1 Introduction

At an urban signalized intersection, a traffic signal timing plan assigns the right-of-way to the conflicting traffic movements to eliminate the potential of collisions at the intersection. However, as traffic volume continuously increases, it has been a common phenomenon that drivers who arrive at the signal at the end of the green period speed up to proceed the intersection. This makes signalized intersections being one of the places with the highest crash rate. According to the Federal Highway Administration (FHWA), about 20% of crashes occurred at signalized intersections, among which rear-end crashes and right-angle crashes are the most common types (National Highway Traffic Safety, A, 2001). Therefore, minimizing the crash risk at signalized intersections has been one of the most critical issues for city planners. In recent decades, signal coordination has been considered as an efficient method to improve the performance of the traffic control system in terms of reducing delay and increasing roadway capacity, especially in high-density urban areas.

On the other hand, as signal coordination has been extensively implemented into urban signalized arterials, safety concerns were raised by both transportation engineers and the public. Since traffic coordination may result in higher mainline speeds than non-coordinated conditions, most of these concerns came up with the worries that a higher speed may increase the risk of involving traffic crashes, particularly injury or fatal crashes. To date, there is neither solid evidence from the field to support the concern, nor theoretical-level models to analyze this issue. In practice, safety is always considered as the top priority, and transportation engineers need to balance the trade-offs between the operational benefits and safety risks introduced by traffic signal coordination, which is also of significant importance to eliminate the public’s concerns about the safety issues caused by signal coordination. In this regard, this research effort aims at assessing the effects of traffic signal coordination on the safety performance of urban arterials. To achieve this research goal, this research developed a TranSync-VISSIM-SSAM microsimulation modeling framework, where TranSync (Transintelligence, 2020) was adopted as the offline signal optimization tool, VISSIM as the microsimulation modeling testbed, and the Surrogate Safety Assessment Model (SSAM) (Pu et al., 2008) was employed for safety performance assessment. To provide practical results, this research employed three real-world urban arterials with various speed limits as case study sites.

The remainder of this paper is organized as follows: first, a literature review on the state-of-the-practice of safety research in signal coordination; after that, description of the testbed corridors and simulation models; then, analysis of the safety performance; afterward, discussion on the potential reasons to the simulation modeling results; finally, conclusions of this research and directions for future works.

2 Literature review

The literature review indicates the majority of previous studies on signal coordination were devoted to upgrading the operational efficiency of urban arterials (Yue et al., 2019), a large number of studies have been focused on signal optimization (Yang et al., 2018; Yang et al., 2019), timing strategy (Tian et al., 2004; Tian et al., 2005; Tian & Urbanik, 2007; Yue et al., 2020), fine-tune strategy (Yue, 2020), and performance evaluation (Yue et al., 2021). Whereas, there’s a relatively small volume of literature has been done to investigate the safety impacts. From these studies, there has been a long-lasting debate on signal coordination’s impact on safety. These studies can generally be divided into three categories: effects of the basic timing parameters on driver behavior and safety, statistical modeling at signalized intersection via empirical traffic crash data, and surrogate measures of safety through microscopic simulation modeling.

A research conducted by Moore et al. (Moore & Lowrie, n.d.) analyzed the traffic accidents in the field. In the same area, crashes have a reduction of 23% under coordinated conditions. In 1986, Berg et al., (1986) evaluated both safety and operational aspects using the NETSIM model, in which the accident records were collected and transferred to the frequency of stops, then a model was proposed for the accident number estimation. However, in this research, they did not find the coordinated arterial is safer than uncoordinated arterials. Some consultants preferred to use a crash reduction factor, however, the accuracy of this approach is undetermined since the crash reduction rate varies. In 2010, Guo et al., (2010) investigated 170 signalized intersections and found those with coordination experience fewer crashes than those without coordination. Li et al., (2011) also investigated the coordination effects on safety via a novel multinomial logit model, the crash likelihood can then be estimated. He also mentioned that short cycle lengths are associated with a lower risk of crashes. However, he still did not compare the coordination effects to the free operation. Fan et al., (2020a) explored the spatiotemporal correlation among traffic crashes on coordinated arterials. He also considered the correlated heterogeneity and multivariate spatial correlation (Fan et al., 2020b). Three models were generated and compared and the MPLN-MCAR model outperformed the other two models and can well capture the relationship. Roshandeh et al., (2016) in 2016 researched the safety issue caused by signal optimization to pedestrians and vehicles using an empirical Bayesian analysis method. Results show the crashes are decreased after the optimization.

For safety concerns, the inter-green (yellow+ all red) time was frequently discussed. Tracing back to 2005, Zimmerman et al., (2005) comprehensively investigated its impact on the dilemma zone and violations of traffic lights. In 2007, Wong et al., (2007) researched the impact of the cycle length and the number of phases on crashes, however, no strong evidence showed they have a direct impact on crash rate. Besides parameters, control strategies can also impact safety. Midenet et al., (2011) in 2011 compared the adaptive real-time control and a time-plan control, results indicated the adaptive real-time control performed better. Also in 2011, Lin et al., (2010) investigated the optimal length between adjacent intersections that can satisfy the needs from the safety and efficiency aspect, which aimed to reduce the rear-end crashes. In 2019, Zhang et al., (2019) investigated the signal coordination influence to different driving characteristics, they found when signal coordination is implemented, drivers of smaller ages, male drivers, and pick-up drivers take more responsivity when the crash happened. Aggressive driving behaviors are more likely to be seen on coordinated corridors than non-coordinated corridors.

There have been several practices in establishing traffic conflicts and accidents. Instead of using historical data, the surrogate methods carried the expectation of generating more convincible results. Considerable research mentioned the surrogate approach and historical data can yield similar results, it can be referred to estimate the crash rate but cannot be alternatively used to predict real crash numbers. In recent years, microsimulation analysis raised more attention, where the SSAM (Yang et al., 2020) was one of the most prevalent software. Its original idea came from Gettman’s research (Gettman et al., 2013). It integrated algorithms and software, based on which the SSAM was developed. In 2010, Sabra (Sabra et al., 2010) developed a method to balance safety and capacity in the adaptive signal control system, their practice use SSAM to yield the occurrence of conflict points between different signal timing parameters. Results also indicate the SSAM was capable of yielding convincible outcomes. However, the research did not compare the safety between adaptive signal control and free operation, therefore it did not mention which performed better in the safety aspect. Stevanovic et al., (2013) investigated how to balance the safety and efficiency based on surrogate measures of safety, a VISSIM-SSAM-VISGOAST system structure was composed. Their results showed that the total number of conflicts can be eliminated as the cycle length increases. However, it did not mention the performance difference between coordination plans and free operation.

In summary, the results from the above research are quite controversial. As the actuated coordinated control dominated the most signalized intersections in the past two decades, it is necessary to re-evaluate the safety impacts from a new framework.

3 Methodology

In practice, the use of historical traffic crash data at signalized intersections is usually limited by the available sample size, which may not be able to represent different traffic conditions. Besides, historical crash data might not interpret the safety effects of signal coordination plans, since a signal coordination plan is usually periodically updated according to the prevailing traffic condition. To enlarge the sample size, some research adopted the near-crash data; however, this is still not a practical way to collect near-crash accidents under different volume conditions. Therefore, this research employed microscopic simulation modeling as an alternative approach.

3.1 Simulation modeling testbeds

Three urban corridors in Reno, Nevada were selected as testbeds: Mill Street, North McCarran Boulevard, and Pyramid Way. Their geometric details are shown in Figs. 1, 2, and 3 accordingly with the distance between each pair of adjacent intersections. Their signal coordination plans were already recently updated by the research team; currently, no issues or complaints were reported from the public regarding the signal coordination plan.

Fig. 1
figure 1

Graphical illustration of the Mill St signalized corridor

Full size image
Fig. 2
figure 2

Graphical illustration of the North McCarran Blvd signalized corridor

Full size image
Fig. 3
figure 3

Graphical illustration of the Pyramid Way signalized corridor

Full size image

The Mill St signalized corridor is situated in the vicinity of downtown Reno with commercial areas alongside the corridor, as shown in Fig. 1. The designed speed limit is 35 mph (55 km/h). Three time-of-day (TOD) signal coordination plans have been applied for this corridor to accommodate the morning peak (AM), mid-day off-peak (MD), and evening peak (PM) traffic conditions. However, these plans are dedicated to specific volume patterns, instead of generalized conditions, therefore, they are not directly adopted in the test. While their basic signal timing information was still adopted, three signal timing plans with cycle lengths of 80s, 100 s, 130 s were designed for simulation use.

The North McCarran Blvd is located at the north of downtown Reno, as shown in Fig. 2. Commercial areas and residential areas are sparsely distributed along the corridor. The average segment spacing of this corridor is larger than those on Mill Street, and the design speed of this corridor is 45 mph (70 km/h). The investigated segment is constituted by 7 intersections, which connect the University of Nevada, Reno to the City of Sparks. Among the 7 signalized intersections, two intersections are ramp terminals. Therefore, this segment contains various intersection types and is an ideal test segment. The cycle lengths of designed signal timing plans for north McCarran Blvd are 60s, 80s, 100 s, and 130 s.

The Pyramid Way locates at the north of downtown Sparks, as shown in Fig. 3. The segment selected for simulation contains four intersections. As a major arterial connecting the residential areas and the downtown Sparks, the arterial usually has a heavy volume. Due to fewer pedestrian movements and larger segment length, the designed speed limit of the arterial is 55 mph (90 km/h). In this research, the cycle lengths of designed signal timing plans are 100 s, 150 s, and 200 s.

3.2 Simulation modeling framework

The developed simulation modeling framework contains four major steps: acquire basic timing data from ATMS, develop signal timing plans using TranSync, conduct simulation analysis to collect trajectories in VISSIM, and estimate the number of conflicts using SSAM.

In general, the basic signal timing information of each intersection on each arterial were extracted from the city’s Advanced Transportation Management System (ATMS). Then, this information was imported to the TranSync software for signal coordination optimization. After obtaining the optimized signal coordination plans, the arterials were constructed in VISSIM following the real scale measured in the field. For each simulation scenario, the trajectory files along with simulation results were collected. Eventually, these trajectories were analyzed by SSAM software to estimate the conflict points (near-crash events) during the simulation period. Based on the simulated number of conflicts, the research results and recommendations are concluded. Details of each step are described as follows.

3.2.1 Acquire basic signal timing information

In current practice, traffic signal facilities are usually operated by a central management system, where city traffic engineers can modify or change the signal status in the platform. The city of Reno currently uses the Cubic TrafficWare ATMS. In this research, signal timing parameters, including minimum green, split, cycle length, vehicle extension time, max green, yellow, all-red, coordination phases, phase sequences, and other useful information were collected for simulation modeling.

3.2.2 Develop signal coordination plans

This research employed the TranSync signal timing optimization software, which is a non-volume-based optimization tool with the MAXBAND algorithm incorporated. In comparison with the commonly used Synchro macroscopic analysis and optimization software, TranSync has the advantage of develo** an optimized signal coordination plan in absence of field-collected traffic volume data, which is more efficient for generating signal coordination plans under various traffic demand levels.

3.2.3 Conduct simulation modeling

This research adopted the VISSIM microsimulation package to generate vehicle trajectory data under various traffic control and demand conditions. For each tested arterial, the road networks were constructed on the same scale as the real-world size. Desire speeds were set to the posted speed limits. Signal coordination plans were coded in VISSIM’s Ring Barrier Controller setting.

One of the advantages of microsimulation modeling is it allows for conducting sensitivity analysis of system performance under one or more controlled parameters such as traffic demand. Since the three arterials under investigation have different speed limits and segment spacings, the capacities of the three corridors are also different. Therefore, the designed traffic demands also vary. To simplify the designed volume pattern, instead of using the Origin-Destination (OD) matrix, side street directional volumes and major street volumes were defined. The directional volumes include any possible left turn volume, right turn volume, and through volume. Major street volume only contains the volume directly input into the major street. The range of side street volume is 0 to 300 vehicles per hour (vph) with an iteration interval of 50 vph. An exception is the Mill St, due to its capacity is less than the North McCarran Blvd and the Pyramid Way, the range of its side street volume is 0 to 250 vph with an interval of 50 vph. For all corridors, the major street volume ranges from 0 to 800 vph with 100 vph as the interval. The three road networks were tested under different major street volumes and side street volumes.

For each simulation scenario, five simulation runs were conducted with different random seeds. The first 500 s was treated as the warm-up period. An hour simulation was conducted, and the data were collected from the 501st second to the 4100th second. Both operational data and safety data were collected, including but not limited to the average delay, average travel time, average speed, and maximum queue length for the entire road network. Data from individual intersections were also collected. These can be used to identify the possible oversaturation parts in investigated segments.

3.2.4 Estimating number of conflict points

After the microsimulation modeling, vehicle trajectories associated with each simulation scenario were imported to the SSAM software to estimate the number of conflicting points generated by the vehicle trajectory data. The simulated near-crash events include four conflict types: rear-end, crossing (angle), lane changing, and unclassified conflicts. For comparison purposes, this research employed the default safety performance analysis parameters, the time-to-collision (TTC) was set to 1.5 s, the maximum post encroachment time (PET) was set to 5 s, rear-end angle was set to 30 degrees, and the crossing angle was set to 80 degrees.

4 Results

With the estimated number of conflicts of each scenario generated from SSAM, this research compared the safety performance between coordinated signal schemes and the free signal operation strategy for the three case study sites.

4.1 Mill street

Figure 4a) illustrates the simulated number of conflict points under different volume scenarios for the free signal operation scheme. From the figure, these scenarios could be divided into three ranges: the number of conflicts less than 2000, between 2000 and 4000, and between 4000 and 6000. It was found that the majority of these scenarios have conflict points of less than 2000. Moreover, as expected, with the increases in traffic volume on both the minor street and the major street, the number of conflicts increases.

Fig. 4
figure 4

Simulated Number of Conflicts at Mill Street under Various Demand Scenarios: a Free Signal Operation; b 80-s Coordination Plan; c 100-s Coordination Plan; d 130-s Coordination Plan

Full size image

Figures 4b-d) presented the conflict points details associated with the designated three signal coordination plans. For the 80s-cycle signal coordination plan, the range of conflict points was significantly increased in comparison with the free signal operation strategy, especially under high volume conditions. When the major street volume exceeds 400 vph and minor street volume is larger than 150 vph, the conflict points increase sharply beyond 10,000, indicating that the 80s-cycle signal plan is not able to safely accommodate heavy volume conditions. When the cycle length increased to 100 s, the conflict points generally decrease compared to the 80s-cycle signal plan. The scenarios with the number of conflicts less than 2000 is almost the same as the free signal operation condition. However, it still has more conflict points when traffic volume is heavy. As the cycle length increased to 130 s, there are more scenarios with the number of conflict points less than 2000. Meanwhile, the maximum number of conflicts (which is associated with the heaviest volume case) decreased to under 8000.

Figure 4 presents a macroscopic overview on the general trend of the effects of traffic demand on the simulated number of conflicts, while it is unclear that whether those signal coordination schemes outperform the free signal operation strategy in terms of reducing the number of conflicts. Therefore, this research made comparisons of the number of conflicts between coordinated signal plans and the free signal operation strategy, as documented in Table 1.

Table 1 Comparison of Simulated Number of Conflicts between Coordinated Signal Plans and Free Signal Operation Scheme at Mill St
Full size table

Table 1a-c) show the percentage of reduction (or increase) in the number of conflicts of each signal coordination plan compared to the free signal operation strategy. The italic numbers indicate that these scenarios had an oversaturated traffic flow situation, which are identified based on the analysis of the simulated average delay and speed. There exist thresholds in average delay and speed in the whole dataset. Average delay may dramatically increase as the volume increase to the capacity of the roadway. In Table 1, all the oversaturation cases are located in the lower right, which corresponds with heavy volume scenarios. For the Mill St case, as the cycle length increases, a larger volume is required to reach an oversaturated condition; in other words, a larger cycle length can accommodate more vehicles (e.g., the 130 s cycle length coordination plan has fewer cases that caused oversaturation than the 80s cycle length coordination plan). Overall, under unsaturated conditions, there are considerable improvements in safety performance for all three signal coordination plans, particularly when the cycle length is longer.

The comparisons were also visualized in Fig. 5. The green areas indicate scenarios where signal coordination performs better than the free signal operation strategy, while the red areas mean free signal operation performs better. From Fig. 5a), for scenarios with a low major street volume and a low minor street volume, the coordinated signal scheme generated fewer conflicts, which is considered safer than the free signal operation strategy. However, due to the short cycle length, it is not able to accommodate heavy demand scenarios, and only 16 out of the 35 simulated scenarios have a better performance than the free signal strategy. When the cycle length increases to 100 s and 130 s, signal coordination performs better in 22 and 28 scenarios, respectively. Therefore, the 130 s-cycle signal coordination scheme can accommodate most of the demand scenarios except the several very heavy demand scenarios. This indicates that the risk of crashes can be significantly reduced with an appropriate timing plan.

Fig. 5
figure 5

Performance Comparison between Coordination Plans and Free Operation, a 80-s Coordination Plan; b 100-s Coordination Plan; c 130-s Coordination Plan

Full size image

4.2 North McCarran Blvd

Figures 6a-f) illustrate the simulated number of conflicts for the North McCarran Blvd under free operation, 60s-, 80s-, 100 s-, 130 s-, and 150 s-cycle length signal coordination schemes. For the free signal operation strategy, although the maximum number of conflicts reaches 20,000, most scenarios have less than 3000 conflicts. Under the 60s-cycle signal coordination scheme, the maximum number of conflicts peaked at 30,000 which is significantly higher than that of the free signal operation strategy. Moreover, there are quite a few scenarios that have a higher number of conflicts than the free signal operation strategy, indicating that the 60s-cycle coordination plan is not compatible. A similar trend was observed for the 80s-cycle coordination plan.

Fig. 6
figure 6

Simulated Number of Conflicts at North McCarran Blvd under Various Demand Scenarios. a Free Signal Operation. b 60-s Coordination Plan. c 80 s-Coordination Plan. d 100-s Coordination Plan. e 130-s Coordination Plan. f 150-s Coordination Plan

Full size image

As the cycle length increased to 100 s and beyond, the simulated number of conflicts are significantly decreased under all traffic demand scenarios compared to the free signal operation strategy and the short cycle length coordination plans. Moreover, the maximum number of conflicts was significantly reduced to 5000 and below. Among the tested signal coordination plans, the 150 s-cycle plan resulted in the lowest number of conflicts, which tends to be the optimal option.

To verify the findings from Fig. 6, details of the percentage of conflict points reduction or increase of each signal coordination plan compared to the free signal operation strategy are presented in Table 2. The heavy volume cases are still located at the lower right corner. However, only 60s- and 80s-cycle coordination plans have oversaturated conditions; the 100 s-, 130 s-, and 150 s- cycle coordination plans can accommodate all volume cases without leading to oversaturated cases. Results show that all coordination plans resulted in improvements in arterial safety performance under unsaturated situations. Moreover, it was found that for oversaturated situations, the free operation scheme performs better. It is worth pointing out that longer cycle lengths, such as the 100 s-, 130 s-, and 150 s-cycle coordination plans generally outperformed the shorter cycle lengths such as the 60s- and 80s-cycle coordination plans.

Table 2 Comparison of Simulated Number of Conflicts between Coordinated Signal Plans and Free Signal Operation Scheme at North McCarran Blvd
Full size table

Based on the results presented in Table 2, Fig. 7 illustrates the comparison between the signal coordination plans and the free signal operation strategy. From Fig. 7a and b, it was found that under a short cycle length signal coordination plan, signal coordination outperforms the free signal operation strategy under low demand scenarios. Among the tested 42 traffic demand scenarios, the 60s- and 80s-cycle signal coordination plans can accommodate 26 and 30 scenarios, respectively. As the cycle length increases to 100 s and beyond, all the signal coordination plan outperforms the free signal operation strategy in all the tested traffic demand scenarios. Therefore, with an appropriate signal coordination plan being implemented at the North McCarran Blvd, the risk of crashes can be reduced in comparison with the free signal operation strategy.

Fig. 7
figure 7

Performance Comparison between Coordination Plans and Free Operation. a 60-s Coordination Plan. b 80-s Coordination Plan. c 100-s Coordination Plan. d 130-s Coordination Plan. e 150-s Coordination Plan

Full size image

4.3 Pyramid way

Figure 8 overviews the simulated number of conflicts under free signal operation, 100 s-cycle length plan, the 150 s-cycle length plan, and the 200 s-cycle length plan at the Pyramid Way testbed. From Fig. 8a), it was found that under the free signal operation strategy, although the maximum number of conflicts is approximately 2000, the majority of scenarios have conflicts less than 1000. Under the 100 s-cycle signal coordination plan, the maximum number of conflicts dropped to 1500; meanwhile, there are more scenarios with the number of conflicts that is lower than 500, suggesting that the safety performance of the 100 s-cycle tends to be better than the free signal operation strategy. Nevertheless, by increasing the coordinated cycle length to 150 s and 200 s, there are fewer scenarios with the number of conflicts that is lower than 500 in comparison with the 100 s-cycle coordination plan.

Fig. 8
figure 8

Simulated Number of Conflicts at Pyramid Way under Various Demand Scenarios: a Free Signal Operation; b 100-s Coordination Plan; c 150 s-Coordination Plan; d 200-s Coordination Plan

Full size image

Again, this research made comparisons of the number of conflicts between coordinated signal plans and the free signal operation strategy, as documented in Table 3 and visualized in Fig. 9. In Table 3, there is no significant oversaturation in the three signal coordination plans. At the lower right corner of each sub-table, there is no negative value, because all coordination plans can accommodate the designed volume cases. Note that three are three negative values in the 200 s-cycle plan, which are considered as the random errors caused by the stochastic nature of simulation, since they are not consistent. Results show that overall, signal coordination considerably reduced the number of conflicts in comparison with the free signal operation strategy. The only exception is the 200 s-cycle signal coordination plan, where three of the tested scenarios underperformed the free signal operation strategy.

Table 3 Comparison of Simulated Number of Conflicts between Coordinated Signal Plans and Free Signal Operation Scheme at Pyramid Way
Full size table
Fig. 9
figure 9

Performance Comparison between Coordination Plans and Free Operation, a 100-s Coordination Plan; b 150-s Coordination Plan; c 200-s Coordination Plan

Full size image

5 Discussions

Simulation results from all three testbeds indicate that signal coordination with an appropriate cycle length has the potential of bringing considerably safety benefits to signalized arterials. The reason behind this phenomenon could be interpreted as coordinating traffic signals enables drivers to drive smoothly and proceed multiple intersections without stop**. Under such a condition, drivers do not experience frequent speed changes. This leads to reduced speed variances in traffic flow, thus reducing the potential of rear-end collisions on the major street.

Nevertheless, although signal coordination has the potential of reducing the number of conflicts, there are concerns from the public that the coordinated signals would bring up the average speed on the arterial, which might increase the severity of crashes. While this concern is beyond the scope of this research effort, we would like to emphasize that, drivers are requested to follow the posted speed limit, which was determined incorporating the general safety considerations and keep an adequate headway distance to the leading vehicle. In terms of pedestrian safety, a valid signal coordination plan also needs to satisfy the pedestrian cross timing requirement before field implementation. Therefore, under a normal operational condition, coordinated signals do not seem to introduce additional safety hazards in comparison with the free signal operation strategy.

6 Conclusions

This research aims at investigating the effect of traffic signal coordination on the safety performance of urban arterials. Findings from this research are expected to answer the concerns of the public regarding the potential increases in travel speed on the safety performance of signalized corridors. To better capture the safety indications, three real-world corridors with different speed limits were selected as the testbeds. Basic signal timing data were in line with the ATMS data. Road networks were constructed in VISSIM for generating vehicle trajectories, which were subsequently applied to the SSAM software to estimate the number of near-crash events. From the simulation modeling results, two major findings are summarized as follows:

  1. 1.

    In general, coordinating the signals along urban arterials has considerable potentials of reducing the risk of vehicular conflicts compared to under non-coordinated signals.

  2. 2.

    For a congested arterial where traffic demands exceed the capacity of the intersections, the free operation may perform better in terms of reducing the risks of conflicts.

Apart from these findings, there are also several limitations in this research. This research severs as a pilot exploration of a barely investigated topic, at this stage, it only involved vehicular crash analysis. Crashes between vehicle and pedestrian, vehicle and bicycle were beyond the scope of this research. This research did not employ an OD matrix to specify vehicles on each route due to the unavailability of detailed turning movement counts at each intersection. Therefore, this research designed the volume differently: the major street and side street volumes are uniformly increased from 0 vph to the capacity of the street, which can capture most of the real-world traffic volume patterns. Also, being limited by the availability of real-world traffic crash data under the designed scenarios, this research was based purely on simulation analysis. Future works can further differentiate the crash number and crash severity. Pedestrian and vehicle interactions should be incorporated to more reasonably represent real-world traffic operation features. The optimal cycle length should also be further investigated by corridor types, based on which, develop practice guidelines on the selection of signal coordination plans.