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.
Chapter 9. Advanced Signal Systems CHAPTER 9 ADVANCED SIGNAL SYSTEMS CONTENTS 9.1 SYSTEMS ENGINEERING ...................................................................................................... 9-1 9.1.1 Concept of Operations ............................................................................................................. 9-1 9.1.2 System Requirements ............................................................................................................. 9-2 9.1.3 Design and Implementation ................................................................................................. 9-2 9.1.4 VeriÂication Plan ........................................................................................................................ 9-2 9.1.5 Validation Plan ........................................................................................................................... 9-3 9.2 ADVANCED COORDINATION FEATURES ....................................................................... 9-3 9.2.1 Actuating the Coordinated Phase ....................................................................................... 9-3 9.2.2 Dynamic Phase Length ............................................................................................................ 9-3 9.2.3 Phase Re-Service ....................................................................................................................... 9-4 9.3 TRAFFIC RESPONSIVE PLAN SELECTION SYSTEMS ................................................... 9-4 9.3.1 TrafÂic Responsive Algorithms ............................................................................................ 9-4 9.3.2 TrafÂic Responsive Detection ............................................................................................. 9-10 9.4 ADAPTIVE SIGNAL CONTROL TECHNOLOGY SYSTEMS .......................................... 9-11 9.4.1 Agency Objectives ................................................................................................................... 9-11 9.4.2 Operational Characteristics ................................................................................................ 9-12 9.4.3 Installation and ConÂiguration ........................................................................................... 9-15 9.4.4 Maintenance and Operations ............................................................................................. 9-16 9.4.5 VeriÂication and Validation of Effectiveness ................................................................ 9-17 9.5 REFERENCES ......................................................................................................................... 9-18 Signal Timing Manual, Second Edion
Chapter 9. Advanced Signal Systems LIST OF EXHIBITS Exhibit 9-1 System Requirements Categories .......................................................................... 9-2 Exhibit 9-2 Example Trafic Responsive Thresholds ............................................................. 9-7 Exhibit 9-3 Example Trafic Responsive Detector Locations ............................................. 9-8 Exhibit 9-4 Example Offset Targets (Based on Arterial Trafic Flow) ............................ 9-8 Exhibit 9-5 Example Cycle/Splits Targets (Based on Critical Intersection Volumes)........................................................................................................................... 9-9 Exhibit 9-6 Example Pattern Assignments Based on Types of Trafic Flow and Timing Plans ........................................................................................................... 9-9 Exhibit 9-7 Example Pattern Signatures .................................................................................. 9-10 Exhibit 9-8 ASCT System Objectives .......................................................................................... 9-12 Signal Timing Manual, Second Edion
Chapter 9. Advanced Signal Systems LIST OF EXHIBITS Exhibit 9-1 System Requirements Categories .......................................................................... 9-2 Exhibit 9-2 Example Trafic Responsive Thresholds ............................................................. 9-7 Exhibit 9-3 Example Trafic Responsive Detector Locations ............................................. 9-8 Exhibit 9-4 Example Offset Targets (Based on Arterial Trafic Flow) ............................ 9-8 Exhibit 9-5 Example Cycle/Splits Targets (Based on Critical Intersection Volumes)........................................................................................................................... 9-9 Exhibit 9-6 Example Pattern Assignments Based on Types of Trafic Flow and Timing Plans ........................................................................................................... 9-9 Exhibit 9-7 Example Pattern Signatures .................................................................................. 9-10 Exhibit 9-8 ASCT System Objectives .......................................................................................... 9-12 Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-1 CHAPTER 9. ADVANCED SIGNAL SYSTEMS While earlier chapters largely address âstandardâ signal timing features, which are fairly similar across all modern trafic signal controllers, Chapter 9 describes advanced signal systems. These systems are able to make signal timing adjustments based on detection information, thus modifying operations during varying trafic low conditions. Applications can range from the use of advanced coordination features to the deployment of adaptive signal control technology (ASCT). While there are many variations, most advanced systems require a considerable investment from local agencies. As such, this chapter begins by describing a process known as âSystems Engineeringâ that can be used to determine whether an advanced signal system is most appropriate for a given location based on local needs and requirements. The following sections focus on several types of advanced signal systems that can be applied, including advanced coordination features, trafic responsive systems, and ASCT systems. 9.1 SYSTEMS ENGINEERING The main objectives of the Systems Engineering process are identifying and deining an agencyâs needs in concrete terms so that a signal technology can be selected that meets those needs. It is recommended that an agency completely assess the capabilities of its existing system before considering alternatives. This ensures that the capabilities of the current system are maxed out before an upgrade is considered. Most advanced signal systems are a considerable investment, so this process is intended to minimize risk for an agency throughout the design and implementation of a new system. Systems Engineering ultimately helps an agency develop ive reports that document the decision-making process: ⢠Concept of Operations (ConOps) ⢠System Requirements ⢠Design and Implementation ⢠Veriication Plan ⢠Validation Plan Systems Engineering was initially developed to design and implement large-scale projects, so the level of effort required to complete the full method is signiicant. The practitioner should tailor the level of application to the size and/or complexity of the project. The intent of this section is to provide a high-level overview of the Systems Engineering process, highlighting some of the key elements. For additional information and detailed guidance, please see Systems Engineering for Intelligent Transportation Systems (1) and Model Systems Engineering Documents for Adaptive Signal Control Technology Systems (2). 9.1.1 Concept of Operaons The Concept of Operations (ConOps) is a non-technical document that deines a systemâs needs (i.e., what problem needs to be addressed) and, more importantly, describes how the system will be used. The project needs in this document will While Systems Engineering is oen used to assess advanced signal systems, the process can be applied in many situaons (e.g., evaluang preferenal treatment opons). Signal Timing Manual, Second Edion
9-2 Chapter 9. Advanced Signal Systems eventually be connected to technical requirements; all elements of a project are ultimately deined by and referenced back to the ConOps. It is the foundation upon which the rest of the Systems Engineering process builds. The ConOps is written to, and is the responsibility of, the stakeholders of the project. Typically, stakeholders include anyone impacted by the project and may encompass owners, operators, maintainers, and users. Identifying appropriate stakeholders, clearly articulating project needs, and obtaining consensus on the project needs are important activities in this stage of Systems Engineering. 9.1.2 System Requirements The System Requirements document is a technical report written for technical staff, vendors, users, and system operators. The elements within this document link directly back to speciic needs in the ConOps and describe what needs to be achieved with the project. The requirements do not, however, describe how the system will be built. Exhibit 9-1 summarizes various requirement categories that should be considered, sourced from Model Systems Engineering Documents for Adaptive Signal Control Technology Systems (2). Requirement Category Descripon Funconal Requirements What the system is to do Performance Requirements How well it is to perform Non-Funconal Requirements Under what condions it will perform Enabling Requirements What other acons must be taken in order for the system to become fully operaonal Constraints Limitaons imposed on the design by agency policies and pracces, such as type of soÂware, type of equipment, and external standards Interface Requirements Definions of the interfaces between subsystems or with external systems Data Requirements Definions of data flow between subsystems or with external systems Composition of the System Requirements document is the responsibility of the same stakeholders responsible for the ConOps. During development of the System Requirements document, stakeholders must select requirements as well as analyze, document, validate, and manage the requirement statements (within the System Requirements document and how they are referenced to the ConOps). 9.1.3 Design and Implementaon Up to this point in the Systems Engineering process, the activities have focused on deining the problem to be solved but not on the solution. This is the purpose of the design activity. This portion of the process may include conducting alternatives analysis and product evaluations, documenting interfaces and standards, and developing detailed designs and speciications. Implementation of the project involves activities associated with installation, transition, and delivery of the project components. 9.1.4 Verificaon Plan Once the project is designed and implemented, a Veriication Plan is used to conirm that all requirements (outlined in the System Requirements document) have been met. The Veriication Plan is typically a technical report written for technical staff, vendors, users, and system operators. It describes how the system will be tested and documents Exhibit 9-1 System Requirements Categories Signal Timing Manual, Second Edion
9-2 Chapter 9. Advanced Signal Systems eventually be connected to technical requirements; all elements of a project are ultimately deined by and referenced back to the ConOps. It is the foundation upon which the rest of the Systems Engineering process builds. The ConOps is written to, and is the responsibility of, the stakeholders of the project. Typically, stakeholders include anyone impacted by the project and may encompass owners, operators, maintainers, and users. Identifying appropriate stakeholders, clearly articulating project needs, and obtaining consensus on the project needs are important activities in this stage of Systems Engineering. 9.1.2 System Requirements The System Requirements document is a technical report written for technical staff, vendors, users, and system operators. The elements within this document link directly back to speciic needs in the ConOps and describe what needs to be achieved with the project. The requirements do not, however, describe how the system will be built. Exhibit 9-1 summarizes various requirement categories that should be considered, sourced from Model Systems Engineering Documents for Adaptive Signal Control Technology Systems (2). Requirement Category Descripon Funconal Requirements What the system is to do Performance Requirements How well it is to perform Non-Funconal Requirements Under what condions it will perform Enabling Requirements What other acons must be taken in order for the system to become fully operaonal Constraints Limitaons imposed on the design by agency policies and pracces, such as type of soÂware, type of equipment, and external standards Interface Requirements Definions of the interfaces between subsystems or with external systems Data Requirements Definions of data flow between subsystems or with external systems Composition of the System Requirements document is the responsibility of the same stakeholders responsible for the ConOps. During development of the System Requirements document, stakeholders must select requirements as well as analyze, document, validate, and manage the requirement statements (within the System Requirements document and how they are referenced to the ConOps). 9.1.3 Design and Implementaon Up to this point in the Systems Engineering process, the activities have focused on deining the problem to be solved but not on the solution. This is the purpose of the design activity. This portion of the process may include conducting alternatives analysis and product evaluations, documenting interfaces and standards, and developing detailed designs and speciications. Implementation of the project involves activities associated with installation, transition, and delivery of the project components. 9.1.4 Verificaon Plan Once the project is designed and implemented, a Veriication Plan is used to conirm that all requirements (outlined in the System Requirements document) have been met. The Veriication Plan is typically a technical report written for technical staff, vendors, users, and system operators. It describes how the system will be tested and documents Exhibit 9-1 System Requirements Categories Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-3 the activities and results of the veriication process. The Veriication Plan does not provide detailed descriptions of data collection or analysis techniques, as that is part of the validation process (described next). The Veriication Plan is typically the responsibility of the system developer, vendor, or supplier, and is overseen by the agency. 9.1.5 Validaon Plan The Validation Plan is written after the veriication process is complete and describes how the performance of the system will be measured. The plan deines the measures of effectiveness (MOEs), the data requirements and collection procedures, and the type of analysis required to validate the system against all needs identiied in the ConOps. It also documents validation activities, results, and any necessary corrective actions. The Validation Plan is typically developed by the agency and is written for the stakeholders of the project. 9.2 ADVANCED COORDINATION FEATURES In most cases, standard coordinated signal timing plans suficiently meet the needs of a corridor or network. However, there are times when standard features are no longer able to respond eficiently to trafic demand or luctuating patterns. When these situations occur, a common response is to invest in new signal technology, such as ASCT. While an ASCT system may beneit signal operations, the practitioner should irst investigate available advanced coordination features and detection upgrades that may address local problems without the substantial investment required by an ASCT system. This section highlights advanced coordination features typical of most signal systems, but it is not intended to be comprehensive. It is an introduction to select features, and highlights situations where these coordination features are both available and applicable. Some trafic signal controllers will have advanced coordination features not discussed in this section. 9.2.1 Actuang the Coordinated Phase Actuating the coordinated phase (or fully-actuated coordination) is a signal timing treatment that actuates a portion of the coordinated phase split. Using this feature, the coordinated phases are permitted to gap out during the later portion of their split time in the event of low demand. This allows additional time to be given to the minor street and left-turn phases. Additional information about this particular advanced coordination feature is provided in Chapter 7, as its application is becoming more typical. 9.2.2 Dynamic Phase Length Dynamic phase length (also known as extended split or adaptive split) allows controller irmware to alter the length of a phase through dynamic movement of the force-off point. This feature allows effective use of unused green time within a cycle and can reduce the queue lengths experienced at an intersection. Note that this feature does not alter cycle length or shorten the coordinated phase(s), so it will not disrupt coordinated operations. The nomenclature and funconality of advanced features differs from vendor to vendor. Signal Timing Manual, Second Edion
9-4 Chapter 9. Advanced Signal Systems In order for an intersection to beneit from the dynamic phase length feature, a system must be capable of estimating vehicle demand. Vehicle demand can be measured and forecast using queue information from vehicle detectors. It is important to ensure that detection is functional and passage time properly set (particularly at setback detectors located upstream of queued vehicles) when using dynamic phase length. 9.2.3 Phase Re-Service The phase re-service feature (also known as repeated phase service or conditional service) allows a phase to be served more than once during a cycle. Most beneits of phase re-service are realized during low to medium trafic volume periods, such as the shoulder of peak periods or midday periods. Phase re-service can also be a useful treatment at minor intersections in a coordinated system where the system cycle length is longer than what is needed to serve local trafic demand (and half-cycling is not an option). Engineering judgment and ield observation are essential when determining whether phase re-service will beneit a system. 9.3 TRAFFIC RESPONSIVE PLAN SELECTION SYSTEMS Trafic responsive plan selection systems are able to change coordination plans to it trafic conditions. Based on detector data returned from the ield, trafic responsive algorithms select a preconigured signal timing plan from a library of plans. The current timing plan must usually have been running for a minimum amount of time before a new plan can be implemented, and the new timing plan must typically be a certain percentage improvement over the currently running plan. If those conditions are met, the signals in a group are directed to begin using the new plan at the same time, in order to ensure that coordination is reestablished quickly. Most, if not all, trafic signal systems allow trafic responsive plan selection to be implemented on a time-of-day basis. In this case, the schedule includes a baseline plan that is initiated during a corresponding time-of-day period. If detection data indicate that the scheduled plan is responding ineficiently to the observed trafic patterns, the system switches the scheduled plan to a more suitable alternative. Some systems allow the trafic responsive mode to be directed manually or to be scheduled as a special event override of the normal time-of-day schedule. Signal timing plans for trafic responsive systems are typically conigured to cover a wide range of ield scenarios. It is not necessary to have a fully comprehensive set of timing plans for trafic responsive systems to operate. However, all timing plans and pattern data must be conigured in the controller prior to ield operations. Trafic responsive plan selection systems do not calculate new plans (or patterns). 9.3.1 Traffic Responsive Algorithms There are two primary categories of trafic responsive algorithms that are used to select timing plans: (1) algorithms that use targets and (2) algorithms that use thresholds. Target-based methods are typically available for NEMA and 170/2070 irmware, while threshold-based methods are typically only available for NEMA controller irmware. Modern NEMA controllers now support both plan- and pattern- based operations. Most traffic signal system vendors offer a so ware module (whether standard or oponal) that will operate a traffic responsive plan selecon system. Signal Timing Manual, Second EdiÂon
9-4 Chapter 9. Advanced Signal Systems In order for an intersection to beneit from the dynamic phase length feature, a system must be capable of estimating vehicle demand. Vehicle demand can be measured and forecast using queue information from vehicle detectors. It is important to ensure that detection is functional and passage time properly set (particularly at setback detectors located upstream of queued vehicles) when using dynamic phase length. 9.2.3 Phase Re-Service The phase re-service feature (also known as repeated phase service or conditional service) allows a phase to be served more than once during a cycle. Most beneits of phase re-service are realized during low to medium trafic volume periods, such as the shoulder of peak periods or midday periods. Phase re-service can also be a useful treatment at minor intersections in a coordinated system where the system cycle length is longer than what is needed to serve local trafic demand (and half-cycling is not an option). Engineering judgment and ield observation are essential when determining whether phase re-service will beneit a system. 9.3 TRAFFIC RESPONSIVE PLAN SELECTION SYSTEMS Trafic responsive plan selection systems are able to change coordination plans to it trafic conditions. Based on detector data returned from the ield, trafic responsive algorithms select a preconigured signal timing plan from a library of plans. The current timing plan must usually have been running for a minimum amount of time before a new plan can be implemented, and the new timing plan must typically be a certain percentage improvement over the currently running plan. If those conditions are met, the signals in a group are directed to begin using the new plan at the same time, in order to ensure that coordination is reestablished quickly. Most, if not all, trafic signal systems allow trafic responsive plan selection to be implemented on a time-of-day basis. In this case, the schedule includes a baseline plan that is initiated during a corresponding time-of-day period. If detection data indicate that the scheduled plan is responding ineficiently to the observed trafic patterns, the system switches the scheduled plan to a more suitable alternative. Some systems allow the trafic responsive mode to be directed manually or to be scheduled as a special event override of the normal time-of-day schedule. Signal timing plans for trafic responsive systems are typically conigured to cover a wide range of ield scenarios. It is not necessary to have a fully comprehensive set of timing plans for trafic responsive systems to operate. However, all timing plans and pattern data must be conigured in the controller prior to ield operations. Trafic responsive plan selection systems do not calculate new plans (or patterns). 9.3.1 Traffic Responsive Algorithms There are two primary categories of trafic responsive algorithms that are used to select timing plans: (1) algorithms that use targets and (2) algorithms that use thresholds. Target-based methods are typically available for NEMA and 170/2070 irmware, while threshold-based methods are typically only available for NEMA controller irmware. Modern NEMA controllers now support both plan- and pattern- based operations. Most traffic signal system vendors offer a so ware module (whether standard or oponal) that will operate a traffic responsive plan selecon system. Signal Timing Manual, Second EdiÂon Chapter 9. Advanced Signal Systems 9-5 9.3.1.1 Target-Based Systems Target-based systems transition to new signal timing plans based on volumes detected in the ield. These systems require a practitioner to program the timing plans that will be applied during certain trafic conditions (deined by target volumes at each detector). The systems then calculate the difference between real-time trafic volumes and the target volumes (or âsignaturesâ) preselected for each detector. Based on the relationship between the real-time conditions and predetermined signatures, the central system or master controller applies the timing plan that is most appropriate for the ield conditions. Typically, target values for each detector are expressed as equivalent hourly volumes. Because detector data yield a combination of volume and occupancy data, Equation 9-1 must be applied to compare real-time volumes to predetermined target volumes. The timing plan (and, more speciically, the associated signature) that produces the lowest value in Equation 9-1 is compared to the value produced by the currently running plan. If the yielded value is lower than a percentage of the current value, the new timing plan is implemented. where err = difference between ield volume and target volume (vehicles per lane per hour), N = number of detectors, i = ield detector, j = target detector, w = weighting factor for detector, V = volume (vehicles per lane per hour), K = weighting factor for detector occupancy (vehicles per lane per hour per percent occupancy, typically a global parameter for all detectors), and O = detector occupancy (percent). In free-low conditions, the occupancy measured by a system detector is often low and does not contribute to the calculation. However, when traffic conditions are congested and queues form on the detector, detector volume measurements will be constant and much lower than actual trafic demand. Without the use of the KO factor in this scenario, the algorithm logic inaccurately assumes that volumes are low. The KO factor provides a way to scale the occupancy measurement to an equivalent volume. A rule of thumb is simply to use a K value that scales the occupancy (reported as between 0 and 100 percent) into an equivalent hourly volume. For example, a K value of 17 would scale an occupancy value of 100 percent to 1700 vehicles per lane per hour. A K value of this magnitude might be appropriate for detectors at the exit side of an intersection that infrequently experiences substantial occupancy. For detectors closer to the stop bar that experience more frequent queuing, lower K values (i.e., approximately Target-based systems were developed by the Federal Highway Administraon (FHWA) in the early 1970s during the development of the Urban Traffic Control System (UTCS). Equaon 9-1 Detector occupancy is the percentage of the analysis period that a vehicle was stopped on the detector. Signal Timing Manual, Second Edion
9-6 Chapter 9. Advanced Signal Systems between 5 and 7) should be used, since occupancy will frequently average non-zero values. In many cases, detector data are weighted to scale the importance of the information from one detector as compared to that of another. In practice, weights are often used to amplify small changes in volume when a detector does not experience variability as high as other detectors in the system. That being said, there is limited guidance on the conÂiguration of detector weights. 9.3.1.2 Threshold-Based Systems Threshold-based systems calculate which timing plan (or signal timing parameters) to select based on Âield detector data crossing certain threshold values. Most threshold- based systems use computation channels (CCs), which are parameters associated with groups of Âield detectors (e.g., a set of inbound and outbound detectors). In general, CC parameters are the sum of weighted detector data and can be calculated using the general form of Equation 9-2. However, each vendor will use different types and numbers of CC parameters, as well as different weights and smoothing factors, and will apply different combinations of maximums and averages to compare to the timing plans (or parameters). The practitioner should refer to the speciÂic manufacturer for details about CC parameter calculations. where CC = computation channel (vehicles per lane per hour), N = number of detectors, i = Âield detector, w = weighting factor for detector, V = volume (vehicles per lane per hour), K = weighting factor for detector occupancy (vehicles per lane per hour per percent occupancy, typically a global parameter for all detectors), and O = detector occupancy (percent). The conÂiguration of weights can be particularly important in order to match the Âield data with appropriate control parameters. A detailed methodology for the conÂiguration of threshold-based systems can be found in Methodology for Determination of Optimal Trafic Responsive Plan Selection Control Parameters (3). This methodology uses statistical analyses (i.e., discriminant analysis and principal component analysis) of the detector data to identify reliable weights for the linear combination of volume and occupancy data. After CC values have been calculated for various groups of detectors, they are compared to the threshold values set by the practitioner. Different CC values will trigger different signal timing plans (or parameters including cycle length, splits, and offsets). Rather than using a single target value, threshold-based systems allow the practitioner to assign CC values that the algorithm will use to both enter into and exit out of new Equaon 9-2 Signal Timing Manual, Second Edion
9-6 Chapter 9. Advanced Signal Systems between 5 and 7) should be used, since occupancy will frequently average non-zero values. In many cases, detector data are weighted to scale the importance of the information from one detector as compared to that of another. In practice, weights are often used to amplify small changes in volume when a detector does not experience variability as high as other detectors in the system. That being said, there is limited guidance on the conÂiguration of detector weights. 9.3.1.2 Threshold-Based Systems Threshold-based systems calculate which timing plan (or signal timing parameters) to select based on Âield detector data crossing certain threshold values. Most threshold- based systems use computation channels (CCs), which are parameters associated with groups of Âield detectors (e.g., a set of inbound and outbound detectors). In general, CC parameters are the sum of weighted detector data and can be calculated using the general form of Equation 9-2. However, each vendor will use different types and numbers of CC parameters, as well as different weights and smoothing factors, and will apply different combinations of maximums and averages to compare to the timing plans (or parameters). The practitioner should refer to the speciÂic manufacturer for details about CC parameter calculations. where CC = computation channel (vehicles per lane per hour), N = number of detectors, i = Âield detector, w = weighting factor for detector, V = volume (vehicles per lane per hour), K = weighting factor for detector occupancy (vehicles per lane per hour per percent occupancy, typically a global parameter for all detectors), and O = detector occupancy (percent). The conÂiguration of weights can be particularly important in order to match the Âield data with appropriate control parameters. A detailed methodology for the conÂiguration of threshold-based systems can be found in Methodology for Determination of Optimal Trafic Responsive Plan Selection Control Parameters (3). This methodology uses statistical analyses (i.e., discriminant analysis and principal component analysis) of the detector data to identify reliable weights for the linear combination of volume and occupancy data. After CC values have been calculated for various groups of detectors, they are compared to the threshold values set by the practitioner. Different CC values will trigger different signal timing plans (or parameters including cycle length, splits, and offsets). Rather than using a single target value, threshold-based systems allow the practitioner to assign CC values that the algorithm will use to both enter into and exit out of new Equaon 9-2 Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-7 timing plans. Exhibit 9-2 illustrates example CC thresholds for the selection of an offset value. In this example, if the CC value is currently 52, then Offset 3 would be the current setting. If the CC value falls below 49, then Offset 2 would be selected, and if the CC value rises above 68, then Offset 4 would be selected. Offset Change CC Enter Value CC Exit Value 1 to 2 25 18 2 to 3 52 49 3 to 4 68 64 4 to 5 75 70 Having separate entry and exit thresholds prevents the system from Âluctuating frequently into and out of timing plans. This limitation is sometimes referred to as a hysteresis, or the delay between change in state (detector data) and the system response (change to signal timing plan). The separate thresholds in a threshold-based system provide the same type of operation as percent improvement targets in a target- based system Some threshold-based systems calculate ratios of detector data for the selection of cycle length, splits, and offsets, referred to as âindices.â A few general principles apply to this method of combining CC parameter information: ⢠The cycle length index is a function of the arterial volume (V+KO). The cycle length increases as the volume detected on both the inbound and outbound detectors increases. ⢠The split index is proportional to the ratio of arterial volume (V+KO) to crossing street volume (V+KO). ⢠The offset index is proportional to the ratio of inbound to outbound volume (V+KO). 9.3.1.3 Traffic Responsive Process Example There are several steps required to set up and conÂigure a trafÂic responsive system. This section walks through an example using a target-based system. The same principles can be applied to threshold-based systems; the practitioner will simply need to deÂine the âintoâ and âout ofâ thresholds. The Âirst step when setting up a trafÂic responsive system is to identify the group of signals that will operate together. Typically, the group of signals is composed of intersections along an arterial that have substantial Âluctuations in trafÂic volumes, requiring a range of cycle times. For this example, it will be assumed that the signalized intersections shown in Exhibit 9-3 experience variable trafÂic Âlow. The next step in the setup process is to establish target volumes for each detector that are associated with speciÂic types of cycle/split/offset combinations. TrafÂic data should be collected for several days at the input and output detector locations (and summarized in daily charts) to determine these target values. Detection stations must be installed at key input and output points. While they are not as critical, detection stations located midway along the arterial can also be conÂigured as system detectors. The detector locations for this example are highlighted in Exhibit 9-3. (Note that the exhibit is oriented so north is at the top of the Âigure.) Exhibit 9-2 Example Traffic Responsive Thresholds Signal Timing Manual, Second Edion
9-8 Chapter 9. Advanced Signal Systems Exhibit 9-4 provides an example of how the relationship between inbound and outbound trafic can be used to identify different types of trafic low patterns (that will be associated with different offset values). For example, the first row of the table describes the signature for a heavy inbound low scenario, while the third row indicates a scenario in which the level of trafic in both directions is similar and moderate. The ratios of inbound to outbound trafic can be used as the trafic responsive system targets. Once an arterial moves from balanced trafic low (with inbound and outbound trafic nearly equal) to a moderate inbound trafic low (with inbound trafic 1.5 times that of outbound trafic), the system is notiied that it is time to select a new timing plan (with different offsets). Type of Offsets (Based on Traffic Flow) Rao of Traffic Volume Detected at Inbound Detector to Traffic Volume Detected at Outbound Detector (Vehicles:Vehicles) Inbound Detector 1: Outbound Detector 1 Inbound Detector 2: Outbound Detector 2 Inbound Detector 3: Outbound Detector 3 Heavy Inbound 2:1 2:1 2:1 Moderate Inbound 1.5:1 1.5:1 1.5:1 Balanced 1:1 1:1 1:1 Moderate Outbound 1:1.5 1:1.5 1:1.5 Heavy Outbound 1:2 1:2 1:2 Exhibit 9-3 Example Traffic Responsive Detector Loca ons Exhibit 9-4 Example Offset Targets (Based on Arterial Traffic Flow) Signal Timing Manual, Second Edion
9-8 Chapter 9. Advanced Signal Systems Exhibit 9-4 provides an example of how the relationship between inbound and outbound trafic can be used to identify different types of trafic low patterns (that will be associated with different offset values). For example, the first row of the table describes the signature for a heavy inbound low scenario, while the third row indicates a scenario in which the level of trafic in both directions is similar and moderate. The ratios of inbound to outbound trafic can be used as the trafic responsive system targets. Once an arterial moves from balanced trafic low (with inbound and outbound trafic nearly equal) to a moderate inbound trafic low (with inbound trafic 1.5 times that of outbound trafic), the system is notiied that it is time to select a new timing plan (with different offsets). Type of Offsets (Based on Traffic Flow) Rao of Traffic Volume Detected at Inbound Detector to Traffic Volume Detected at Outbound Detector (Vehicles:Vehicles) Inbound Detector 1: Outbound Detector 1 Inbound Detector 2: Outbound Detector 2 Inbound Detector 3: Outbound Detector 3 Heavy Inbound 2:1 2:1 2:1 Moderate Inbound 1.5:1 1.5:1 1.5:1 Balanced 1:1 1:1 1:1 Moderate Outbound 1:1.5 1:1.5 1:1.5 Heavy Outbound 1:2 1:2 1:2 Exhibit 9-3 Example Traffic Responsive Detector Loca ons Exhibit 9-4 Example Offset Targets (Based on Arterial Traffic Flow) Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-9 The cycle length and splits that are selected will often be based on traf ic characteristics at the critical intersection (in this example, Detector 2 detects the critical intersection major street volumes and Detector 4 detects the minor street volumes). Example target values for major and minor street traf ic volumes are illustrated in Exhibit 9-5. In this example, if both the major street and minor street detectors indicate low, approximately equal traf ic volumes, a timing plan with a shorter cycle length and balanced splits will be selected. If both the major street and minor street experience high traf ic volumes, but the volume on the major street exceeds that of the minor street, then a timing plan with a longer cycle length and splits that favor the major street will be selected. Type of Cycle and Splits Traffic Volume Detected at Crical Intersecon (Vehicles) Major Street Volume (Inbound and Outbound) Minor Street Volume (Northbound and Southbound) Short Cycle/Direconal Splits 800 400 Short Cycle/Balanced Splits 800 800 Medium Cycle/Direconal Splits 1600 800 Medium Cycle/Balanced Splits 1600 1600 Long Cycle/Direconal Splits 3200 1600 Long Cycle/Balanced Splits 3200 3200 These two matrices are merged (as shown in Exhibit 9-6) to correlate trafic patterns with speciic timing plans. In other words, a speciic combination of cycle/splits/offsets is assigned a pattern number. Each pattern is deined by the target volumes at each system detector. Example pattern signatures are illustrated in Exhibit 9-7 for Patterns 1, 8, 19, 22, and 30. Type of Cycle and Splits Type of Offsets Heavy Inbound Moderate Inbound Balanced Moderate Outbound Heavy Outbound Short Cycle/Direconal Splits Pa ern 1 Pa ern 2 Pa ern 3 Pa ern 4 Pa ern 5 Short Cycle/Balanced Splits Pa ern 6 Pa ern 7 Pa ern 8 Pa ern 9 Pa ern 10 Medium Cycle/Direconal Splits Pa ern 11 Pa ern 12 Pa ern 13 Pa ern 14 Pa ern 15 Medium Cycle/Balanced Splits Pa ern 16 Pa ern 17 Pa ern 18 Pa ern 19 Pa ern 20 Long Cycle/Direconal Splits Pa ern 21 Pa ern 22 Pa ern 23 Pa ern 24 Pa ern 25 Long Cycle/Balanced Splits Pa ern 26 Pa ern 27 Pa ern 28 Pa ern 29 Pa ern 30 As a inal step in the coniguration process, these signatures are entered into the signal system module, and the algorithm then selects the pattern that most closely matches the real-time, smoothed detection data from the system detectors. Most practitioners have had better success with trafic responsive systems when they deined a reasonable number of intersections and timing plans. Care must be taken to set up parameters appropriately and ensure that detection systems are functional. With proper calibration and set up, trafic responsive operations can provide signiicant reductions in travel time and delay by better matching the signal timing to the underlying trafic patterns. Exhibit 9-5 Example Cycle/Splits Targets (Based on CriÂcal IntersecÂon Volumes) Exhibit 9-6 Example Pa ern Assignments Based on Types of Traffic Flow and Timing Plans Signal Timing Manual, Second Edion
9-10 Chapter 9. Advanced Signal Systems Pa ern Informaon Pa ern Signatures Pa ern 1 Pa ern 8 Pa ern 19 Pa ern 22 Pa ern 30 Type of Paern Heavy Inbound Volume Requiring Short Cycle/ DirecÂonal Splits Balanced Volumes Requiring Short Cycle/ Balanced Splits Moderate Outbound Volume Requiring Medium Cycle /Balanced Splits Moderate Inbound Volume Requiring Long Cycle/ DirecÂonal Splits Heavy Outbound Volume Requiring Long Cycle/ Balanced Splits Rao of Inbound Volume to Outbound Volume on Major Street 2:1 1:1 1:1.5 1.5:1 1:2 Crical Intersecon Major Street Volume (Inbound and Outbound) 800 800 1600 3200 3200 Cri cal Intersec on Minor Street Volume (Northbound and Southbound) 400 800 1600 1600 3200 Detector 1 Targets Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 2 Targets (Cri cal Intersec on Major Street) Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 3 Targets Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 4 Targets (Cri cal Intersec on Minor Street) Northbound Volume 200 400 800 800 1600 Southbound Volume 200 400 800 800 1600 9.3.2 Traffic Responsive Detec on Trafic responsive algorithms rely on detector data from âsystemâ detectors to determine which timing plan or pattern to implement. Detectors must be located reasonably upstream from the stop bar and should be very short (typically zones measuring 6 feet by 6 feet) to reduce the effect of trafic queues. Detectors are frequently located on the exit side of an intersection and are separated by lane in order to yield a more accurate assessment of trafic conditions. Tube counters or other temporary detection equipment may be used to determine coniguration parameters before permanent system detection stations are installed. Knowledge of local conditions is beneicial when determining detection station placement. While functioning detectors are crucial to the effectiveness of a trafic responsive system, many systems allow plan selection calculations to be completed even when some of the detection stations are not functional or are not reporting valid data. These failed detector outputs either yield a measurement of zero, or are simply ignored, so the calculations are not affected. If too many detectors fail, however, the trafic responsive system will be disabled. Exhibit 9-7 Example PaÂern Signatures Signal Timing Manual, Second Edi on
9-10 Chapter 9. Advanced Signal Systems Pa ern Informaon Pa ern Signatures Pa ern 1 Pa ern 8 Pa ern 19 Pa ern 22 Pa ern 30 Type of Paern Heavy Inbound Volume Requiring Short Cycle/ DirecÂonal Splits Balanced Volumes Requiring Short Cycle/ Balanced Splits Moderate Outbound Volume Requiring Medium Cycle /Balanced Splits Moderate Inbound Volume Requiring Long Cycle/ DirecÂonal Splits Heavy Outbound Volume Requiring Long Cycle/ Balanced Splits Rao of Inbound Volume to Outbound Volume on Major Street 2:1 1:1 1:1.5 1.5:1 1:2 Crical Intersecon Major Street Volume (Inbound and Outbound) 800 800 1600 3200 3200 Cri cal Intersec on Minor Street Volume (Northbound and Southbound) 400 800 1600 1600 3200 Detector 1 Targets Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 2 Targets (Cri cal Intersec on Major Street) Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 3 Targets Inbound Volume 530 400 640 1920 1070 Outbound Volume 270 400 960 1280 2130 Detector 4 Targets (Cri cal Intersec on Minor Street) Northbound Volume 200 400 800 800 1600 Southbound Volume 200 400 800 800 1600 9.3.2 Traffic Responsive Detec on Trafic responsive algorithms rely on detector data from âsystemâ detectors to determine which timing plan or pattern to implement. Detectors must be located reasonably upstream from the stop bar and should be very short (typically zones measuring 6 feet by 6 feet) to reduce the effect of trafic queues. Detectors are frequently located on the exit side of an intersection and are separated by lane in order to yield a more accurate assessment of trafic conditions. Tube counters or other temporary detection equipment may be used to determine coniguration parameters before permanent system detection stations are installed. Knowledge of local conditions is beneicial when determining detection station placement. While functioning detectors are crucial to the effectiveness of a trafic responsive system, many systems allow plan selection calculations to be completed even when some of the detection stations are not functional or are not reporting valid data. These failed detector outputs either yield a measurement of zero, or are simply ignored, so the calculations are not affected. If too many detectors fail, however, the trafic responsive system will be disabled. Exhibit 9-7 Example PaÂern Signatures Signal Timing Manual, Second Edi on Chapter 9. Advanced Signal Systems 9-11 Trafic responsive methods typically smooth detector data through time-based weighted averages in order to prevent an excessive number of timing plan changes caused by short-term surges or reduced trafic volumes. Smoothing the detection data reduces the ability of algorithms to respond quickly to surges in trafic, so the application of trafic responsive systems is most effective when changes in trafic low are both large and persistent (e.g., 30 minutes or more). A smoothing factor of 50 percent is commonly identiied as a reasonable setting. 9.4 ADAPTIVE SIGNAL CONTROL TECHNOLOGY SYSTEMS ASCT is a signal system technology that uses detection data and algorithms to adjust signal timing parameters for current conditions. Unlike trafic responsive plan selection systems, which use predetermined timing plans, ASCT is able to adjust various timing parameters (within certain constraints) based on what the traffic requires. Over time, trafic evolvesâwhether quickly in minutes or slowly in days, months, or yearsâand renders signal timing plans increasingly less effective. Traditional signal operations are often unable to adjust to this variability, resulting in degradation of trafic performance. Trafic responsive systems may require a large library of plans to address highly variable trafic conditions. Conceptually, ASCTâs continual adjustment of signal timing parameters provides incremental beneits over time-of-day plans. It is important to understand that ASCT systems are not âset-and-forgetâ systems. They require ongoing ine-tuning and higher levels of maintenance than traditional systems, in order to keep the detection and communications infrastructure working at a high level of performance. While ASCT is beneicial in certain applications, there are many other sound trafic engineering principles included in this manual that may improve trafic operations without the expense and complexity associated with ASCT. This section provides an overview of agency objectives that are often applied to ASCT systems, as well as operational characteristics and various factors worth considering during the selection and deployment process. For more information on agency experiences with ASCT systems, refer to NCHRP Synthesis 403 (4). 9.4.1 Agency Objecves A wide range of operational objectives serve as motivation for ASCT deployment, but variable trafic demand is the most common characteristic that agencies want to address through an adaptive system. While advanced coordination and trafic responsive systems can accommodate predictable variability, ASCT systems are able to accommodate less predictable variations in trafic, which may be caused by trafic incidents, special events, or short- or long-term trafic low changes. ASCT systems apply solutions according to speciied objectives embedded in the program. The degree to which an agency is satisied with ASCT operations often depends on how closely the ASCTâs optimization objectives match the agencyâs objectives. Common high-level objectives for the deployment of ASCT are summarized in Exhibit 9-8. It is often the case that different objectives are appropriate at different times of the day and under different trafic conditions. An arterial road that provides access between a freeway and large residential areas but also contains trafic generators (e.g., retail centers or schools) may require a peak-period pipeline objective during most times of the day but a smooth low or access equity objective during business hours and on weekends. Under these conditions, the ASCT may be required to accommodate the Too many transions between ming plans can have an adverse effect on operaons. It is crical to acknowledge that ASCT cannot solve underlying system capacity issues. Signal Timing Manual, Second Edion
9-12 Chapter 9. Advanced Signal Systems transition of objectives at different times of day. Most ASCT systems today do not explicitly include features or conigurations to address this, although some systems may be modiied to transition between objectives with the detection of user-speciied ield conditions. Objecve Descripon Where Objecve Is Typically Applied How Objecve Is Applied Pipeline Minimizing the number of stops experienced by a preferred movement on a criÂcal route (from one end of the corridor to the other). Linear, arterial routes. Less commonly applied to progress heavy turning movements. Maintaining large splits for the coordinated phases. Smooth Flow PrioriÂzing simultaneous bidirecÂonal movements on a criÂcal route in order to maximize throughput. Suburban arterials. Maintenance of large splits for the coordinated phases. Equitable Access Providing sufficient arterial access to traffic generators along a corridor by placing increased emphasis on minor street demand. Areas with significant le- turn and minor street demand (e.g., suburban retail shopping districts). Appropriate applicaÂon of split Âmes to prevent long delays for minor movements (including pedestrians). Manage Queues MiÂgaÂon of queues and congesÂon caused by blocked intersecÂons or movements. Note: Most ASCT systems do not include features that specifically migate queues. LocaÂons where queues block upstream intersecÂons or movements. Constraints on cycle and phase duraÂons to ensure that large platoons progress at appropriate Âmes. âGaÂngâ to store vehicles at locaÂons upstream of criÂcal links may also be required. MiÂgate OversaturaÂon Prevent, delay the onset of, or limit the duraÂon of oversaturated condiÂons. If oversaturated condiÂons do persist, clear overflow queues quickly. LocaÂons with oversaturated movements. Note: ASCT systems can migate some oversaturaon due to short- term capacity limitaons but are not a substute for capacity improvements. AdjusÂng green Âme allocaÂon for saturated phases. Accommodate Long-Term Variability Update signal Âming more frequently than tradiÂonal systems and reduce deterioraÂon of traffic operaÂons over Âme. Areas with changing traffic paÂerns, parÂcularly growing communiÂes. While ASCT systems require some adjustments when major traffic paÂern changes occur, they can adjust automaÂcally to many changes. Manage Events and Incidents Manage surges in traffic (both planned and unplanned). Areas with recurring planned special events (e.g., concerts, sporÂng events, or community acÂviÂes). Adjust signal Âming parameters to match traffic surges over Âme. However, they generally cannot adjust rapidly to large surges. 9.4.2 Operaonal Characteriscs ASCT systems adjust trafic signal settings based on current trafic conditions (or nearly current, as most systems lag changes in demand by one or more cycles). However, available systems vary widely in levels of responsiveness, algorithmic framework, and detection requirements. The practitioner should refer to the Exhibit 9-8 ASCT System ObjecÂves Signal Timing Manual, Second Edion
9-12 Chapter 9. Advanced Signal Systems transition of objectives at different times of day. Most ASCT systems today do not explicitly include features or conigurations to address this, although some systems may be modiied to transition between objectives with the detection of user-speciied ield conditions. Objecve Descripon Where Objecve Is Typically Applied How Objecve Is Applied Pipeline Minimizing the number of stops experienced by a preferred movement on a criÂcal route (from one end of the corridor to the other). Linear, arterial routes. Less commonly applied to progress heavy turning movements. Maintaining large splits for the coordinated phases. Smooth Flow PrioriÂzing simultaneous bidirecÂonal movements on a criÂcal route in order to maximize throughput. Suburban arterials. Maintenance of large splits for the coordinated phases. Equitable Access Providing sufficient arterial access to traffic generators along a corridor by placing increased emphasis on minor street demand. Areas with significant le- turn and minor street demand (e.g., suburban retail shopping districts). Appropriate applicaÂon of split Âmes to prevent long delays for minor movements (including pedestrians). Manage Queues MiÂgaÂon of queues and congesÂon caused by blocked intersecÂons or movements. Note: Most ASCT systems do not include features that specifically migate queues. LocaÂons where queues block upstream intersecÂons or movements. Constraints on cycle and phase duraÂons to ensure that large platoons progress at appropriate Âmes. âGaÂngâ to store vehicles at locaÂons upstream of criÂcal links may also be required. MiÂgate OversaturaÂon Prevent, delay the onset of, or limit the duraÂon of oversaturated condiÂons. If oversaturated condiÂons do persist, clear overflow queues quickly. LocaÂons with oversaturated movements. Note: ASCT systems can migate some oversaturaon due to short- term capacity limitaons but are not a substute for capacity improvements. AdjusÂng green Âme allocaÂon for saturated phases. Accommodate Long-Term Variability Update signal Âming more frequently than tradiÂonal systems and reduce deterioraÂon of traffic operaÂons over Âme. Areas with changing traffic paÂerns, parÂcularly growing communiÂes. While ASCT systems require some adjustments when major traffic paÂern changes occur, they can adjust automaÂcally to many changes. Manage Events and Incidents Manage surges in traffic (both planned and unplanned). Areas with recurring planned special events (e.g., concerts, sporÂng events, or community acÂviÂes). Adjust signal Âming parameters to match traffic surges over Âme. However, they generally cannot adjust rapidly to large surges. 9.4.2 Operaonal Characteriscs ASCT systems adjust trafic signal settings based on current trafic conditions (or nearly current, as most systems lag changes in demand by one or more cycles). However, available systems vary widely in levels of responsiveness, algorithmic framework, and detection requirements. The practitioner should refer to the Exhibit 9-8 ASCT System ObjecÂves Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-13 manufacturer for speciic information about each system. This section summarizes a few operational characteristics that are common across many systems. In particular, systems share the same basic structure: ⢠Detector collection of trafic conditions information, ⢠Calculation of timing plans based on inputs from detection, and ⢠Selection of the next settings to be applied. 9.4.2.1 Adapve Groups Adaptive systems often require the division of an ASCT-controlled network into subsystems (or regions) of intersections that either need to be coordinated or should operate using a common cycle length. This grouping is determined during the design phase. Typical rules for grouping intersections using traditional signal timing also apply to the identiication of groupings for adaptive operations (refer to Chapter 3). Some ASCT systems allow for âcross-coordinationâ on multiple, crossing routes, while others only allow coordination along a single route (e.g., north/south, east/west) by time of day. Some ASCT systems allow a bordering intersection in one subsystem to leave its current subsystem and join a neighboring subsystem according to a suitability factor calculation. These types of settings need to be calibrated during system installation and coniguration. 9.4.2.2 Calculaon Methods ASCT uses two primary methods to adjust signal timing: (1) downloading new settings and (2) overriding the local controller. Some ASCT systems compute new signal timing parameters and download those new settings to the local controller. ASCT systems that download new parameters allow the local controller to use actuated operational features to adjust green splits according to gap out and ixed or loating force-off logic. Alternatively, ASCT systems can override the operation of the local controller through hold, force-off commands, or through controlling the presence or absence of phase calls. Through override methods, some ASCT systems use command messages to communicate the changes to the controller, while others interface to the cabinet directly using either hardwire electrical connections (for Type 332, Type 336, and NEMA TS-1 cabinets) or serial interface units (for NEMA TS-2 and ITS cabinets). ASCT systems that download new timing parameters are typically unconstrained by algorithm processing time, but limit the breadth of their search for better timing to a limited range around the current settings. This is done for two reasons: (1) to limit the effects of signal transition from the current settings to the new settings and (2) to minimize âover-reactionâ to very short-term anomalies in lows. For example, if a current split value is 15 seconds, an ASCT optimization method might consider values from 10 seconds to 20 seconds for the next split value. If the current cycle time is 100 seconds, cycles between 110 seconds and 120 seconds might be considered next (even if a better value were 130 seconds). Some ASCT systems are not as concerned about minimizing the difference between the last phase split and the next phase split because they have complete control of phase durations, and no transitions are necessary. This is sometimes referred to as a rolling- horizon optimization process. These systems constantly re-optimize the next signal settings in reaction to short-term luctuations by trying to predict what the best Signal Timing Manual, Second Edion
9-14 Chapter 9. Advanced Signal Systems possible sequence and duration of phases should be in the next few seconds to minutes. For example, a rolling-horizon system might evaluate hundreds of different options for the duration and sequence of each phase over the next 60 seconds and then implement only the decision trajectory for the next 5 seconds. The process then repeats. Maximum and minimum phase durations are typically enforced much like a controller running in free mode. All ASCT systems have settings that need to be Âine-tuned to customize the level of adaptability. These settings can include maximum and minimum green times for each phase, adherence to pedestrian crossing time requirements (or the ability to ignore pedestrian times if there is no pedestrian demand), the amount each timing parameter may be changed incrementally, the largest and smallest values for the timing parameters (e.g., cycle lengths or offsets), and the timing parameters that the system is prohibited from changing (e.g., phase sequence, ability to skip phases). 9.4.2.3 Detecon Detection installation, conÂiguration, and accuracy are critical to the effective operation of ASCT because the results are used by the optimization algorithms to determine appropriate signal timing settings. However, ASCT systems are able to receive input from various detection technologies (e.g., microwave, magnetometer, inductive loop, and video). Some systems are even integrated with speciÂic types of detection devices. Many ASCT systems may have a preferred (and some a required) detection conÂiguration, but can operate a subset of adaptive algorithms with a different conÂiguration. Generally, there are Âive locations where ASCT detectors may be needed: ⢠Stop bar detectors, ⢠Setback detectors (typically used for decision zone protection and phase extension in actuated operations), ⢠Mid-block detectors, ⢠Upstream detectors located at the upstream intersection (on the downstream side), and ⢠Setback detectors at the entrance points of left- and right-turn storage bays. No ASCT system requires detection at all Âive locations, but detection is generally required on all phases. While many ASCT systems often require lane-by-lane detection on all phases (which provides the best ASCT operations), some are able to accommodate detection that is reported across multiple lanes for each phase. Typically, stop bar detectors estimate the degree of saturation for each phase and measure output for queue length estimation. These detectors often need to be added to coordinated phase lanes if the existing operation is coordinated but not actuating the coordinated phase(s). Shorter stop bar detection zones yield more accurate count and occupancy information; although in practice, newly deployed ASCT systems often accommodate existing stop bar detection zones that are 50 feet or longer. Setback detectors can estimate queue lengths and measure the arrival proÂile of approaching platoons of trafÂic. The farther upstream these detectors are located, the better the quality of the arrival proÂile information. It is worth mentioning that mid- block arrival Âlows entering the link downstream of the detection point will not be The amount of me required for calibrang and fine- tuning models and adapve operaons is directly related to the number of signals that the ASCT manages. ASCT systems oÂen require extensive detecon that must be highly reliable and well maintained. Signal Timing Manual, Second Edion
9-14 Chapter 9. Advanced Signal Systems possible sequence and duration of phases should be in the next few seconds to minutes. For example, a rolling-horizon system might evaluate hundreds of different options for the duration and sequence of each phase over the next 60 seconds and then implement only the decision trajectory for the next 5 seconds. The process then repeats. Maximum and minimum phase durations are typically enforced much like a controller running in free mode. All ASCT systems have settings that need to be Âine-tuned to customize the level of adaptability. These settings can include maximum and minimum green times for each phase, adherence to pedestrian crossing time requirements (or the ability to ignore pedestrian times if there is no pedestrian demand), the amount each timing parameter may be changed incrementally, the largest and smallest values for the timing parameters (e.g., cycle lengths or offsets), and the timing parameters that the system is prohibited from changing (e.g., phase sequence, ability to skip phases). 9.4.2.3 Detecon Detection installation, conÂiguration, and accuracy are critical to the effective operation of ASCT because the results are used by the optimization algorithms to determine appropriate signal timing settings. However, ASCT systems are able to receive input from various detection technologies (e.g., microwave, magnetometer, inductive loop, and video). Some systems are even integrated with speciÂic types of detection devices. Many ASCT systems may have a preferred (and some a required) detection conÂiguration, but can operate a subset of adaptive algorithms with a different conÂiguration. Generally, there are Âive locations where ASCT detectors may be needed: ⢠Stop bar detectors, ⢠Setback detectors (typically used for decision zone protection and phase extension in actuated operations), ⢠Mid-block detectors, ⢠Upstream detectors located at the upstream intersection (on the downstream side), and ⢠Setback detectors at the entrance points of left- and right-turn storage bays. No ASCT system requires detection at all Âive locations, but detection is generally required on all phases. While many ASCT systems often require lane-by-lane detection on all phases (which provides the best ASCT operations), some are able to accommodate detection that is reported across multiple lanes for each phase. Typically, stop bar detectors estimate the degree of saturation for each phase and measure output for queue length estimation. These detectors often need to be added to coordinated phase lanes if the existing operation is coordinated but not actuating the coordinated phase(s). Shorter stop bar detection zones yield more accurate count and occupancy information; although in practice, newly deployed ASCT systems often accommodate existing stop bar detection zones that are 50 feet or longer. Setback detectors can estimate queue lengths and measure the arrival proÂile of approaching platoons of trafÂic. The farther upstream these detectors are located, the better the quality of the arrival proÂile information. It is worth mentioning that mid- block arrival Âlows entering the link downstream of the detection point will not be The amount of me required for calibrang and fine- tuning models and adapve operaons is directly related to the number of signals that the ASCT manages. ASCT systems oÂen require extensive detecon that must be highly reliable and well maintained. Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-15 captured by the ASCT, so care must be taken to locate setback detection appropriately on a link-by-link basis. It is also important to locate setback detection away from driveways so that vehicles exiting the street do not introduce error into the ASCT calculations. 9.4.2.4 Communicaons Communication among the detectors, controllers, and ASCT system is vital. Some ASCT systems interact with detection directly, either (1) through direct hardwire integration with the devices or (2) through wireless polling of stand-alone detection. Others rely on communication with the local controller to retrieve detection status. Many systems can accommodate a combination of directly connecting to detection and receiving detection status from the local controller. A detailed understanding of existing communications capabilities and ASCT needs is essential when evaluating alternative systems and their overall cost. Communications infrastructure has been a challenging requirement for ASCT in the past, although the emergence of Internet Protocol (IP) technology in trafÂic operations has improved the deployment options. (IP communications can either be achieved through wireless technology or through hardwire technology.) 9.4.2.5 Phase Configuraon The practitioner should be aware of the phase conÂiguration limitations of the ASCT system. Some systems can currently only operate with eight-phase, dual-ring controller architecture, while other technologies can accommodate more than eight phases. Many ASCT systems also use stage-based operation instead of phase-based operation, requiring conÂiguration of all allowable combinations of phases as stages (e.g., phase pairs 1 and 6, 1 and 5, 2 and 5, 2 and 6, etc. are conÂigured as separate stages). 9.4.3 Installaon and Configuraon The installation of an ASCT system can be, but does not have to be, a lengthy and difÂicult process. The condition and suitability of the existing infrastructure (i.e., detection, hardware, and communications) are the most critical elements of the deployment. Common activities during installation and conÂiguration of an ASCT system are described below. 9.4.3.1 ASCT Equipment Installaon ⢠Installation of any special ASCT equipment in existing trafÂic cabinets or construction and installation of new cabinets as needed. ⢠Installation and conÂiguration of IT systems, databases, servers, and computer peripherals for ASCT operation. 9.4.3.2 ASCT System Configuraon ⢠Installation and conÂiguration of a plan and schedule that is both realistic and tolerant of technology-related issues. ⢠ConÂiguration and calibration of ASCT adaptive and model parameters. ⢠IdentiÂication and conÂiguration of adaptive on and off times, if desired. Signal Timing Manual, Second Edion
9-16 Chapter 9. Advanced Signal Systems ⢠Con iguration of interaction with third-party systems such as preemption, transit priority, and pedestrian countdown timers. 9.4.3.3 Detecon Installaon ⢠Installation of new detection or recon iguration of existing hardware by quali ied contractors. ⢠Separation of detection lead-in cables into lane-by-lane operations. ⢠Addition of detector cards, detector racks, video system processors, and other detection equipment. ⢠Veri ication of all detection operations. 9.4.3.4 Communicaons Installaon ⢠Installation of new communications or upgrade of existing communications media between traf ic signals and from signals in the ield to a central or master location. ⢠Con iguration of communications equipment (i.e., modems, switches, irewalls, and central server computers). ⢠Veri ication of all communications operations. 9.4.4 Maintenance and Operaons ASCT systems are more operationally demanding and require more agency support than traditional traf ic signal systems. As discussed in NCHRP Synthesis 403, ASCT systems rely on experienced staff that understand and can adjust operations (4). The need for training and ongoing operational attention of agency staff needs to be recognized in the early stages of ASCT consideration. One solution for an agency may be to obtain operational support from outside sources, including engineering consultants or system vendors. Whether the adaptive system is maintained through contracted support, development of existing staff, or newly hired talent, it is critical that agencies never consider an ASCT system to be a âset-and-forgetâ system that eliminates the need for sound traf ic engineering and expertise. Key activities during the maintenance and operations phase of ASCT deployment are described below. 9.4.4.1 ASCT System Review ⢠Review and analysis of signal timing policies and underlying signal settings. ⢠Frequent review of ASCT decision-making for recalibration. ⢠Frequent analysis of MOEs for adaptive parameter ine-tuning. ⢠Development of backup timing plans. 9.4.4.2 Equipment Maintenance ⢠Continual preventative maintenance checks on ield hardware, communications, and detection systems. ⢠Timely repair of malfunctioning detection and communications hardware. Signal Timing Manual, Second Edion
9-16 Chapter 9. Advanced Signal Systems ⢠Con iguration of interaction with third-party systems such as preemption, transit priority, and pedestrian countdown timers. 9.4.3.3 Detecon Installaon ⢠Installation of new detection or recon iguration of existing hardware by quali ied contractors. ⢠Separation of detection lead-in cables into lane-by-lane operations. ⢠Addition of detector cards, detector racks, video system processors, and other detection equipment. ⢠Veri ication of all detection operations. 9.4.3.4 Communicaons Installaon ⢠Installation of new communications or upgrade of existing communications media between traf ic signals and from signals in the ield to a central or master location. ⢠Con iguration of communications equipment (i.e., modems, switches, irewalls, and central server computers). ⢠Veri ication of all communications operations. 9.4.4 Maintenance and Operaons ASCT systems are more operationally demanding and require more agency support than traditional traf ic signal systems. As discussed in NCHRP Synthesis 403, ASCT systems rely on experienced staff that understand and can adjust operations (4). The need for training and ongoing operational attention of agency staff needs to be recognized in the early stages of ASCT consideration. One solution for an agency may be to obtain operational support from outside sources, including engineering consultants or system vendors. Whether the adaptive system is maintained through contracted support, development of existing staff, or newly hired talent, it is critical that agencies never consider an ASCT system to be a âset-and-forgetâ system that eliminates the need for sound traf ic engineering and expertise. Key activities during the maintenance and operations phase of ASCT deployment are described below. 9.4.4.1 ASCT System Review ⢠Review and analysis of signal timing policies and underlying signal settings. ⢠Frequent review of ASCT decision-making for recalibration. ⢠Frequent analysis of MOEs for adaptive parameter ine-tuning. ⢠Development of backup timing plans. 9.4.4.2 Equipment Maintenance ⢠Continual preventative maintenance checks on ield hardware, communications, and detection systems. ⢠Timely repair of malfunctioning detection and communications hardware. Signal Timing Manual, Second Edion Chapter 9. Advanced Signal Systems 9-17 ⢠Identiication of preferred failure modes if detection, communications, or ASCT equipment fails (e.g., coordination versus free operation). ⢠Database backup and software and hardware maintenance plan development. 9.4.4.3 Training and Vendor Contact ⢠Initial and continued training on system operations (particularly new training for staff added after ASCT deployment). ⢠Identiication of a speciic person as the ASCT vendor point of contact. 9.4.5 Verificaon and Validaon of Effecveness Veriication is necessary if a system is being procured with new features that have not been deployed previously in order to determine whether the agencyâs requirements have been met. This would not be necessary if the procurement is for an existing ASCT that has been previously veriied. Validation is the process of assessing how the deployed system actually performs relative to agency operational objectives and actual trafic conditions. The assessment of ASCT system performance is highly dependent on both the quality of timings in place before ASCT is deployed and the duration of time that occurs between the collection of before data and after data. That is, if timings are updated just before ASCT deployment, the system may experience only minor improvements or, in some situations, signal timing with actuated control may outperform the adaptive system. The performance of adaptive systems may be evaluated through one or more performance measures. The validation process helps to ensure that these performance measures are selected and applied under the guidance of sound trafic engineering practices and principles. Further discussion on validation and MOEs may be found in Every Day Counts: Validate Traf ic Signal Operational ObjectivesâDraft MOE and Evaluation Approach Plan (5). Some common types of validation activities include ⢠State operational objectives for ASCT installation clearly and measure performance with respect to those objectives; match MOEs closely with operational goals. ⢠Validate minor street performance, particularly in lieu of short-lived manual trafic studies that only capture a snapshot of performance. Most ASCT and controller hardware can now continuously provide detailed, high-resolution, event-based signal timing data. ⢠Supplement probe travel time runs with vehicle re-identiication technologies, which can provide continuous measurement of point-to-point travel times. Validation activities using only probe data are limited by the number of runs and the time periods in which they are conducted. ⢠Evaluate ASCT effectiveness before and after peak conditions. In these time periods, ASCT effectiveness may be most dramatic, potentially revealing the ability to shorten the peak period or more quickly dissipate queues. Performance data during peak periods, particularly when conditions are oversaturated, may not indicate superiority of ASCT. ⢠Simulate incident conditions in order to assess how ASCT will manage abnormal conditions. Signal Timing Manual, Second Edion
9-18 Chapter 9. Advanced Signal Systems ⢠Collect before-and-after performance data as close to the same date as possible. Separation of before-and-after data by many months can skew results. Consider an on/off approach where days are alternated with and without ASCT operation. ⢠Collect enough performance data for both on and off periods to estimate the reliability of performance with and without ASCT. ⢠Consider the sensitivity of performance due to changes in volumes. Collect volume data on key routes to identify whether conditions are similar during on and off periods. ⢠Consider that retiming signals just before ASCT deployment will result in less impressive ASCT validation results. 9.5 REFERENCES 1. National ITS Architecture Team. Systems Engineering for Intelligent Transportation Systems. Report FHWA-HOP-07-069, Federal Highway Administration, United States Department of Transportation, 2007. 2. Fehon, K., M. Krueger, J. Peters, R. Denney, P. Olson, and E. Curtis. Model Systems Engineering Documents for Adaptive Signal Control Technology Systems. Report FHWA-HOP-11-027, Federal Highway Administration, United States Department of Transportation, 2012. 3. Abbas, M. M., N. A. Chaudhary, A. Sharma, S. P. Venglar, and R. J. Engelbrecht. Methodology for Determination of Optimal Trafic Responsive Plan Selection Control Parameters. Report FHWA/TX-04/0-4421-1, Texas Department of Transportation, Austin, TX, 2004. 4. Stevanovic, A. NCHRP Synthesis 403: Adaptive Trafic Control Systems: Domestic and Foreign State of Practice. Transportation Research Board of the National Academies, Washington, D.C., 2010. 5. Gettman, D. Every Day Counts: Validate Trafic Signal Operational ObjectivesâDraft MOE and Evaluation Approach Plan. Report FHWA-HOP-12-034, Federal Highway Administration, United States Department of Transportation, 2012. Signal Timing Manual, Second Edion
