Module 12: Freeway Management Concepts

MODULE 12: FREEWAY MANAGEMENT CONCEPTS (Prepared by: N. H. Gartner)

This chapter discusses freeway management concepts, based on well-known references such as the FHWA Traffic Control Systems Handbook [1] and the ITE Traffic Engineering Handbook [2]. Additional reference materials can be found in standard textbooks such as McShane et al [3].

1. Introduction

Freeway management consists of freeway traffic control supplemented by such freeway operations activities as incident management, aid to stranded motorists, and driver information messages. These activities are designed to provide for improved operation and safety during off-peak as well as peak periods.

Purpose and Scope

The freeway network is an essential component of many urban area highway networks, carrying as much as 20% of the area’s vehicle-miles on as little as 5% of the area’s road-miles. Freeways were originally conceived and designed as free-flowing, limited-access facilities with little advance consideration given to the possibility of needing traffic control systems. Continued growth in traffic demand and the resultant congestion have necessitated the installation of freeway surveillance and control systems. Such systems along with other freeway management concepts, have become accepted as cost-effective, even indispensable, tools for improving traffic operations on freeways.

Freeway congestion is the primary problem being addressed by freeway traffic control. For this reason, congestion – its measurement and causes – is described first in this chapter. Then specific control concepts are described to treat these causes. These include ramp control, mainline control, and corridor control. The following concepts are typically considered to be part of a freeway management system:

 Restricted entry (closure, metering)  Traveler information (speed, travel time, route guidance, diversion)  Priority treatment (HOV operation)  Surveillance  Incident management (detection and response)

Congestion – A Daily Occurrence

Congestion occurs on a facility when demand exceeds capacity. A section of a freeway where traffic demand exceeds capacity is termed a bottleneck. A bottleneck or congestion of this type occurs as a result of conditions which either have demand increased to a level greater than capacity, or have capacity decreased to a level less than demand.

Congestion is a daily occurrence on many portions of freeway networks in urban areas. Even casual observation can define specific points at which congestion can be expected. Congestion routinely expected at predictable locations during specific periods of time is termed recurrent congestion. In contrast, other forms of congestion result from random or less predictable events. Such nonrecurrent congestion results most frequently from unpredictable accidents and incidents. Congestion from special situations (sporting events, weekend travel, maintenance

1 and construction activities) may also be considered as nonrecurrent congestion. Maintenance and construction activities, if not properly planned, may result in recurrent congestion. In defining strategies to deal with the problems of congestion, some perspective of the relative magnitude of each type is worth while. Table 1 summarizes typical percentages of the types of congestion which were reported in a metropolitan area.

Congestion whether recurrent or nonrecurrent, is manifested in terms of slow travel speeds, erratic speeds characterized by stop-and-go movement, increased and inconsistent travel times, increased accident potential, inefficient operation, and other undesirable conditions which cause user dissatisfaction. If users know to expect a certain level of congestion during peak periods, they can plan their trips accordingly. However, nonrecurrent congestion may severely impact traveltimes during peak or off-peak periods. Inability to provide a reliable, even though sometimes lower, level of service is a serious problem.

Table .1

Types of congestion by percentage for typical metropolitan area.

Type of Congestion Percentage Recurrent 43 Nonrecurrent Incidents 47 Weekends 6 Holiday weekends 3 Others (e.g. special events) 1 Total (nonrecurrent) 57 Total 100

Congestion – Its Measurement

Motorists experience congestion in terms of overcrowded freeways, stop-and-go driving conditions, and the frustration and discomfort of restricted maneuverability. The traffic engineer, on the other hand, expresses congestion in terms of traffic variables such as flow rate, density (or occupancy), and average space speed, together with the fundamental relationships among them. Figure-1 illustrates the fundamental relationship between flow rate and density. As traffic density increases from zero to some value, k1, traffic flow rate increases, and resulting operation in this range is defined as non-congested. As density increases from k1 to k2, however, though the flow rate tends to increase, traffic operations become unstable, and the probability of serious breakdown increases. Further increases in density above k2 result in a decrease in the flow rate until, theoretically, it reaches zero at jam density, kj. The traffic flow regimes occurring at densities greater than k2 is classified as congested.

The specific value of density used to define congestion depends on such factors as the geometrical characteristics of the freeway section, traffic composition, and weather conditions. The values that have been used are on the order of 50 vehicles per mile per lane, usually corresponding to a space mean speed of 35 to 50 mph and a flow rate of 1,400 to 1,800 vph per lane. It has been reported that lane occupancies, in the ranges of 0 to 20 percent, 20 to 30 percent, and 30 to 100 percent indicate noncongested, unstable (impending congestion), and congested operations, respectively. Figure 2 shows generalized freeway traffic-operations curves, which relate lane occupancy to flow rate and average space speed. As lane occupancy

2 exceeds 20 percent, speeds decrease because of fewer and shorter gaps between vehicles and increased difficulty of lane changing, and generally more restrictive flow conditions.

Congestion – Its Causes

The factors that have been identified to contribute to congestion can be considered within the categories of:

 Geometric design factors

 Traffic operations factors

 Random factors

Geometric Design Factors. The capacity of a freeway is not constant along the entire freeway length. Certain physical features result in capacity restrictions at the location where they occur. Upstream and downstream from this location, capacity is usually slightly higher, resulting in bottleneck conditions. Proper design standards make it possible to avoid in the construction phase many such bottlenecks causes. Examples of physical features which contribute to capacity reductions, and thereby congestion, include the following:

 Reduction in lanes – Congestion may occur at locations where the number of freeway lanes is reduced. Even though the lane drop may occur at an exit ramp, the through volume may exceed the remaining capacity, thus resulting in congestion. Also, the act of weaving out of the dropped lane may create turbulence, causing speed to decease with a resultant decrease in throughput. This may be a particular problem where multiple freeways merge without maintaining the same number of through lanes.

 Horizontal curvature – Capacity may be reduced by a moderately sharp horizontal curve. Under conditions of heavy flow, a vehicle may cross into the next lane, causing hesitation and speed decreases in adjacent lanes. Even momentary speed decreases during periods of unstable flow can result in congestion.

 Vertical alignment – The effects of steep grades have been well documented, particularly their effects on trucks. Accepted design standards impose limitations on acceptable grades. Even so, vertical alignment may cause small, but imperceptible, speed changes, which can affect following traffic and also result in congestion. Upgrades in tunnels, for example, are the limiting point of tunnel system capacity.

 Other physical features – Lane widths, lateral clearance, ramp design, and surface conditions can all affect freeway capacity and, in turn, contribute to congestion. Traffic lanes on interstate freeways are generally 12-ft wide, but older freeways may have more restricted lanes. Likewise, lateral obstructions, in the form of bridge abutments, retaining walls, illumination poles, sign supports, etc., may reduce capacity if located closer than 6 ft to a traveled lane.

 Number of lanes – The number of freeway lanes is also a significant factor. The 6- and 8- lane freeways show greater average capacities per lane than do the 4-lane freeways. The TRB Highway Capacity Manual [4] shows the effect of number of lanes on freeway capacity,

3 or maximum service volumes. Also, a number of studies have observed that capacity is not uniform along a freeway section, even though the number of lanes may be constant.

Traffic Operations Factors. Traffic operations factors, which can contribute to congestion, include the following:

 Volume/capacity relationship – Congestion can be expected as a direct result of the nature of supply and demand. Freeways with too much traffic attempting to use the limited capacity of the facility will experience congestion. The demand placed on a facility can be defined simply as the number of vehicles seeking to use the facility. However, measuring the flow on a freeway in an attempt to determine demand is often futile. In this case, the parameter measured is observed demand. Frequently, a considerable volume of latent demand has been diverted to other time periods, other routes, or other modes because of present conditions. Thus, a substantial difference between actual demand and observed demand may exist. The actual demand is best measured by origin-destination surveys where the times of the desired trips can be obtained.

The relationship of capacity and demand at a specific bottleneck is shown in the top portion of Figure 3. There, the demand exceeds capacity for a 30-min period. In the lower portion of the figure, congestion is shown to begin at the point of time where capacity is first exceeded, and then to extend for a time long enough to dissipate the backlog of demand. Thus, the congestion period is longer than the time period for which the demand exceeds capacity.

As previously indicated, capacity is a constant property of a roadway section, while its ability to service traffic (service rate) varies over short time periods. When a section is operating at capacity (i.e., k = k1) as shown in Figure 4 then all is well. If the incoming volume exceeds capacity, then vehicles accumulate in that section, increasing the density to perhaps, k = k2. But at k = k2, the service rate is below capacity, is the major factor causing congestion. This relationship explains why a relatively small excess demand over capacity can produce prolonged, serious congestion.

 Unrestrained ramp access – Freeway bottlenecks and the attendant congestion is frequently the result of unrestrained ramp access. In such situations, ramp traffic, when added to the volume of traffic already on the freeway, creates a demand in excess of the capacity. The frictions generated by the ramp-trafficmerging maneuver further complicate this condition.

 Exit ramp queues – Freeway mainline congestion is occasionally the result of exit ramp queues. Where the exit ramp demand exceeds the ability to process traffic trough a merging area or an intersection downstream of the ramp, or where there is a lack of storage on the ramp itself, the end result is a queue of vehicles extending back onto the freeway. Such conditions can develop not only because of congestion on the surface street system, but also as a result of heavy exit ramp demand occasioned by special activities such as sporting events.

 Weaving maneuvers – The presence of heavy weaving movements frequently cause freeway mainline congestion. Heavy weaving may result from relative ramp location as

4 determined during design and from the development of heavy traffic flows to or from specific generators.

Non-recurrent Events. Examples of non-recurrent events include traffic accidents, incidents, and other unpredictable situations, which cause nonrecurrent congestion. Experience has shown that such events occur with alarming frequency. For example, it has been observed that, within a metropolitan freeway network, at least one lane could be expected to be closed or blocked for 7.2 percent of the time as a result of accidents; for 6.7 percent of the time because of vehicle stalls or breakdowns; and for 27.8 percent of the time because of maintenance operations; for a total of 41.7 percent of the time.

Maintenance and construction operation can also cause congestion with significant delay. Where possible, such activities should be confined to periods when low volumes are anticipated, even to schedule night or weekend work. Adequate onsite traffic control and advance information to motorists must be provided.

Other factors of an environmental nature frequently contribute to capacity reductions and thereby congestion. Conditions such as wet pavement and sunrise and sunset aligned with driver’s visibility can have a significant impact, particularly during peak periods.

2. ENTRANCE RAMP CONTROL

Entrance ramp control is the most widely used form of freeway traffic control. It has been applied in numerous cities in the United States as well as in other countries. The objective is the elimination, or at least the reduction, of the operational problems resulting from freeway congestion.

Entrance ramp control is predicted on limiting the number of vehicles entering the freeway so that the demand on the freeway itself does not exceed capacity. Consequently, some of the traffic desiring to use the freeway is required to wait at the entrance ramps before being allowed to enter the freeway. Some vehicles may choose not to use the freeway at all, to enter at another location, or to enter at another time of day. Also, some drivers may decide to use public transit (if available) or use carpools. This traffic, which takes another route or time or mode of transportation, is referred to as “diverted traffic.”

The objective of entrance ramp control is to maintain uninterrupted, noncongested flow on the freeway for as long as possible by transferring the delay factor from the freeway location to the entrance ramp location, and thus diverting the excess vehicular demand to either alternate routes in the corridor, or other time periods with less demand, or other modes of transportation.

Entrance ramp control strategies which attempt to reduce entrance ramp demand generally result in some diversion of trips to other routes. Such strategies, called diversion strategies, are used in most freeway control systems in the United States. In some cases, where alternate routes are not available to bypass a bottleneck condition, a nondiversion strategy may be used. In a nondiversion strategy, ramps are metered to accommodate the full traffic demand. By releasing vehicles one or two at a time, experience has shown the nondiversion strategy to be effective in reducing the friction (jockeying for position) in the area where ramp traffic merges with mainline flows. As a result of reduced merging friction, the onset of peak-hour congestion is delayed due to the ability to maintain free-flow conditions for a longer period.

5 Entrance ramp control has been found to offer several direct benefits. Generally, it provides a higher and more predictable level of service on the freeway. By maintaining noncongested freeway flow, or by at least reducing freeway congestion, this type of control improves the efficiency and safety of freeway operation. It can also improve the safety of the merging operation, particularly at entrance ramps with inadequate sight distances. Ramp metering also smoothes the input of traffic over short periods of time even without diverting traffic. In terms of safety, freeway management projects have produced remarkable accident reductions, averaging 31 percent for seven representative projects.

In addition, entrance ramp control can help achieve maximum utilization of total corridor capacity. By making traffic wait, it makes using the freeway less attractive for some trips than using either available alternate routes or the freeway at another time, or transferring to another mode of transportation. In this way, ramp control can help distribute the total traffic load among all the routes and modes of transportation in the corridor in a more efficient manner and help steady the demand within the peak periods.

From a practical standpoint, improved traffic operations on a freeway must more than offset problems, which may be created or intensified for traffic on alternative routes. The success of entrance ramp control, then, is dependent on the extent to which the following conditions are satisfied:

 When diversion strategies are used, additional capacity (alternate routes, time periods, or modes of transportation) should be available in the corridor. The capacity should be great enough to accommodate not only the traffic to be diverted from the freeway, but also the normal level of traffic already using it. If this extra capacity is not available in the corridor, congestion will occur elsewhere, even though entrance ramp control might be preventing the congestion from occurring on the freeway. This transferred congestion might then occur on the alternate routes and in the vicinity of the entrance ramps, offsetting the benefits of freeway traffic control. Since ramp control measures do not create capacity, additional capacity must be added to the corridor to prevent congestion. This can be achieved in many ways such as by retiming traffic signals, installing a traffic-responsive signal system, adding turn lanes at intersections (by eliminating parking, adding a left-turn lane through use of paint striping or channelization with narrow lanes), reversible lane operation, and/or roadway widening.

 A priority system of entrance ramp control can be used to public transit and /or carpools. The system can be designed to provide preferential treatment to buses and/or carpools in such a way that these vehicles are allowed to enter the freeway with little or no delay, while the other ramp vehicles are required to wait.

 Adequate storage space should exist at the entrance ramps. The storage capacity should assure that queues of waiting vehicles would not seriously affect nonfreeway traffic. Ramp traffic should not be allowed to back up so as to block frontage roads or surface streets. Often, queue detectors are installed on the ramp just downstream of the frontage road, surface street, or other critical point. When queue buildup across the detector is observed, vehicles are metered at a higher rate to avoid spillback of congestion into the critical area.

 Public acceptance is important for the success of entrance ramp control. Although the system may offer significant overall benefits to the public, it is difficult to relate these benefits to individual freeway users. Users who are impacted by entrance ramp control may only

6 see the delay caused by the entrance ramp signals. It is important to inform the public about the basic reasons for installing the system (severe congestion, inefficient freeway operation, etc.), a realistic expectation of the system’s benefits to users (reduced delays and user costs), and the alternative choices that are available to system users. It must be stressed, that the system responds to the allocation of available capacity in an equitable manner. The operation should be closely monitored and fine-tuned as required to achieve these objectives, and the public should be kept informed of their achievement.

The following types of entrance ramp control are in common use:

 Closure  Ramp metering  Traffic-responsive metering  Gap-acceptance merge control  Integrated ramp control

2.1 Ramp Closure

The closure of an entrance ramp during peak traffic conditions is the simplest and most positive form of entrance ramp control. It is also the most restrictive. Therefore, it is usually the least popular and it is also subject to public opposition. However, it has been used successfully as part of a system in a number of cities in the United States and Japan. Closure may be the appropriate measure where an entrance ramp introduces serious weaving problems. However, if applied inappropriately, it could result in underutilizing freeway capacity, with the overloading of alternate routes. Because of its limitations, entrance ramp closure should not be considered except under the following circumstances:

 Adequate storage is not available at the entrance ramp to prevent queues of vehicles waiting to enter the freeway from blocking surface street traffic. The closure of the entrance ramp would eliminate the storage problem.

 Traffic demand on the freeway immediately upstream of the entrance ramp is at capacity, and an alternate route with adequate capacity is available. The closure of the entrance ramp would prevent the demand from exceeding capacity on the freeway section immediately downstream from it, and it would divert traffic at the ramp to the alternate route. The required demand-capacity relationship could occur because of nonrecurrent congestion as well as because of recurrent conditions. Therefore, closure might also be used as a response to incidents on the freeway, as is done in Japan.

Methods of entrance ramp closure that have been used include the following:

 Manually placed barriers

 Automated barriers

 Signing

7 Experience has indicated that signs alone cannot effect a positive entrance ramp closure. Automated barriers enable an entrance to be closed and opened automatically, which increases the flexibility of closure as a means of control. Since manual placement of barriers is laborious, this approach is best suited for short-term or trial control projects.

2.2 Ramp Metering

Metering is a method of restricting traffic flow. When applied as a form of entrance ramp control, metering is used to limit the rate at which traffic can enter a freeway. Metering rates used at an entrance ramp will usually be less than a practical maximum of 750 to 900 vph and greater than a practical minimum of 180 to 240 vph. When the metering rate is not directly influenced by mainline traffic conditions, the control is referred to as “pretimed metering.”

Metering Rates. The calculation of metering rates depends on the purpose for which the metering is being used. Normally, metering is used either to eliminate congestion on the freeway or to improve the safety of the merging operation as follows:

 Congestion – If the metering system is intended to eliminate or reduce congestion, the demand must be kept less than capacity. Therefore, the calculation of the metering rate at a ramp is based on the relationship between upstream demand, downstream capacity, and the volume of traffic desiring to enter the freeway at the ramp. Downstream capacity may be determined by the merging capacity at the ramp or by the capacity of the freeway section downstream, it the sum of upstream demand and ramp demands are less than or equal to downstream capacity, metering is not needed to prevent congestion. On the other hand, if the upstream demand alone is greater than downstream capacity, metering at the ramp does not eliminate congestion. Otherwise, the metering rate is set equal to the difference between upstream demand and downstream capacity.

For example, in the situation shown in Figure 1.5, the upstream demand is 5,100 vph, the downstream capacity is 5,400 vph, and the ramp demand is 500 vph. Since the total demand (5,600 vph) is greater than the downstream capacity, congestion will occur on the freeway unless the total demand is limited to capacity. Since the upstream demand is less than downstream capacity, ramp metering is a feasible solution. If a metering rate equal to the difference between upstream demand and downstream capacity (300 vph) is used, the freeway can accommodate the upstream demand and maintain noncongested flow while also handling 300 vph of the ramp demand.

The feasibility of ramp metering at a rate of 300 vph involves the following considerations:

 Is adequate additional capacity available in the corridor for the 200 vph that are likely to be diverted?  Is adequate storage available at the ramp to accommodate the queue of vehicles that will wait at the ramp before entering the freeway? If adequate storage is not available at the ramp, alternatives to be considered are closure of the ramp, or metering at other ramps upstream to reduce upstream demand.  Is the specified metering rate (300 vph) too restrictive? If so, consideration should be given to closing or metering other ramps upstream to reduce upstream demand, which would permit a higher metering rate at the ramp.

 Safety – If metering is to be used only as a means of improving the safety of the merging operation, then the metering rate is simply set at a maximum consistent with merging

8 conditions at the particular ramp. The primary safety problem of the merging operation is incidence of rear-ends and lane-change collisions caused by platoons of vehicles on the ramp competing for gaps in the freeway traffic stream. Metering is used to break up these platoons and to enforce single-vehicle entry. The metering rate selected must assure that each vehicle has time to merge before the following vehicle approaches the merge area. For example, if the average time to merge is 6 sec, the metering rate would be 10 vpm or 600 vph.

2.3 Pretimed Metering

Pretimed metering is the simplest form of entrance ramp control. Essential components of a pretimed metering system include the following:

 Ramp metering signal – Usually a standard 3-section (red-yellow-green), or 2-section (red- green), signal display which controls the ramp traffic.

 Local controller – A standard controller, with sufficient logic to vary metering rates by time of day and/or accept input from checkin and queue detectors.

 Advance ramp control warning sign with flashing beacon – A sign that indicates to traffic approaching the ramp that it is being metered.

The following components are sometimes also included:

 Checkin (demand) detector – In some applications of pretimed metering a checkin detector is placed on the approach to the ramp metering signal so that the signal will remain red until a vehicle is detected at the stop line.  Checkout (passage) detector – In some systems, a checkout detector has been used to ensure single-vehicle entry. When a vehicle is permitted to pass the ramp-metering signal, it is detected by the checkout detector, which is installed just beyond the stop line (about 8 ft past it). The green interval is then terminated as soon as the checkout detector senses the vehicle. In this way, the length of the green interval is made sufficient for the passage of only one vehicle.  Queue detector – In some pretimed metering systems, a queue detector is used to prevent ramp traffic from blocking frontage roads or surface streets. The queue detector is placed at a strategic point on the ramp, or on the frontage road, in advance of the ramp-metering signal. As soon as it senses a vehicle occupying the loop for a selectable period of time, indicating that the queue of vehicles waiting at the ramp metering signal is sufficient to interfere with traffic on the frontage road or surface streets, a higher metering rate is used to shorten the queue length.  Merge detector – A device that senses the presence of vehicles in the primary merging area of the ramp and freeway main lanes.

The location of these components on a ramp is illustrated in Figures 6 and 7.

System Operation. In a pretimed metering system, the ramp signal operates with a constant cycle in accordance with a metering rate prescribed for the particular control period. Timing the red, yellow, and green intervals of the cycle (many systems use ramp signals which

9 have only red and green intervals) depends on the type of metering used, single-entry metering or platoon metering.

 Single–entry metering – In the case of single-entry metering, the ramp metering signal is timed to permit only one vehicle to enter the freeway per green interval. Therefore, the green-plus-yellow (or just green if yellow is not used) interval is just long enough (usually about 3 sec) to allow one vehicle to proceed past the signal. Then red is used for the remainder of the metering cycle. For instance, if a metering rate of 600 vph or 10 vpm were to be used, the green-plus-yellow interval would be 3 sec and a red interval of 3 sec would be used. If the metering rate were 300 vph, or 5 vpm, a green-plus-yellow interval of 3 sec and a red interval of 9 sec would be used.

 Platoon metering – When metering rates greater than 900 vph are required, platoon metering, which permits the release of 2 or more vehicles per cycle, must be used to achieve such high metering rates. For pretimed platoon metering, the cycle length it’s determined on the basis of the desired metering rate and the average number of vehicles to be released per cycle. For example, in the case of a metering rate of 1,080 vph, or 18vpm, and a release of 2 vehicles per cycle, 9 cycles/min would be required. Therefore, the cycle length would be 6.67 sec. Similarly, if a release of 3 veh/min was used instead, the cycle length would be 10 sec. However, the timing of the cycle intervals (i.e., green, yellow, red) would depend on the form of platoon metering used, tandem or 2-abreast. In the case of tandem metering, the vehicles are released one behind another. With two-abreast metering, two vehicles are released side by side per cycle.

Compared to single-entry metering, platoon metering has a number of disadvantages such as greater driver confusion, greater probability of rear-end accidents, and greater possibility of disrupting freeway flow. Therefore, single-entry metering should always be given first consideration, and platoon metering should not be used unless it is necessary to achieve higher metering rates.

Benefits. Benefits of pretimed metering systems, when compared to a system without control, are usually observed in terms of the following:

 Higher freeway speeds

 Less delay to freeway users

 Higher freeway-service volumes

 Reduced travel time to freeway users

 Safer merging operation (fewer rear-end and lane-change collisions)

 Reduced operating costs to freeway users

 Reduced fuel consumption and emissions.

Benefits that have been reported for some ramp metering installations are given in Ref. [5]. It has been shown that entrance ramp control can be very cost effective. Experience has

10 indicated that the biggest net gain in benefits is realized in going from no control to pretimed metering, and that further increases in control sophistication obtain reduced returns.

Pretimed metering offers both advantages and disadvantages. The most important advantages are that it gives the driver a dependable situation to which he can readily adjust, and lower costs associated with pretimed metering. The major disadvantage is that the system can neither respond automatically to significant changes in demand, nor adjust to unusual traffic conditions resulting from incidents.

3. Traffic-Responsive Metering

In contrast to pretimed metering control, traffic- responsive metering is directly influenced by the mainline and ramp traffic conditions during the metering period. Metering rates are selected on the basis of real-time measurements of traffic variables indicating the current relation between upstream demand and downstream capacity.

Fundamental Traffic Flow Relations. In order to determine or predict demand- capacity conditions on the basis of real-time measurements of traffic variables, a description or model of traffic flow is necessary. Most frequently used as indicators of operating conditions for traffic-responsive metering are the traditional functional relationships between the flow rate, q; space-mean speed, u; and density, k. A typical form of the relationship among q, u, and k is depicted in Figure 8 (cf. also Fig. 3.1 in Chapter 3). Congestion occurs whenever demand exceeds capacity. Therefore, as shown in Figure 8, the values of qm , um, and km define boundaries between noncongested flow and congested flow. The purpose of metering is to prevent or reduce congestion, or, in other words, to keep the values of the fundamental traffic flow variables at levels that define points on the uncongested-flow portions of the traffic flow curves. Thus, a basic strategy of traffic-responsive metering is as follows:

 Obtain real-time measurements of traffic variables on the freeway.

 On the basis of these measurements, determine where on the fundamental traffic flow curves the freeway section is operating with respect to capacity.

 Determine the maximum ramp metering rate at which vehicles can be permitted to enter the freeway without exceeding capacity.

A refinement, which is often made to this strategy, is to select the traffic entering traffic flow curves on the basis of traffic composition and weather conditions.

Lane occupancy (a surrogate measure for density) and flow rate (volume) are the two traffic variables generally used to describe freeway conditions for traffic-responsive metering. These are the control parameters usually used because they can be measured in real time using vehicle detectors.

The principal traffic-responsive strategies remain the following :

 Demand-capacity control

 Occupancy control

11 Demand-Capacity Control. Demand-capacity control involves the selection of metering rates on the basis of real-time comparison of upstream volume and downstream capacity. The upstream volume is measured in real time and compared with either a preset value of downstream capacity determined from historical data or a real-time value computed from downstream volume measurements. To be most effective, the downstream capacity used should account for the effects on capacity of weather conditions, traffic composition, and incidents.

The difference between the upstream volume and the downstream capacity is then determined and used as the allowable entrance ramp volume. This ramp volume is expressed as a metering rate to be used during the next control interval (usually 1 min). If the upstream volume is greater than the downstream capacity, a minimum metering rate is used (e.g., 3-4 veh/min). Theoretically, if the upstream volume is greater than the downstream capacity, a zero metering rate, or ramp closure, should be used in order to prevent congestion. However, it has been found that metering rates lower than 3 veh/min are not effective, because vehicles waiting at the ramp will consider the ramp metering signal to be malfunctioning and disobey it.

Since a low upstream volume could occur in congested as well as noncongested conditions, volume alone does not indicate degree of congestion. Therefore, an occupancy measurement has to be made to determine whether noncongested or congested flow prevails. If the occupancy measurement is above a preset value (e.g., 18 percent) which is determined from historical data, congested flow would be assumed to exist and a minimum metering rate used.

Occupancy Control. Occupancy control utilizes real-time occupancy measurements generally taken upstream of the entrance ramp. One of a number predetermined metering rates is selected for the next control interval (usually 1 min) on the basis of occupancy measurements taken during the current control interval. For a given entrance ramp, the metering rate to be used for a particular value of occupancy is based on a plot of historical volume-occupancy data collected at each measurement location. From such a plot, an approximate relationship between volume and occupancy at capacity is determined. For each level of occupancy measured, a metering rate can be computed which corresponds to the difference between the predetermined estimate of capacity and the real-time estimate of volume. If the measured occupancy is greater than, or equal to, the preset capacity occupancy (e.g. 23%), a minimum metering rate is selected instead of a zero rate or ramp closure.

System Components. A traffic-responsive metering system contains the same components as described for pretimed metering. These include ramp metering signal(s), local controller, advance warning sign with flashing beacon, and detectors. The local controller unit for traffic-responsive metering requires additional logic in order to monitor traffic variable measurements, select or calculate metering rates, and respond to override-type conditions such as excessive queues. Queue, checkin, checkout and merge detectors are typically also included in traffic-responsive metering systems. The typical location of these components on a ramp is shown in Figure 7.

System Operation. The operation of a traffic-responsive metering system is similar to that of a pretimed metering system, except in regard to the following:

 Metering rate selection – Single-entry metering is normally used to time the red-yellow- green (or red-green) intervals for a given metering rate. However, if high metering rates (e.g., greater than 13 vpm), are required, platoon metering might be used.

12 Metering-rate selection – In a traffic-responsive metering system, the selection is based on real-time measurements of traffic variables, which describe traffic flow conditions on the freeway. The control interval, which is the time period during which a selected metering rate remains in effect, is much shorter for a traffic-responsive metering system (e.g., 1 min) than for a pretimed metering system (e.g., 30 min, 1 hr, or the entire peak period).

System Evaluation. As with pretimed metering, the following benefits have been experienced with traffic-responsive systems:

 Higher freeway speeds

 Less delay to freeway users

 Higher freeway-service volumes

 Reduced travel time to freeway users

 Safer merging operations

 Reduced operating costs to freeway users

 Reduced fuel consumption and emissions

Traffic-responsive metering system have been found to be cost-effective, producing improvements that are generally 5 to 10 percent greater than those of pretimed metering. With incremental costs only in the 10 to 15 percent range, traffic-responsive metering is considered to be a cost-effective approach. The primary advantage of traffic-responsive metering is that it automatically adapts to changes in traffic flow. Most importantly, the metering system helps to minimize the adverse effects caused by short-term variations in traffic demand and reductions in capacity caused by incidents.

3.1 Gap-Acceptance Merge Control

The merge-control concept of entrance ramp metering is intended to enable a maximum number of entrance ramp vehicles to merge safely and smoothly without causing significant disruptions to freeway traffic. The concept involves maximum utilization of gaps in the traffic stream of the freeway lane into which ramp vehicles are to merge. It does not, however, involve the calculation of ramp metering rates in accordance with the demand-capacity relationship. Instead, the problem is merely one of inserting entrance ramp vehicles into existing freeway gaps.

The concept of gap acceptance at freeway entrance ramps is important in describing the interaction of the freeway and ramp traffic. It is assumed that the ramp driver measures each gap in the adjacent freeway lane and compares it with an acceptable gap, which he/she judges to be large enough for safe merging.

The minimum acceptable gap is dependent on several factors, such as the following:

 Entrance ramp and freeway geometric

13  Vehicle performance characteristics

 Driver behavior

 Traffic conditions

 Weather conditions

Merge-control systems are designed to improve the merging operation at the entrance ramp by providing the driver with the information needed to coordinate in time and space entry onto the freeway. These systems operate according to the following basic guideline procedures:

 Detection of an acceptable gap on the freeway into which a ramp vehicle could merge

 Projection of the arrival of the acceptable gap at the merging point of the entrance ramp

 Release of the ramp vehicle in sufficient time to accelerate and merge into the moving gap

System Components. Gap-acceptance merge-control systems use many of the same components as described for pretimed metering, to include ramp metering signals, local controller, advance warning sign with flashing beacon, and detectors. A main lane gap/speed detector is located in the shoulder lane of the freeway upstream of the ramp merge to provide data from which the controller unit can determine presence and speed of available gaps in which to insert merging ramp traffic. Queue, checkin, checkout, and detectors are normally also included in gap-acceptance merge-control systems.

System Operation. A gap-acceptance merge-control system does not normally operate in accordance with a constant metering rate for a specified control interval, as do pretimed and traffic-responsive metering systems. Instead, it operates in response to the availability of acceptable gaps in the lane of the freeway into which ramp vehicles are to merge. Usually, the system is designed to operate in a single-entry metering mode, with the ramp metering signal resting on red when no vehicles are waiting on the ramp.

Procedures can be summarized as follows for the nominal operation of a gap-acceptance merge-control system, with single-entry metering and the ramp metering signal resting on red:

 A vehicle stops at the ramp metering signal and actuates the checkin detector.

 The controller begins to measure gaps and vehicle speeds, which are, sensed by the gap/speed detector, which is located upstream from the ramp in the lane of the freeway, into which ramp vehicles are to merge.

 The controller compares each measured gap to a preset minimum gap size to determine whether or not it is an acceptable gap.

 If a gap is not acceptable, the controller considers the next gap. If it is acceptable, the controller computes the time at which the vehicle at the ramp metering signal should be released in order to arrive at the merge point at the same time, as does the acceptable gap.

14  At the proper instant, the controller causes the ramp metering signal to change to green.

 The ramp metering signal remains on green for a fixed interval long enough to release a single vehicle. Then, it changes to yellow for a short fixed interval before it changes to red. The green-plus-yellow (or green only) interval is usually about 3 sec long. It is necessary that the ramp metering signal remain on red long enough to give the next vehicle in line time to pull up to the signal. Thus, the minimum length for a full green-yellow-red (or green-red) cycle should be 4 to 5 sec.

System Evaluation. Benefits to be realized from a gap-acceptance merge-control system are similar to those of a traffic-responsive metering system. In general, for entrance ramps that have well-designed geometric, a gap-acceptance merge control is less cost-effective than either pretimed or traffic-responsive metering systems. However, gap-acceptance control might be warranted at locations where the geometric are substandard and the primary concern is to improve the safety of the merging operation.

3.2 Integrated Ramp Control

Integrated ramp control refers to the application of ramp control to series of entrance ramps where the interdependency of entrance ramp operations is taken into account. The primary objective of integrated ramp control is to prevent or reduce the occurrence of congestion on the freeway. Therefore, the control of each ramp in the control system is based on the demand- capacity considerations for the whole system rather than on the demand-capacity constraint at each individual ramp.

If congestion is to be prevented or reduced on the freeway system, the concept of integrated ramp control must be used in the design of a system of controls for a section of freeway with more than one entrance ramp. It is applied in the following types of systems:

 Integrated pretimed metering (including ramp closure)  Traffic-responsive metering  Gap-acceptance merge control

A discussion of integrated ramp control applied to each of these systems follows.

Integrated Pretimed Metering. Integrated pretimed metering refers to the application of pretimed metering to a series of entrance ramps. The metering rate for each of these ramps is determined in accordance with demand-capacity constraints at the other ramp as well as its own local demand-capacity constraint. These metering rates, which are computed from historical data pertaining to each control interval, require the following information:

 Mainline and entrance ramp demands  Freeway capacities immediately downstream of each entrance ramp  The traffic pattern within the freeway section to be controlled

This information provides the basis for establishing the demand-capacity constraints of the entrance ramps and their interdependencies.

 Metering rate calculations - Given the requi9red data, the fundamental procedure for computing metering rates involves five steps:

15 1) Start with the entrance ramp, which is farthest upstream.

2) Determine the total demand (upstream mainline demand plus ramp demand) for the freeway sec6tion immediately downstream of the ramp.

3) Compare the total demand to the capacity of the downstream section, and proceed as follows:

a) If the total demand is less than the capacity, metering is not required at this ramp by the demand-capacity constraint. Skip Step 4 and go immediately to Step 5.

b) If the total demand is greater than the capacity, metering is required at this ramp by the demand-capacity constraint. Proceed to Step 4.

4) Compare the upstream mainline demand to the capacity of the downstream section, and proceed as follows:

a) If the upstream mainline demand is less than the capacity, the allowable entrance ramp volume (or metering rate) is set equal to the difference between the capacity and the upstream mainline demand.

b) If the upstream mainline demand is greater than or equal to the capacity, then the allowable entrance ramp volume is zero, and the ramp must be closed. The volumes permitted to enter at ramps upstream must be reduced accordingly.

5) Select the next entrance ramp downstream and go back to Step 2.

Integrated Traffic-Responsive Metering. Integrated traffic-responsive metering is the application of traffic-responsive metering to a series of entrance ramps where the metering rates at each ramp are selected in accordance with system, as well as local, demand-capacity constraints.

 System operation – During each control interval, real-time measurements are taken of traffic variables (usually volume, occupancy, and/or speed). The data are used to define the demand-capacity conditions at each entrance ramp. Then, on the basis of these measurements, both an independent and an integrated metering rate are calculated for each entrance ramp. Of these two metering rates, the one that is the more restrictive is selected to be used during the next successive control interval.

 Metering rates – The methods used to calculate independent and integrated traffic- responsive metering rates are basically the same as those used to compute independent and integrated, pretimed metering rates. Usually, instead of metering rates being calculated in real time, a set is precomputed for the range of demand-capacity conditions expected, from which the metering rates are then selected in real time. A linear programming model is often used to calculate predetermined sets of integrated, traffic-responsive metering rates. Also, the metering rates are usually subject to the merge-detector, queue-detector, and maximum-red-time overrides used in traffic-responsive metering.

16 Integrated Gap-Acceptance Merge Control. Integrated, gap-acceptance merge control is the application of gap-acceptance merge control to a series of entrance ramps where such control at the individual ramps is subject to system demand-capacity constraints.

Metering rates are computed by the procedure described for calculating integrated, pretimed metering rates. These integrated metering rates are computed for each entrance ramp on the basis of the real-time measurements of traffic variables which are used to define the system’s demand-capacity conditions. Then, for each entrance ramp, a minimum acceptable gap setting is determined which will yield a ramp volume corresponding to the integrated metering rate. The minimum acceptable gap is the smallest gap in freeway main lane traffic into which a ramp vehicle will merge. Larger settings of a minimum acceptable gap have the impact of reducing the metering rate.

Integrated vs. Independent Ramp Control. Comparisons of integrated and independent entrance ramp control indicate that increased benefits are realized with integrated ramp control. Improvements occur in terms of the following:

 Lower travel time

 Higher total travel

 Fewer accidents

In addition, in the case of traffic-responsive metering, the greater system flexibility provided by integrated ramp control enables an optimal system response to individual variations in traffic demands and capacities resulting from incidents on the freeways.

Controller Interconnection. A significant feature of integrated ramp control is the interconnection among local ramp controllers, which permits conditions at one location to affect the metering rate imposed at one or more other locations. Real-time metering plans are computed and updated by a central controller, which issues metering rates to the respective local ramp controllers, based on freeway traffic information obtained from vehicle detectors throughout the system.

1. EMERGING TECHNOLOGIES

Most of the advances and emerging technologies in freeway management systems are in the computing and communications technologies. Development of those fields will continue to enhance the freeway ramp control process. Specific advancements may include:

 System Operation. As freeway systems expand with more communication links and detector data become more readily available, there will be expanded operation of entrance ramp metering with the metering rates being determined on a system or subsystem basis.  Integrated Systems. Earlier freeway management systems operated, generally, independently of the operation of surface street signal systems. State traffic engineers and local system managers sometimes communicated informally to coordinate the operation of the two systems. Future systems will likely by fully integrated, with data exchange and control decisions being made adaptively based on real-time information.

17  Information to Motorists. Advanced information systems will be installed as part of the ITS deployment and expansion program. These systems will assist motorists in selecting or bypassing ramps where the metering rates may be restrictive. Such diversion strategies will be considered explicitly in integrated freeway and surface street systems.  Advanced Control Algorithms. As part of the National ITS Architecture Implementation Strategy there will be opportunities for deployment of “processing technology and advanced algorithms that enable advanced vehicle and traffic control applications.”  Advanced Ramp Metering Concepts. Queue management becomes critical under heavy ramp demand conditions. Improved queue management algorithms based on multipoint detection are under development. Also, traffic-responsive activation of ramp control will likely be used to manage traffic during off-peak or weekend incident conditions.

5. References:

1. Traffic Control Systems Handbook. Report FHWA-SA-95-032. U.S. Dept. of Transportation, February 1996. 2. Inst. of Transportation Engineers. Traffic Engineering Handbook, Prentice Hall, 1992. 3. McShane et al: TRAFFIC ENGINEERING, 2nd edition, Prentice Hall, 1996. 4. Highway Capacity Manual. Transportation Research Board, 2000. 5. Freeway Management Handbook. Report No. FHWA – SA – 97 – 064. USDOT, August 1997.