Innovative intersection design shows promise

Transportation Management Article October 08, 2014
Printer-friendly version



At signalized intersections, daily fluctuations in demand can be handled by the amount of green time allocated to the movement.


As traffic increases over time and the demand cannot be handled by signal-timing changes, the resulting congestion can be handled by adding lanes to the problematic movement to directly increase its capacity or adding lanes to the other movements to minimize the amount of green time they “steal” from the problematic movement. Availability and cost of right-of-way often influences the decision about which movements can be widened. The process of evaluating congested intersections, designing the solution and then implementing it is relatively straightforward, and traffic engineers have a good understanding of the mitigation options for conventional four-way intersections with typical traffic movements. 


A conventional four-way intersection in Morgantown, W.Va., with non-typical traffic movements presented an opportunity to develop an unconventional solution. At this specific intersection, the major mainline movement is not a through movement, which presents some unique challenges, and there was available right-of-way in one of the intersection quadrants that could be incorporated into the solution. After investigating a number of conventional solutions to the congestion, Dr. Jason Chou suggested an option that would turn the intersection into a triangle by constructing a hypotenuse in the open quadrant. This idea was derived from an intersection he had observed in his home country of Taiwan. Flow of traffic is controlled at the three vertices by traffic signals and flow within the triangle is one-way counterclockwise, similar to a roundabout. 


Intersection problems


Morgantown is the home to West Virginia University. The basketball arena, football stadium, part of WVU campus, WVU housing, three major hospitals, a pharmaceutical manufacturer and many other commercial developments are located along S.R. 705 on the north side of Morgantown, so there is significant traffic volume on a daily basis. The intersection of Chestnut Ridge Road (westbound), Van Voorhis Road (northbound/southbound) and Burroughs Street (eastbound) in Morgantown is located along the S.R. 705 corridor. The mainline arterial movement is the westbound left turn and the northbound right turn. The eastbound and southbound approaches are considered to be the minor approaches. Peak-hour data collection revealed that the westbound left-turn demand was 65% of the overall westbound volume and 72% of the northbound traffic was for the right-turn movement. The overall intersection level of service (LOS) based on control delay was D in the a.m. peak hour and E in the p.m. peak hour. 


This particular intersection is a turning point for S.R. 705, running northbound-eastbound (right-turn movement) and westbound-southbound (left-turn movement). Since these two turning traffic movements can run concurrently, the allocated green time can be increased to account for the increasing demand. However, ongoing construction of student housing north of this intersection continues to increase the level of congestion to a point that signal-timing adjustments are not effective, and significant queueing occurs during peak hours. The pedestrian movements at this intersection operate as an exclusive phase due to the complexity of traffic movements. Due to the size of the intersection, the pedestrian walk and clearance times are a combined 32 seconds, which is approximately 15% of the peak-hour cycle length. Therefore, every cycle that the pedestrian phase is served during the peak hour causes significant queueing. 


At typical intersections where the mainline movement flows straight through the intersection, vehicles from the side street can enter the mainline by turning right on red or with a green arrow that runs concurrently with the mainline left-turn movement, which helps reduce the amount of green time needed to serve the side street. Furthermore, the side-street movements are commonly served concurrently, or running at the same time. If the mainline movement isn’t a through movement, these capacity-increasing strategies do not exist. Since the side-street movements are adjacent to each other (westbound and southbound in this case), they must be served sequentially (i.e., split phase), which further reduces the amount of green time available for the mainline movement. 


Investigating solutions


Conventional mitigation alternatives were evaluated for the intersection, including modified signal timings, adding lanes and construction of a roundabout. Signal-timing modification alone was not able to provide enough capacity to reduce the queueing during the peak hour. Right-of-way constraints dictated that lanes could only be added on the eastbound and northbound approaches. Multilane roundabouts were evaluated for the intersection, particularly to utilize the available right-of-way in the southeast quadrant of the intersection. The roundabout would not work from a geometric or operational standpoint. Geometrically, the roundabout could not be properly constructed with the southbound approach having the necessary width and entry angle without impacting the properties in the northwest and northeast quadrants. Operationally, significant queueing was projected for the eastbound approach because there were minimal gaps remaining in the traffic flow due to the heavy westbound left-turn and southbound movements. A signalized alternative with additional lanes on the eastbound and northbound approaches was identified and carried forward as an acceptable alternative from an operational standpoint, and is referred to as the “Add 2 Lanes” alternative in this article.


In addition to the quantitative objectives of reducing delay and queueing (derived from simulation), there were qualitative objectives of improving pedestrian safety, reducing vehicle crash potential and improving access to the businesses in the vicinity of the intersection. With the recent emphasis placed on Unconventional Arterial Intersection Designs (UAID) by the FHWA Every Day Counts program, other solutions for this intersection were investigated. Many of the UAIDs improve intersection capacity by displacing the left turns from the main intersection so that there are fewer signal phases (no left-turn arrow) and more green time to be allocated to the other heavy movements. Examples include median U-turns, jughandles and quadrant intersections, all of which have been constructed and are operational in the U.S. The quadrant intersection design is intended to divert left-turn movements away from the main intersection by creating a new two-way roadway in one quadrant of the intersection that connects two adjacent intersection legs. This additional roadway is primarily used to divert the left-turn traffic from these two roads away from the main intersection so that the corresponding left-turn phases can be eliminated and decrease lost time at the intersection. The two additional T-intersections created by this road are typically signalized to still provide dedicated right-of-way for left-turning vehicles, but are synchronized in a way to facilitate major mainline movements. Typically, the displaced left-turn movements are minor movements. 


For the study intersection, the left turn is the major movement, so the desire was to displace other traffic movements from the main intersection to increase the green time for the left-turn movement. After some brainstorming regarding the performance of these alternatives and the benefits regarding some of the UAIDs, a new alternative design was identified that provides safety benefits similar to a roundabout through reduced conflict points and operates similar to a roundabout with counterclockwise flow, but provides signalized right-of-way for entering vehicles. A diagonal road is constructed across one of the quadrants, which will serve to divert the minor movements at the intersection away from the main intersection where the heavy left-turn movement will occur. Since the resulting shape is a triangle and the vehicle flow is counterclockwise, similar to a roundabout, the name “Triangabout” was adopted to describe it. Signalized intersections with a non-through arterial movement exist in other locations, so the solution developed for this intersection may be applicable at other locations. In fact, S.R. 705 makes another turn at another intersection along this corridor.


The Triangabout


The proposed Triangabout design for the study intersection is illustrated in Figure 4. This layout will utilize currently available right-of-way in the southeast quadrant of the intersection. In this figure, ‘A’ denotes the location of the original intersection, and ‘B’ and ‘C’ are the two new signalized intersections created by the one-way “quadrant” road. The flow along the three legs of the Triangabout is one way in a counter-clockwise manner. All left-turn movements at the three signalized intersections are dual lanes with protected-only signalization. Pedestrian access is provided to all four quadrants of the intersection with pedestrian signalization and crosswalks as shown in Figure 4a. In order to maintain the flow of traffic circulating within the Triangabout, dedicated pedestrian right-of-way is only provided across the external approaches, which allows pedestrian movements to run concurrently with other signalized phases, rather than exclusive phases. 



Figure 4. A conventional signalized intersection alternative (Add 2 Lanes).


Since the three intersections are in close proximity, the signals must operate in a coordinated fashion to facilitate continuous movements through the intersections. Although this can be accomplished with three separate traffic-signal controllers, it is proposed to run all three intersections with a single controller to avoid congestion that might occur when a controller goes into transition (e.g., changing time-of-day plans or pre-emption). The proposed signal phasing for the three intersections running on one controller is depicted in Figure 4b, where the first, second and third rings show the plans for intersection C, A, and B, respectively.


The one-way conversion of the two legs AB and AC allows the capacity on those links to be increased by re-striping existing pavement. This results in the ability to configure the westbound approach to C and northbound approach to B as three entering lanes. The original northbound right-turn (mainline) traffic should only be stopped at one intersection, at most. They will be stopped at B when the southbound left-turn movement is active or when a pedestrian activates the pedestrian signal. The pedestrian phase is served concurrently with the southbound left-turn movement to minimize vehicle delays. The northbound right-turn vehicles then proceed through C without stopping because it is a continuous movement without conflicting vehicles or pedestrians. Therefore, the original northbound right-turn movement should not experience a significant difference in delay, and actually may be improved due to the additional lane capacity.


The original westbound left-turn (mainline) traffic might get stopped at C, but should proceed through A if the signals are coordinated correctly. Unless a pedestrian is crossing at B, which stops the southbound through movement, those vehicles should continue through unimpeded. Therefore, the Triangabout should provide as good, if not improved performance for this major movement.  


At the main intersection A, there are only three approaches using this intersection, which can operate as a three-phase signal with overlaps, which will greatly reduce the overall delay. Intersections B and C can both operate as two-phase signals with overlaps, so lost time will be minimal. The two pedestrian movements at A will run concurrently with other vehicle movements. The westbound right turn at A will flow continuously as an overlap, except when a pedestrian crosses the southbound approach. Another benefit of this alternative is that widening the eastbound approach is not required. Once the eastbound movement is green, all traffic will turn left into four lanes. The eastbound vehicles that ordinarily go straight are diverted to B to make a left-turn movement. The eastbound vehicles that ordinarily turn left are impacted the most because they are diverted to make left turns at B and C, but will generally make a free-flow westbound right turn back at A. The southbound vehicles that ordinarily turn right or go straight will be relatively unaffected. The southbound vehicles that ordinarily turn left will be diverted to B to make a left turn, but will proceed through C unimpeded. If the signal coordination is effective, vehicles queued at A should not have to stop at B to make a left-turn movement. The original northbound left-turn and straight movements also are impacted because they must divert to C. 


The Triangabout will not reduce the delay for every movement at the intersection; however, it should favor the mainline movements and can be timed to benefit other heavy movements. The benefits that are derived at UAIDs by diverting the left-turn traffic are experienced with the Triangabout by diverting through movements, due to the non-through arterial movement. Similarly, the safety benefits of a roundabout are experienced due to the reduction of the number of crossing (i.e. T-bone) vehicle paths. 


Stream success


The system benefits main traffic streams significantly by increasing capacity for westbound left-turn and northbound right-turn movements. It simplifies phase design at intersection A so that green time can be allocated for a dramatic main traffic stream increase. Phases omitted include all northbound movements, EBTH and eastbound left turns and southbound left turns. In addition, the exclusive pedestrian phase is removed. Both B and C intersections are running with a two-phase design so that green times allocated for the main traffic stream are sufficient for coordination with adjacent intersections. Consequently, the system enhances arterial functionality for the entire corridor. Meanwhile, safety also is improved by reducing the total number of conflict points. After converting traffic lanes among intersection A, B and C into a one-way system, right-angle crossing conflicts—the most severe type—are all eliminated. However, it is possible that rear-end collision rates could increase with the Triangabout design due to the addition of two signalized intersections. In addition, the system does not need much land to build additional lanes. It makes the solution possible for areas with similar traffic patterns and limited land.


On the other hand, the routing scheme of a Triangabout induces longer distances for minor traffic movements. The most complicated movement is the eastbound left turn, where traffic is diverted through all Triangabout intersections to complete the maneuver. Although the main traffic stream along the arterial is well maintained, intersection performance might be degraded due to diverted minor traffic streams if such demand is high. It is possible that the eastbound left-turn movement could be served directly, but that additional phase would increase delay for the WBTH and westbound right-turn movements. It is not feasible for this location due to the volume distribution and because the eastbound left-turn storage is not adequate, which would cause left-turn vehicles to impede the flow of the eastbound right-turn vehicles. This alternative could be explored in other applications of this design.


The design itself is considered to be unconventional because there is no known similar installation. It might also take time for road users to become familiar with the routing scheme, and could initially result in an increase in crashes until drivers become familiar.


Circle it as a possibility


This study applied VISSIM simulation to quantify performance of the Triangabout compared to both the existing operations and the operations of the conventional signal with two additional lanes (i.e., Add 2 Lanes). The a.m. peak period at the study intersection is from 7:45-8:45 a.m. and the p.m. peak is from 4:30-5:30 p.m. To evaluate the general safety of the alternatives, conflict points resulting from crossing, diverging or merging vehicle paths were identified and compared. 


Figure 5a and Figure 5b illustrate the a.m. peak hour and p.m. peak hour delays by movement. Figure 5c illustrates the overall intersection delay. The LOS along the northbound and westbound approaches will be improved to B and C levels in the a.m. and p.m., respectively, with implementation of the Triangabout. At the intersection level, the Triangabout can save up to 55% and 54% of travel delay in the a.m. and p.m. peak hours, respectively. The Add 2 Lanes alternative shows comparable performance in terms of delay reduction for this intersection. The Triangabout design can reduce the average delay for the major westbound left turns by 63% and 59% during the a.m. and p.m. peak hours, respectively. The major northbound right-turn delay is reduced by the Triangabout alternative 53% and 59% during the a.m. and p.m. peak hours, respectively. Additionally, the Triangabout benefits all other minor movements except the northbound left turns during the a.m. peak hour. Although the Add 2 Lanes alternative can improve traffic operations for this intersection as well, the improvement to the mainline traffic stream is not as significant as the Triangabout alternative.


Figure 5a and Figure 5b illustrate the a.m. peak hour and p.m. peak hour delays by movement. Figure 5c illustrates the overall intersection delay.


The results from the mainline travel time analysis along the mainline corridor showed the Triangabout significantly reduced travel time in both directions by more than 50% compared to the existing scenario. It also outperforms the Add 2 Lanes alternative during the a.m. peak in both directions and the p.m. peak in the westbound direction.


Even though the Triangabout alternative adds two signalized intersections, it reduces conflict points by 34% as compared to the existing network. The Add 2 Lanes alternative, on the other hand, increases the conflict points by 13%. The benefits of the Triangabout are similar to a roundabout where the more severe crossing type conflicts are eliminated due to the unidirectional flow of traffic.


Using estimated construction costs of each alternative and the benefits—derived from simulation of the reductions of travel delay, emissions and fuel consumption—benefit-cost ratios were calculated for each alternative. Based on only one year of operation, the Triangabout and Add 2 Lanes alternatives resulted in benefit-to-cost ratios of 2.88 and 2.66, respectively. 


The proposed Triangabout design can significantly improve mobility and safety of the study intersection in Morgantown, which is an uncommon configuration where the mainline movement is not a through movement and operates as concurrent left and right turns. Results from quantitative analysis produced with VISSIM indicated the Triangabout can save more than 50% of travel delay and reduce 34% of the conflict points at this intersection. This alternative not only enhances the arterial functionality along the corridor, but also benefits most of the minor movements at this intersection. The operational benefits for the overall intersection between the Triangabout and an improved conventional signalized alternative were not drastic. However, the Triangabout alternative greatly improves motorist safety at the intersection and has other qualitative benefits, including improved pedestrian safety, improved access to adjacent developments and reduced right-of-way impacts. Therefore, the Triangabout is deemed to be a preferable alternative.


Because the Triangabout diverts minor traffic movements, the intersection delay may further degrade if that minor traffic demand increases. Thus, future research will include a sensitivity analysis to investigate this effect. Furthermore, the length of the internal legs of the Triangabout were dictated by the available right-of-way, but future analysis will investigate the sensitivity of these distances. The traffic-signal control for the Triangabout is quite complex, especially if operated with a single controller. The performance is heavily dependent on having the optimal coordination settings. However, no optimization software packages are capable of handling this configuration, so optimization is a manual iterative process. Future research will examine methods to calculate solutions for various volume scenarios, which also will account for the coordination settings of adjacent intersections. 


Currently, both alternatives are being vetted by the West Virginia Department of Transportation and the decision-making committees in the Morgantown area, with a final decision to be made in the coming months. TM&E

About the author: 
Nichols is an associate professor of engineering at Marshall University and director of Intelligent Transportation Systems at the Nick J. Rahall Appalachian Transportation Institute (RTI). Chou is a research associate with RTI.
Overlay Init