterize the type of system(s) that they use to help manage signals. The categories of systems listed in the synthesis survey included the following: • A dedicated traffic signal monitoring system; • A broad-based management system such as a mainte- nance management system (MMS) or a comprehensive transportation infrastructure asset management system (TIAMS) that includes traffic signals as well as other assets; • Simple programs that address traffic signals; 0 10 20 30 40 50 60 70 80 90 100 No Response None of the Above Other Historical Database PMs, Dashboards, Accountability GIS Maps, Reports GIS Interface Est. Asset Impacts on Public Track Public Comments Cost Models for Treatments Other Optimization Procedures Benefit-Cost, LCC Decision Rules or Trees Inspector Recommendations Established Mntce. Schedule Deterioration Models Anticipated Service Life Dates of Inspections, Assess. Asset Age Usage, Traffic Volume Photograph Current Condition, Performance GPS Coordinates Location (e.g., Rte-Milepost) Number/Quantity of Asset Percentage of Responses FIGURE 18 IT capabilities to help manage signal systems. GPS = global positioning system; LCC = life-cycle cost; GIS = geographic information system; PMs = performance measures.

• Workbooks or spreadsheets that address traffic signals; and • Parts of other products or procedures that the agencies were requested to describe briefly. The responses to this survey item are summarized in Figure 19, showing a relatively uniform distribution of use across the several system categories, with no strong distinc- tions among different levels of government. Multiple responses by many agencies suggested that different tools were used at different levels within the organization. The agencies that reported using a signal management system or a maintenance management or transportation infrastructure asset management system that includes signals are listed here. • Signal Management System – Michigan DOT – Minnesota DOT – North Carolina DOT – Ohio DOT – Oregon DOT – Colorado DOT Region 4 – City of Edmonton, Alberta. • Maintenance or Asset Management System That In- cludes Signals – Maryland SHA – New Mexico DOT – Ohio DOT – Oregon DOT – Virginia DOT – Colorado DOT Regions 1 and 5 – Ministry of Transport of Quebec – City of Edmonton, Alberta – City of Portland, Oregon. The “Other Products or Procedures” responses in Figure 19 included mention of a Traffic Signal Information System data- base (Oregon DOT), the tracking of signals by the Kansas DOT through its Audits of Existing Roadways, and the following comment by PennDOT: 30 Some of our Regional Engineering District Offices have traffic signal asset databases. We are embarking upon an effort to develop a statewide traffic signal asset management system. — PennDOT FHWA State-of-Practice Findings and Signal System Framework The FHWA survey of state and local agencies for its state- of-practice review of signal system asset management asked participants what software tools they used for signal system management. More than 95% of respondents identified signal timing optimization and simulation as a tool they used to help generate signal timing plans. Other IT capabilities that were identified by more than 50% of respondents included inventory tracking of system components encompassing identification, location, classification, and date acquired or constructed for each item; maintenance and work order management; and bud- geting for capital, maintenance, and operations expenditures. Applications to track the available inventory of spare parts, exercise version control of signal hardware and software, and conduct system operational performance monitoring were reported in use by 30% to 40% of respondents (“Signal Systems Asset Management . . .” n.d.). The FHWA has proposed the architecture of a generic signal system asset management system (SSAMS) that con- forms to the principles in the AASHTO Transportation Asset Management Guide (Cambridge Systematics, Inc. 2002). This report and associated web document describes the features and decision-support capabilities of a SSAMS, how it could be applied to different signal improvement scenarios, and how a SSAMS compares with other asset management systems. The generic SSAMS is structured of several modules: • Physical characteristics of the signal system—for example, signal components, detectors, controllers, communica- tions, and central control hardware and software. • Operational characteristics of the signal system—for example, timing plans, control strategies, coordination, preemption, and design and placement. Simple Program(s) for this Asset Broad-Based MMS, TIAMS, etc. Signal Management System Percentage of Responses Workbook, Spreadsheet Other Products or Procedures 0 20 40 60 80 100 FIGURE 19 Types of analytic tools to support signal system management. MMS = maintenance management system; TIAMS = transportation infrastructure asset management system.

31 • Operating environment of the signal system—for ex- ample, traffic volume, composition, and flow patterns; development affecting traffic growth rate; intersection geometry; pedestrian flows; and variations in these pa- rameters and the degree to which they are predictable. • Signal system performance—for example, operational reliability and downtime, and impact on traffic as mea- sured, for instance, by throughput, travel time and delay, and effect on safety (e.g., number and severity of crashes). • Actions and resources to manage the system—for ex- ample, the range of actions to correct or improve the system encompassing routine operations and mainte- nance, system preservation, repairs, upgrades, and re- placement; and the labor, equipment, material, and financial resources needed to do identified work. Although these capabilities appear to resemble those of other transportation asset management systems, the FHWA document notes several differences in a SSAMS; for exam- ple, the need to recognize a dynamic operating environment; a shorter service life than assets such as pavements and bridges, and a need to manage potential failure of system components; greater systemwide interdependencies among components; and a redirection in the understanding of what constitutes an “asset” and how one characterizes its behav- ior, moving from things like materials properties, physical condition, and structural capacity that are typical of pave- ments and bridges, for example, to concepts of electrical and electronic technology, and operational characteristics and performance. The FHWA document also provides a sum- mary of interviews with the Minnesota DOT (MnDOT) and the Wisconsin DOT on their experiences with signal system management and operations (Harrison et al. 2004). Analytic Modeling: Group Relamping As an example of computerized analytic methods applied to signal system asset management, Zwahlen et al. (2004) have applied group relamping concepts to incandescent lamps in traffic signals, using data provided by the Ohio DOT. A re- lamping model balances the costs of replacing a group of lamps at one time, before they have failed, versus the bene- fits to the maintenance crew of making one trip to a location rather than multiple trips each time a bulb has failed. The management parameter requiring decision is the relamping interval. If the interval is very much shorter than the expected service life, many bulbs will be replaced while they still have considerable life left, resulting in waste and high cost. As the relamping interval approaches the expected service life, the amount of wasted lamp capacity is much reduced and the ef- ficiencies of one trip versus many trips by maintenance crews take hold, driving costs down. As the replacement interval continues to be lengthened, the number of bulb failures be- fore group replacement increases, requiring a greater number of individual trips by maintenance crews for emergency spot replacement, again increasing costs. There is thus an optimal relamping interval at which the costs of replacing signal bulbs are minimal. Zwahlen et al. have created an Excel® spreadsheet to compute the minimum cost solution—that is, the optimal relamping interval—as a function of several maintenance, intersection, and cost variables. Their study used incandes- cent bulbs rated for 1-year, 7,000 h of service (Ohio DOT considers one year of service as 8,760 h). Actual bulb sur- vival percentages by month were developed over a 24-month study period in Ohio District 4, which were then input to the analysis spreadsheet. The group replacement analysis showed that the optimal relamping interval for this case was 10 months. The sensitivity analysis that can also be done with the spreadsheet showed that the results were not that sensi- tive to changes in bulb performance, maintenance productiv- ity, or cost variables. It should be noted that the study results are an example only; District 4 has since moved to other lamps that comply better with Ohio DOT performance spec- ifications (Zwahlen et al. 2004). It should also be noted that the costs considered in this spreadsheet are limited to agency costs: that is, the materials costs of the lamps; the costs of performing relamping in both the group replacement and the emergency spot replacement situations; and the cost of travel by the maintenance crew between the maintenance facility and an intersection, and between intersections during the day. KNOWLEDGE GAPS AND RESEARCH NEEDS Synthesis Survey Comments Agencies at all levels identified a number of knowledge gaps and resulting needs for research. These comments have been organized by topic area and paraphrased here. Data on Field Performance A number of agencies identified the need for basic data on signal asset management, particularly service-life and per- formance data gained from actual field experience, and a way to organize these data in useable form. [We need to get] accurate field information about the condition of equipment on the street. — Michigan DOT (MDOT) [We need] a comprehensive signals inventory maintenance data- base to track repair and maintenance of the major components of a traffic signal installation. — North Carolina DOT [We need to know how] many signals are owned by the agency versus how many are maintained by the agency; where are they located; annual maintenance cost per signal; [and] physical condition of all signals. How many of the agency’s signals have been re-timed within the last 3 years? When was the last time that the timing at Signal X was updated?

Adding additional loading to an existing pole can be difficult because structural information is not tracked. We need to assign a structure number for each pole, installation date for the struc- ture, and the standard drawing with revision that was used to per- form the installation. — Oregon DOT [What is needed is] the real condition of the structures of traffic signals. — Quebec Ministry of Transport [What is needed is the] life estimate of a signal. — Colorado DOT Region 3 It would be nice to have some information on average service lives of components of traffic signals. — Kansas DOT [What is needed is a] definition of service-life level; e.g., when is a steel structure replacement needed as compared to use of longevity-increasing remedial applications. — Maryland State Highway Administration (SHA) [What is needed is] knowledge of the functional obsolescence of electronic components rather than the life of the hardware. Various options for detection—e.g., preformed inductive loops, saw-cut loops, video, etc. [—also need study]. — Utah DOT [What is needed is a] knowledge of quality of materials used. — City of Tampa, Florida Nature of Asset Respondents believe that, to some degree, the lack of knowl- edge regarding asset management of traffic signals relates to the character of these assets as compared with other trans- portation infrastructure, and to the need for stronger inter- actions with component manufacturers. Asset managers are usually accustomed to large highway main- tenance equipment, not small electronic devices associated with signals. — Ohio DOT [We need] integration of proprietary traffic-signal-controller- unit manufacturer’s data into generic agency databases and programs. — Maryland SHA Advances in Materials and Technology Agencies also referred to advances in the materials and devices that are incorporated within signal systems. However, they had different takes on this issue and its implications for asset management. Several identified a need to understand this issue better. [What is needed are] service lives for the various elements in a signalized intersection, especially since the materials have changed over time. — City of Portland, Oregon 32 There is no method in place for the agency to account for the ef- fect of materials evolution on service life. — Ohio DOT How to account for changes in materials quality? We don’t. — Colorado DOT Region 1 Signals are getting installed at an alarming rate. We are trying to keep up with technology and maintenance of these addi- tions to our inventory. Calculating life cycles is not a high priority. — Colorado DOT Region 4 [What is needed is a] study of the service life and reliability of the newer 2070 controller architecture and fiber optic communi- cations network. — North Carolina DOT Other agencies contended that they already account for technological change, often as part of intersection projects, and typically involving the professional judgment of agency staff. Responses are just estimates of how long the agency expects these components to last without being damaged in some other way. It is known that materials are becoming increasingly durable and re- liable, and those characteristics are taken into account. — Kansas DOT The agency’s estimated longevity of signal structural elements has remained consistent at 30 years by the routine use of steel poles/spans or steel mast arms/poles. Wiring and associated signal heads, with a service life greater than 30 years, would be replaced as part of a structural replacement effort. Further, at-grade, in-pavement loop detection would also be replaced as part of the noted structural replacement effort with current detection strategies (video cameras for presence detection and non-invasive detection for upstream detection). — Maryland SHA We rely on our field personnel to identify any issues regarding the service life of any particular piece of equipment. — MDOT Improvements in technology are observed and factored in based on actual experience. — New Mexico DOT Change in materials quality [is] accounted for by a program re- placement based on the life of the materials. — Quebec Ministry of Transport Change in materials quality [is] accounted for by professional judgment. — City of Portland, Oregon We place the older ones on a priority basis and upgrade those components as part of maintenance and/or system upgrades. — Colorado DOT Region 4 Organizational and Procedural Aspects Some agencies focused on needs for organizational or pro- cedural changes to manage their signal systems better.

33 The major gap [in the existing process] is [the need to assign] someone in the agency to apply asset management principles to the traffic signals that we have on the State Highway System. — Kansas DOT [There is a need] to have structures inspected to [a] set idea of condition, especially older units, to help establish [a] time line [of their service life]. — Colorado DOT Region 3 [There is a need] to generate some hard numbers [developed from actual experience]. [For example:] This item will last this long. Maintenance is described as keeping the system operational. If a bulb burns out you replace it. A service call record is kept. — Colorado DOT Region 4 We do not use service life. Materials are replaced when they fail. We have installations and facilities that are in excess of 40 years old with no plans to replace... . . . Issue is not one of gaps in knowledge. There is no plan, direction, or system in place here to manage traffic [signal] assets. — City of Tampa TRB Millennium Papers The TRB Millennium Paper on traffic signal systems sum- marized the history of traffic signal technological develop- ment and considered current and future needs within a broad context (Bullock and Urbanik 2000). To some degree the technical challenges and emerging issues and opportunities identified in this paper echoed the findings of the NTOC survey discussed earlier, stressing the need to look at traffic signals at a broad systems level, in addition to understanding performance at the levels of individual signal components and signal clusters. Technical Research Needs The Millennium Paper on traffic signals identified several technical areas in which research would add useful knowl- edge (Bullock and Urbanik 2000). • System integration—Although current system products perform well when considered in isolation, they do not necessarily make effective components of an integrated, interoperable system comprising products of several vendors. – Work is needed to reconcile different adaptive con- trol models within a standard architecture. – New sensor technology is needed to estimate queue lengths; detect trains, nonferrous bicycles, and pedes- trians; and sense environmental conditions such as weather and air quality. – This new sensor technology must also be able to pass new information to the control system; for example, bus number and passenger loading for transit priority algorithms. – Standards will be needed to integrate new sensors within existing signal systems, and improved relia- bility and lower cost must be achieved for wider mar- ket acceptance. – The different methodologies and predictive capabili- ties of macroscopic and microscopic traffic models need to be reconciled to provide consistent guidance and gain the confidence of potential users. • Improved design practice—An accepted reference model for signal performance must be agreed on to enable practitioners to evaluate alternatives in traffic signal design and controller settings. This advance would greatly improve current design practice, which now often relies on individual technological preferences. • Coordination of research—Research on signal systems is now conducted by many public and private entities, lead- ing to fragmentation of effort and difficulty in achieving the integrated, interoperable systems described earlier. A new initiative will be needed to coordinate the many research efforts so that increasingly complex signal sys- tems can advance successfully. Emerging Issues and Opportunities The TRB Millennium Paper on signal systems also identi- fied broader issues and opportunities that will shape the evolution of more advanced signal systems (Bullock and Urbanik 2000). • Transportation organizations responsible for signal systems will need to revise their priorities from serving primarily automobile traffic to meeting the broader transportation needs of various categories of users. • These updated organizational objectives will be com- plicated by jurisdictional and institutional issues, a process likely more challenging than even the technical advances discussed earlier. Agencies will need to work together, a broader constituency for signal systems will need to be engaged, and a wider, more advanced set of expected services will need to be provided. • Agencies will need to educate the public on the tech- nical complexities of signal systems, the uncertain- ties inherent in predicting and responding to future traffic demands, and the resulting importance of long-term investments in transportation management and operations. • The wider market for signal systems created by improvements in service may also provide the wider customer base needed for funding support, but will require public outreach and education. Another TRB Millennium Paper on vehicle user charac- teristics identified the need for human factors research regarding drivers’ understanding of, and reactions to, the

different ways in which signal systems operate. Different jurisdictions apply different combinations and phasings of signals for certain traffic movements, particularly protected left-hand turns, which can confuse drivers. Research is needed to identify these problems and recommend solutions (Ranney et al. 2000). 34 The TRB Pedestrians Committee identified two signals- related topics among its top 16 research problem statements: (1) Optimizing Signal Timing for Pedestrians, and (2) Evalua- tion of MUTCD Signing, Markings, and Traffic Signals for People with Visual Impairments, Children, and Elderly Adults (Transportation Research Circular E-C084 . . . 2005).