Abstract — Traffic signal optimization is a lower-cost alternative to fixing traffic congestion than traditional roadway widening, geometry changes, and capacity upgrades. Optimizing an existing road fits with UDOT’s strategic goal to “optimize mobility” through innovated traffic management strategies. Modern software allows for realistic modeling of existing and projected traffic conditions. In this study, Synchro will be used to explore the performance of existing green time settings and offsets along the 900 East corridor in Provo, Utah. This study demonstrates how software can improve level or service and reduce wasted green time. The report includes description of existing conditions, study procedure, results, recommendations and conclusion.


This study intended to recommend optimization for 900 East in Provo during southbound afternoon rush-hour traffic. The study will use Trafficware’s Synchro software and will compare software optimization with existing settings. This report will examine the site’s existing conditions, tell how the study was conducted, offer performance data for several scenarios, and will offer recommendations and conclusions.

Description of the Site

This study examines an important north-south arterial street in Provo. It is the most eastern multi-lane arterial road before reaching the city’s edge at the foothills of the Wasatch Mountains. Provo 900 East is a five-lane arterial with two lanes in each direction (north and sound) with a two-way turning lane in the center. The street serves as an eastern connector to Brigham Young University and as a key mobility road for eastern central Provo and foothill neighborhoods. While the road itself is not a state highway, it does connect US-89 in southern Provo with a city-owned spur of SR-265 which connects Brigham Young University to Interstate 15. According to the Utah Department of Transportation (UDOT), the road carries approximately 22,000 vehicles per day.

The study area examines Provo 900 East bounded by a signal at Center Street to the south and a signal at Temple View Lane to the north. The study area includes eight (8) signals, which include 450 North, 700 North, 900 North, Campus Lane, Heritage Drive, University Parkway, and Temple View Drive (see Figure 1). Intersections turning types include protected lefts, permitted/protected lefts, and permitted lefts. There are no one-way segments along the study corridor, nor any segments with roadway geometry that varies from the five-lane configuration except for a double-left from NB 900 East to WB University Parkway. Please note the map is oriented with east to the top.

Table 1 – Reference List of Studied Intersections

8 7 6 5 4 3 2 1
Temple View Univ. Parkway Heritage Drive Campus Lane 900 North 700 North 450 North Center Street
Figure 1 – Map of Study Area
Figure 1 – Map of Study Area

Study Procedure

In order to optimize the signals along 900 East, calculations are required to determine appropriate green, yellow, and all-red time and setting an appropriate ratio for each direction (north-south versus east-west). This uses turning count data that was collected by class groups in September. Teams of two stood along the street corners during afternoon rush hour. These teams counted the number of left, thru, and right turning vehicles over a period of 75 minutes. They also included pedestrian and bicycle counts.

All calculations were performed using Trafficware’s Synchro software. A background image of the study corridor was loaded into the software and calibrated to represent real-world distance. As links were added to the maps, they were positioned to fit the background image. This automatically set the links to appropriate real-world lengths which are crucial for modeling purposes. Roadway geometry was added to reflect the appropriate lane sizes, the number of lanes, and turning configuration at each intersection. Turning counts were added as well as the number of conflicting pedestrians. Signals were configured to represent protected, permitted-protected, or permitted left turns.

The existing signal timing program was entered into the software. These reflected documents generated for each intersection. Exploring the schedule would identify which schedule plan to follow. That would lead to a day plan and an appropriate split ratio, which was entered. Finally, the actual phase plan could be identified and entered. This included minimum initial green time, walk signal time, flashing “don’t walk” time, vehicle extension time (minimum gap), yellow time, and all-red time. A minimum split was hand calculated from the given minimum initial plus yellow plus all-red times and was also entered into the software.

From there, the file was divided into four versions. One retained simultaneous offsets, all of which were set to zero. A second version of the file had Center Street set as a master controller intersection. Provo’s existing signal offsets were entered into the other intersections, offsetting each signal by a number of sections based upon Center Street. Then a third has some human-made adjustments. A fourth, based on the second version, was optimized by the Synchro software.

Study Results

Each of the four versions was measured for travel time and level of service (LOS) to determine performance. These versions included simultaneous offsets, existing offsets, human optimization, and computer optimization.

Mistakes in Original File

The instructor offered a quality-control checkpoint before road performance was checked to make

certain data was entered properly. There were a number of issues that needed to be corrected. Every intersection needed conflicting pedestrian counts added. Dual entry mode was falsely set on several intersections and needed to be deactivated. Recall mode was set incorrectly on several intersections. A few gray bands on the green/red band indicated some split data was entered incorrectly. This, too, was corrected.

Scenario #1: Simultaneous Offsets

The initial scenario takes into account the existing split settings for each intersection, as well as preset yellow, all-red, and pedestrian walk times. No master intersection is set. All offsets are set to 0 seconds, which makes the initial green happen simultaneously though they don’t remain simultaneous as each intersection’s split time varies. Arterial performance was calculated using metrics generated in Synchro.

Time-space diagram

Figure 2 - Time-Space Diagram for Scenario 1
Figure 2 – Time-Space Diagram for Scenario 1

Travel Time, Delay, and LOS

Under this scenario, the travel time for a southbound-travelling vehicle is 6:47 seconds. Northbound-travelling vehicles take 5:40 to travel the corridor. The intersection at 700 North fails, campus drive gets an E, but most intersections seem to perform within a serviceable manner.

Table 2 – Level of Service Under Scenario 1

Scenario 1 LOS SB Speed SB Delay SB LOS NB Speed NB Delay NB
1. Center Street C 18.4 33.3 D 16.9 20.7
2. 450 North C 19.2 14.1 B 27.1 6.4
3. 700 North B 24.9 4.9 F 9.5 22.5
4. 900 North C 19.7 11.7 B 25.5 4.0
5. Campus Lane E 12.6 13.8 B 25.8 3.9
6. Heritage Drive C 18.9 35.5 C 21.7 9.6
7. Univ. Pkwy. D 18.0 19.4 B 24.5 11.8
8. Temple View D 17.3 18.9 D 15.1 15.3
Overall C 18.6 151.6 C 21.3 94.2

Scenario #2: Existing “Provo City”-set Offsets

This scenario uses the same data as Scenario #1 but sets offsets that the city’s consultant has preset. This reflects the status quo.

Time-space diagram

Figure 3 - Time-Space Diagram for Scenario 2
Figure 3 – Time-Space Diagram for Scenario 2

Travel Time, Delay, and LOS

Under this scenario, travel time remains around 6:44 for southbound traffic, but surprisingly northbound travel time increases by nearly one minute to 6:23. The intersection at 700 North still fails.

Table 3 – Level of Service Under Scenario 2

Scenario 2 LOS SB Speed SB Delay SB LOS NB Speed NB Delay NB
1. Center Street B 24.0 13.1 D 16.9 20.7
2. 450 North D 15.3 24.4 B 24.6 13.2
3. 700 North C 18.0 16.1 F 8.5 26.5
4. 900 North D 17.6 15.9 C 20.4 10.5
5. Campus Lane D 15.1 9.6 C 20.7 10.9
6. Heritage Drive C 19.9 30.5 C 21.7 9.7
7. Univ. Pkwy. D 18.0 19.4 C 23.4 14.7
8. Temple View D 17.3 18.9 C 20.6 15.3
Overall C 18.8 147.9 C 20.2 121.5

Scenario #3: Hand Optimization

This scenario employs Synchro’s time-space diagram function (Figure 4) which allows a human user to “eyeball” and manually adjust green offsets. In this specific case, only southbound commuting traffic was looked at and northbound was ignored. Green splits and other times remain the same.

Time-space diagram

Figure 4 - Time-Space Diagram for Scenario 3
Figure 4 – Time-Space Diagram for Scenario 3

Travel Time, Delay, and LOS

Under this scenario, travel time for southbound-travelling vehicles reduces by over half a minute from the status quo (Scenario #2) to 6:05. Northbound reduces to 6:19. This is a much better-performing scenario than what is currently status quo.

Table 4 – Level of Service Under Scenario 3

Scenario 3 LOS SB Speed SB Delay SB LOS NB Speed NB Delay NB
1. Center Street C 23.0 15.9 D 16.9 20.7
2. 450 North C 19.7 13.1 B 28.2 3.9
3. 700 North B 25.6 4.2 F 7.2 33.9
4. 900 North B 28.4 1.3 C 22.6 7.4
5. Campus Lane D 15.8 8.6 C 19.0 14.0
6. Heritage Drive C 20.5 28.0 B 24.8 5.3
7. Univ. Pkwy. D 18.0 19.4 C 22.6 17.2
8. Temple View D 17.3 18.9 C 20.6 15.3
Overall C 20.8 109.4 C 20.4 117.7

Scenario #4: Computer Optimization

This scenario reflects Synchro’s ability to use software algorithms to adjust signals to be mathematically ideal. Unlike Scenario #3, the computer will balance not only offsets but signal time splits as well. It will also consider northbound traffic as well as southbound when changing offsets.

Time-space diagram

Figure 5 - Time-Space Diagram for Scenario 4
Figure 5 – Time-Space Diagram for Scenario 4

Travel Time, Delay, and LOS

Under this scenario, southbound improves to 6:00 and northbound dramatically improves to 5:29.

Table 5 – Level of Service Under Scenario 4

Scenario 4 LOS SB Speed SB Delay SB LOS NB Speed NB Delay NB
1. Center Street C 19.4 28.7 B 24.9 4.2
2. 450 North C 19.1 14.3 B 26.8 7.1
3. 700 North B 27.7 2.0 E 13.9 11.4
4. 900 North B 26.6 2.9 C 19.3 12.5
5. Campus Lane E 13.1 12.9 C 21.8 9.1
6. Heritage Drive C 23.5 16.5 C 22.5 8.4
7. Univ. Pkwy. C 18.8 17.2 B 25.0 10.6
8. Temple View C 20.8 10.6 E 13.2 20.0
Overall C 21.0 105.1 C 22.0 83.3


The existing settings are not set nearly ideal enough for what is possible for southbound and northbound traffic. Under even non-programmed offsets (Scenario #1), which have everything set simultaneously at 0 seconds, southbound performs nearly equally to the status quo (Scenario #2) and the non-programmed actually performs quite a bit better than the offset program (Figure 6). This is disappointing to see during an afternoon rush hour when southbound traffic is at high volume.

Under Scenario #3, very little attention was paid to northbound travel and all focus was spent on manually caring for offsets to benefit southbound travel. It such a situation, it seems reasonable to expect the drop in travel time (and increase in performance) for southbound vehicles. What came as a surprise was the slight improvement for northbound vehicles. One would expect a spike in travel time as northbound vehicles are “sacrificed” for southbound travel, but that didn’t seem to occur. It stands as an indictment to the poorly-set existing offsets (Scenario #2).

The computer appears to perform the best, which comes as little surprise. Not only did it improve southbound performance, but it offered a huge performance advantage to northbound traveling vehicles, proving that it can be possible to benefit both directions of travel without having to wreck one to benefit the other. It may perhaps have something to do with the block distances seeming to mirror each other, as University Parkway to Heritage Drive mirrors the Center Street to 450 North long block; 450 North to 700 North mirrors the short blocks between Heritage Drive and Campus Drive; Campus Drive to 900 North mirrors 700 North to 900 North – although actual block measurements don’t back up that assertion directly.

It is recommended that Provo City’s engineering staff consider implementing Synchro-based offsets to see if they can better utilize green time along the 900 East corridor. It is also very possible that the Synchro entry this model is based on is faulty, which could explain why the status quo (Scenario #2) appears to perform so poorly. Further study is recommended before implementing any changes.