Coordinated Intersections
Requirements:
· Data: You are to coordinate the four signalized intersections along the W. Temple Ave. The geometry and turning movement data is given at the end of this document.
· Tasks: Perform all optimization steps for designing actuated and coordinated signals. Experiment with as many as possible different combinations of settings for each step, and justify the particular settings applied in your design.
· Software: Both Synchro and SimTraffic will be used in this tutorial.
Actuated Signal Coordination in Synchro
Actuated signal coordination design in Synchro is based on its build in optimization functions. Figure 1 illustrates the major steps for performing the signal coordination design in Synchro.
Figure 1 Actuated Signal Coordination Design Steps in Synchro
Some basic Synchro terminologies associated with these steps are first listed as follows.
Compatible Cycle Length is the same cycle length, double the cycle length, or half the cycle length of a given cycle length.
Master Intersection is the intersection with an offset of zero. In Synchro there is zero or one master intersection for each compatible cycle length used in the network.
Zone is a subsystem of a given network partitioned according to certain criteria. Synchro allows intersections to be assigned to zones. Zones are useful for analyzing parts of a network. Zones are also useful to keep the timings for some intersections constant while the timings for other intersections are changed.
Coordinatability Factor (CF) is a measure of the desirability of coordinating the intersections. CF is used with the network partitioning optimization. Intersections with CF values above the threshold CF are placed in the same zone or signal system. Several criteria are used in an attempt to determine whether coordination is warranted. These criteria are used to determine a CF on a scale from 0 to 100 or more. Any score above 80 indicates that the intersections must be coordinated, while any score below 20 indicates that the coordination is not desirable.
Performance Index (PI) is a measure for determining the performance of designed signal timing plans (including the coordination). PI is computed as a weighted summation of delay, vehicle stops, and queue penalty. Therefore, a lower PI value is more desirable. PI is used to compare the performance of different coordination designs. For a particular design, a MOE report including the PI value can be obtained by choosing the “File - Create Report…” command and then the “Summary Network MOEs” option in the following dialog, as shown in Figure 2.
Figure 2 Save Network MOEs
By clicking the “Save Text” button in Figure 2, an ASCII file will be created with desirable MOEs listed, similarly as those shown below.
Measures of Effectiveness
11/06/2018
Network Totals
Number of Intersections 4
Control Delay / Veh (s/v) 13
Queue Delay / Veh (s/v) 0
Total Delay / Veh (s/v) 13
Total Delay (hr) 44
Stops / Veh 0.52
Stops (#) 6022
Average Speed (mph) 19
Total Travel Time (hr) 114
Distance Traveled (mi) 2112
Fuel Consumed (gal) 152
Fuel Economy (mpg) 13.9
CO Emissions (kg) 10.64
NOx Emissions (kg) 2.07
VOC Emissions (kg) 2.47
Unserved Vehicles (#) 0
Vehicles in dilemma zone (#) 0
Performance Index 60.5
Fixed Cycle Coordinated System is the actuated signal coordination in Synchro. The name “fixed cycle” comes from the fact that a fixed background cycle length is required for all coordinated signals in a given zone or the entire network.
The four optimization steps are described as follows.
Step 1. Individual Intersection design
The first step is to make timing plans for each individual intersection. This includes coding the entire network based on the geometry data, and designing lane groups and inputting the volume data for each intersection. Then, the timing plan for each signal is designed and optimized individually. In other words, each intersection will be designed and optimized separately in the first step. This will provide basic cycle length and split information for later optimization steps.
Step 2. Network Partitioning
The next step is to divide the coded network into subsystems (zones). This step is optional. It is up to the engineering judgment to decide whether to partition a given network. This step will produce one or multiple zones for which later optimization steps, e.g. cycle length, offsets, splits, etc., can be performed relatively independently. The partitioning will be based on the CF value computed automatically by Synchro for each link. All signals in one zone will have compatible cycle lengths. Note that zones are not absolutely independent with each other since Synchro tends to perform optimizations among zones whenever it is possible.
Step 3. Optimization of Network Cycle Length
A system cycle length, also called the background cycle length, is crucial for actuated signal coordination. Therefore, the third step is to determine a system cycle length for each zone. Synchro can set up an optimal cycle length for each zone with the best PI value calculated.
Step 4. Optimization of Offsets and Lead-lag Phasing
After determining the system cycle length for each zone, the last step is to optimize offsets and Lead-lag phasing. In this step, Synchro will perform offset and phasing optimizations for signals both within and across zones.
Synchro online help provides detailed descriptions on the above four steps. You are encouraged to read these materials before actually designing coordinated signals in Synchro. In the rest of this tutorial description, we will go through all the steps using a small example provided by Synchro.
Individual Intersection Design:
The example network contains eight intersections as shown in Figure 3.
Figure 3 Experiment Site
Figure 4 Geometry of Example Network
After coding the network, the next step is to input the turning volume data which is depicted in Figure 5. The table at the end of this tutorial is given with these values. Note that the lane group information is also included in this figure. While designing the timing of each individual intersection, be sure to double-click it and choose the “Control Type” as Actuated-Coordinated or as shown “Actd-Coord”. Also pay attention to the “Offset Settings” for each intersection. The actual value of “Offset” can be determined later by the network wide optimizations. Four options are available for the “Reference Style” which is actually determined by the controller used in the field. In this example, “Begin of Green” is used. This refers to the time that a phase turning green. Another setting is the “Reference Phase” which determines the coordinated phases for the actuated signals. Normally, the directions with the heaviest volumes need coordination. For the example, we have a major arterials, W. Temple Ave. For signals spans WB and EB along the avenue are to be coordinated. Notice that the coordinated phases are generally Phase 2 and 6 (why?).
Figure 5 Volume Data of Example Network
The last step is to optimize each intersection separately. This includes optimizing the cycle length and green splits. Click on the intersection and then choose “Optimize - Intersection - Cycle Length” and “Optimize - Intersection - Split” to optimize the cycle length and splits of the intersection, respectively. The intersection offset can be optimized in later steps.
Network Partitioning:
The network partition will try to divide the entire network into subsystems. Choose “Optimize - Partition Network”, a dialog will pop us as shown in Figure 5-2. You are allowed to choose the coordinatability factor (CF) threshold for the partitioning. Higher value normally results in more partitions. How to choose a proper value is again determined by the engineering judgment. In our example, we simply use CF = 50. For this tutorial, it is recommended that all four options are experimented and the one with the best performance (PI) is chosen.
Figure 5-2 Partition Network Dialog
By clicking the “OK” button, Synchro starts to find the optimal zone partition. The result for the example network can be shown in Figure 6. Note that the zone symbol “A” can be enabled by choosing the button in Home - Display Results.
Figure 6 Network Partition Result
Sometimes, the partition generated by Synchro might not be appropriate based upon local jurisdiction. In this case, one can manually partition the network by assigning each intersection to a proper zone. This can be done by double-clicking on the intersection and change the “Zone” property in the popped up menu. For example, if we know that isolating the signals on the Valley Blvd. and Pomona Blvd. into a different zone would be more beneficial, then we can group signals on them into zone B, as shown in Figure 7.
Figure 7 Alternative Network Partition Scenario
Optimization of Cycle Length:
After partitioning the network, the next step is to optimize the cycle length for each zone. This can be done by choosing the “Optimize - Network - Cycle Lengths” command, as shown in Figure 8.
Figure 8 Cycle Length Optimization (First Pass)
In the dialog in Figure 8, one can enter the minimum and maximum cycle lengths to evaluate. The optimizer will evaluate every cycle length between the minimum and maximum at increment intervals. For example, if values are set to 40, 100, and 10, the optimizer will evaluate cycle lengths of 40, 50, 60, 70, 80, 90, and 100. The “Allow Uncoordinated” option will recommend some intersections to be uncoordinated. The number in parentheses is the threshold CF. When “Allow Half Cycle Length” is checked, half cycle length will also be evaluated for some intersections to improve the delay. The “Preserve Files For Each Cycle Length” option will save a file for each evaluated cycle length. These files can be loaded afterwards for evaluation purposes. The files are given the name “filename-***.syn”, where filename is the name of the Synchro file and *** is the cycle length. For the “Offset Optimization”, either “Quick”, or “Medium”, or “Extensive” option can be selected. Finally, a zone name will be chosen for the optimization. Then if the “Automatic” button is clicked, Synchro will automatically select the best cycle length for the subject zone. On the other hand, if the “Manual” button is chosen, Synchro will create a table of cycle lengths with MOEs listed. The user can choose the best cycle length with the lowest PI value.
The evaluation of each cycle length takes time and it is thus recommended that a two step optimization is performed. The first pass evaluates a wide range of cycle lengths with a large increment. Also the “Quick” offset optimization can be used in this pass. Normally, the first pass will be performed in the “Manual” mode. From the first pass, the best cycle length can be narrowed down to a small range. Then the second pass can focus on the optimization of the best cycle length in this small range with small increment. The offset optimization is usually set to “Medium” or “Extensive” and the “Automatic” mode can be used in this pass.
For the example network, if signals are grouped into only one zone (A) and the dialog in Figure 8 shows options for the first pass, the cycle length table as shown in Figure 9 will be generated for this pass.
Figure 9 Cycle Length Table from Pass 1
By comparing the PI values for different cycle lengths, it is easy to observe that the best cycle length is around 50 seconds. Then the next pass will evaluate cycle lengths from 40 to 60 using 2 as the increment, as shown in Figure 10.
Figure 10 Cycle Length Optimization (Second Pass)
To illustrate the results of the second pass, we use the Manual mode. The resulting cycle length table is depicted in Figure 11. Clearly, cycle length 52 has the minimum PI value and hence is the best cycle length found. Note that the textbox in the right bottom of the figure (marked using a circle) shows the number of uncoordinated and half cycle intersections. For our example, there is no uncoordinated intersection nor signal that has half cycle length, i.e., 26 seconds that is too short to handle the traffic.
Figure 11 Cycle Length Table from Pass 2
Optimization of Offsets and Lead-lag Phasing:
The final step of network optimization is to optimize offsets and phasing. This step should be performed after the cycle length has been determined and by choosing the “Optimize - Network - Offsets” command. Figure 12 demonstrates the dialog for selecting options for optimizing the offset.
Figure 12 Optimize Network Offsets
The splits and lead/lag phasing can be optimized by choosing the “Optimize” and “Optimize Lead/Lag Phasing” option, respectively. The offset optimization speed controls how many passes are performed and the step size of each pass. The speed can vary from “Quicker” to “Best Timing Plans”, as shown in Figure 12. Finally, the offset optimization is usually performed for the entire network which can be set in the “Scope” option. The offset optimization will produce a list of possible improvements, as those shown below, if the delay can be reduced by at least 1800 vehicle seconds per hour.
Consider changing the phase order at Node 3: E Pine ST & Freeman RD
Node #3 Phase:1 WBL, Change to a lagging phase.
Node #3 Phase:5 EBL, Change to a leading phase.
Total Delay is reduced by 10295 vehicle seconds per hour!
Node #3 Intersection delay is reduced by 8395 veh-s/hr!
Node #3 NBR Delay is reduced from 28.2 to 20.3 seconds per vehicle!
Node #3 WBT Delay is reduced from 9.7 to 4.0 seconds per vehicle!
Node #15 EBL Delay is reduced from 28.1 to 25.7 seconds per vehicle!
Then corresponding adjustments to the original signal plan can be performed based on the recommended improvements. In our example network, no significant improvement was found and therefore no adjustment is necessary.
Simulation in SimTraffic:
The coordinated signal plans resulted from the above four steps can be simulated in SimTraffic. The purpose of simulation is to further verify whether all signals work properly such that no heavy congestion is incurred by the designed timing plans.
Interpret your Results and Draw Some Conclusions:
After finishing all above steps, you are now ready to answer the following questions. Write down your answers succinctly.
1. Which intersection, if any, is set as the master intersection in your design? Why?
2. How the zones are defined in your design? Are they produced by Synchro automatically or you manually defined them? If the former, what is the threshold CF value that you applied? If the latter, what are your criteria for the definitions?
3. What are the settings for performing the cycle length optimization (for both Pass 1 and Pass 2)? Why do you think these settings are reasonable?
4. Is there any recommended improvement generated by the offset optimization step? If yes, how did you apply them to improve your original design?
5. Is there any queue on any approach when you simulate your designed signal timing plans in SimTraffic? If yes, how to further improve the signal design to reduce or eliminate the queue?