7.0 Arterial Data Types and Performance Measures
7.1 Overview of NCHRP Arterial Performance Measures
The Highway Capacity Manual [Transportation Research Board. TRB Special Report 209: Highway Capacity Manual. 3rd ed. Washington, DC: Transportation Research Board of the National Academies, 1994.] states that the arterial level of service is based on the stopped delay per vehicle at a signalized intersection. Calculating this measure requires knowledge of the signal phase and the number of vehicles in each lane group on a second by second basis. Table 2 shows the standard information for Arterial LOS [Transportation Research Board. TRB Special Report 209: Highway Capacity Manual. 3rd ed. Washington, DC: Transportation Research Board of the National Academies, 1994.].
Table 2 – Arterial Levels of Service
Level of Service |
Stopped Delay Per Vehicle (Seconds) |
|---|---|
A |
≤5.0 |
B |
>5.0 and ≤15.0 |
C |
>15.0 and ≤25.0 |
D |
>25.0 and ≤40.0 |
E |
>40.0 and ≤60.0 |
F |
>60.0 |
The National Cooperative Highway Research Program (NCHRP) has projects under way to refine the arterial performance measures. NCHRP Project 3-79 [Bonnenson, James A., Anuj Sharma, and Darcy Bullock. Measuring and Predicting the Performance of Automobile Traffic on Urban Streets. National Cooperative Highway Research Program Project 03-79. Washington, DC: Transportation Research Board of the National Academies, 2008.] has several research tasks that focus on real-time performance measures for arterial traffic. The focus of this effort is to identify methods for measuring, calculating, and modeling the LOS performance measures in real-time.
Of particular interest in this research is the work contracted with Purdue University on performance measurements using Input-Output measurements and hybrid techniques for calculating performance measurements using the incoming and outgoing volumes of traffic at individual intersections. These techniques are of interest to ICMS deployments because their focus is on collecting traffic volume data as a means of modeling the current LOS performance measures.
While real-time traffic volumes by segment and intersection are not performance measures of primary interest to the NCHRP 3-79 project, this data, acquired as a byproduct, may be of value to ICMS deployments as a way to measure arterial capacity and the elasticity available for transit pre-emption or capacity shifting from highway to arterial modes.
Several recent publications from the NCHRP have discussed the importance of performance measures. Guidance from NCHRP Report 551 [Cambridge Systematics, PB Consult, and Texas Transportation Institute. NCHRP Report 551: Performance Measures and Targets for Transportation Asset Management. Washington, DC: Transportation Research Board of the National Academies, 2006.] identifies the following criteria for performance measures:
- Policy Driven Measures
- Be sensitive and responsive to policy objectives
- Convey meaningful information about the transportation system
- Strategic Perspective Measures
- Be able to be forecast
- Relate to an economic as well as a technical dimension
- Reflect a combination of outputs and outcomes
- Consideration of Options and Tradeoffs
- Be sufficiently sensitive to reflect impacts of a broad range of options, and potentially, modes
- Help to relate system impacts to factors under the agency's control and to identify impacts of factors not under the agency's control
- Be applicable to scenario testing or "what-if" analysis
- Provide a clear indication of changes in impacts due to different proposed investments, funding levels, and resource allocations
- Enable a linkage analytically between budget and performance while considering the requirements of the Governmental Accounting Standards Board (GASB) 34 modified method
- Be able to relate project outcomes to the program level
- Feedback Measures
- Provide information enabling managers to understand problems and suggest solutions
- Be able to be monitored economically on a periodic basis
- For performance measures dealing with system operations and management, be able to be monitored and provide useful feedback in real-time
- Measures Across Organizational Units and Levels
- Be developed for technical as well as managerial and executive levels within the organization
- Be of a mathematical form that permits aggregation or "rolling up" where appropriate
It should be noted that many traffic engineers do not consider the performance measures defined in the Highway Capacity Manual to be the performance measures of choice for arterial traffic. Many favor the use of volume/capacity ratios for a performance measure. Even this measure becomes problematic when traffic approaches saturation (v/c=1).
7.2 Pioneer Site Performance Measures
Many of the pioneer sites expressed a need for the ability to coordinate ramp metering with adjacent traffic signal controls as a means to improve the performance of both the arterials and the freeways. In order to coordinate these, there must be unused capacity on the ramp and the adjacent traffic signal cycles must be operating such that the arterial has capacity to handle the volume while the ramp meters the vehicles onto or off of the freeway. Performance measures that were identified to assist with this coordination were:
- Vehicle speed
- Intersection approach volumes
- Ramp queues
- Link and spot speeds
- Link and ramp capacity
Each site identified improvement in overall throughput of the corridor as a desired performance measure. In order to demonstrate improvement, the current level of corridor throughput must be calculated across all modes of transportation and all segments of roadway. The following surveillance and detection capabilities as defined in Section 3.2 are needed to calculate the improvement:
- Freeway and Tollway Monitoring
- Transit Monitoring
- Arterial Monitoring
In addition to corridor throughput, each pioneer site identified specific performance measures for the corridor that were desired. These included:
- Travel time including mean, maximum, buffer, and range
- Vehicle speed
- Travel delay time and predictability
- Incident duration and frequency
- Fuel consumption savings
- Pollutant emissions savings
Some of the pioneer sites identified specific performance measures for the freeway, transit, or arterials. The performance measures identified for arterials included:
- Arterial speed based on AVL
- Arterial volume and occupancy
- Arterial capacity
- Arterial segment specific measures which include:
- traffic volume
- travel speeds and times
- level of service
- vehicle miles and vehicle hours traveled
- person-miles and person-hours traveled
- number of incidents and incident rate
- number of fatalities and fatality rate
- number of injuries and injury rate
- incident response and incident clearance time
The key factors in the quality of service, as defined in NCHRP 3-79, are identified as the average speed and the number of stops. Traffic performance monitoring tools use vehicle counts, speeds, and signal status data as inputs in determining the performance of the traffic on a particular segment or link.
7.3 Performance Measures for ICMS
Performance measures figure heavily in the design of arterial transportation systems. Most arterial planning is based on current demand and forecasts of future demand based on demographic measures such as population growth, population shifts, new construction, and economic forecasts of fuel costs. These studies are essentially a demand forecast which is then equated to a capacity requirement. The modeling exercise is focused on determining where to add capacity, how much capacity to add, and how to add the required capacity in the most cost-effective way.
Current models that handle multiple transportation modes typically use a cost per unit of additional capacity as a measure of comparison between alternatives involving multiple modes of transportation within the planning model. As a performance measure for planning, incremental cost of construction for equivalent capacity works for comparing multiple transportation modes.
Incremental cost of construction does not work for real-time management or event planning. This difference between construction planning and operational planning establishes a time-event horizon between construction planning and operational planning and management. Real-time management and event planning must be based on the assets at hand. This means that controls and strategies for operation of corridor assets must be based on using existing assets without exceeding capacity limits.
The forward-looking nature of event planning allows operations staff to move or re-allocate existing transportation resources. For example, real-time responses do not usually result in mode shifts:
- Real-time responses to incidents call for messages on existing Dynamic Message Signs and Highway Advisory Radio in addition to information broadcasts to the public through the media, text messaging, and the Internet.
- Lanes may be closed, and even entire roadways may be closed. In the case of incidents involving transit, schedules may be disrupted or routes dropped for a short time.
- For incidents lasting less than an hour or two, route shifting and travel plan changes are the primary responses to the reduced capacity caused by the incident.
Event planning may involve substantial modifications to the deployment of equipment, and the allocation of lanes, roadways, intersections, and other capacity-related assets. The differences between events (construction, parade, or large venue event) and disaster planning (evacuations, road closures due to flooding, weather, or roadway damage) are more a matter of degree than method. For example, planning can involve capacity changes such as:
- Event related timing plans can be implemented on arterial signal systems
- Intersections can be closed and detour routes established to modify conventional traffic patterns
- Lanes or whole roadways can be closed and detour routes established
- Lanes or roadways can be blocked and put into contra-flow operation to expand capacity in desired directions
- Transit vehicles can be added and/or diverted from established routes to expand transit capacity between desired locations
- Temporary parking areas can be opened to facilitate additional transit capacity
- Portable Dynamic Message Signs and Closed-Circuit Television (CCTV) cameras can be deployed to key locations to provide additional surveillance and direct communication with travelers
The goal of incident management is to keep an incident from cascading into a situation where there is a substantial loss of capacity at a critical time, but incidents do frequently result in major congestion. The goal of ramp metering is to smooth out traffic density and manage volume to avoid recurring congestion on highways (without causing equally detrimental congestion on arterial roadways). The goal of transit priority-based signal strategies is to quickly move people in high occupancy vehicles without creating traffic problems for people in other vehicles. None of these goals are conflicting unless, and until, they negatively impact the capacity of other transportation modes. Viewed as a common goal to maximize the utilized capacity to move goods and people through the corridor, a common feedback value can be derived that is suitable for evaluating and managing use of the corridor transportation resources. This concept establishes a basis for Integrated Corridor Management that can usually be agreed upon by all participants:
Integrated Corridor Management should be based on the concept of reducing or avoiding capacity overloading on all corridor transportation modes and maximizing the volume of people and goods moved through the corridor for any given transportation demand.
To achieve this goal, the ICMS must have two of the following three data types available for arterial roadways:
- The usable capacity (from design or historical data) for each intersection and road segment
- The current traffic volume for each intersection and road segment
- The remaining unused capacity (usable capacity minus current traffic volume)
An effective measure for usable capacity would likely be derived from measured traffic volumes just prior to onset of recurring congestion. The difference between the design capacity and the usable capacity would provide a measure of elasticity.
7.4 Performance Measures for Arterial Management
Arterial networks are highly interconnected meshes of nodes (intersections) connected by road segments. The capacity on the arterial network is a function of the capacity of the intersections and the capacity of the interconnecting road segments. The intersection capacity is a function of geometry, road friction, signal timing, vehicle size, and the number of vehicles in each approach lane. Arterial road segment capacity is a function of geometry, road friction, vehicle size, number of vehicles in each lane, and average speed of the vehicles. Like freeway capacity, the onset of congestion can result in a drop in functional arterial capacity.
To establish real-time monitoring of arterial performance, an ICMS will need the capability to collect current data about speed, volume, and queue lengths at intersections on a lane by lane basis [Balke, Kevin, Hassan Charara, and Ricky Parker. Development of a Traffic Signal Performance Measurement System (TSPMS). Austin: Texas Department of Transportation, 2005.]. In work done by Portland State University in 2007, the arterial signal systems were evaluated for their ability to collect 17 different arterial performance measures. The work in this investigation demonstrates that relatively good performance data can be obtained. This study confirmed that data needs to be collected and sent to an operations center at least once per signal cycle. The study also identified a barrier – that the current generation of signal control hardware is generally incapable of the required level of performance and communication.
One last issue may also need to be considered: the above performance measures treat all vehicles the same. Strictly speaking, high occupancy vehicles and high capacity freight haulers should be given a different weighting factor, as they represent a more efficient movement of people and goods than a single occupancy vehicle used to transport an individual or a small quantity of goods. Signal priority for a loaded transit vehicle makes sense in these terms, but signal priority for an empty transit vehicle may be counterproductive.
7.5 Data for Impact Assessments
Impact assessments are modeling tools that attempt to determine capacity utilization for a given scenario or control algorithm. For planning purposes, the model can be evaluated over a range of values for capacity and under differing conditions to establish a profile of the impact across a range of conditions. Whether the planning is for construction or for a special event, an impact profile can help identify which strategies provide the most desirable outcome.
Impact assessment for real-time control or decision support has different requirements. First, the only input conditions of interest are the current conditions. The only alternative actions or control modes that need to be evaluated are the ones that can be implemented immediately. Responses are usually time constrained. An operator does not have a lot of time to review a complex profile of results.
Data for real-time impact assessment must be timely and accurate. The performance measures discussed in the previous section may be suitable for impact assessments, providing that the measures are current enough to accurately reflect the situation under assessment. The assessment itself must be calculated quickly enough that the recommended control actions are based on input conditions that have not changed substantially.
One interesting phenomenon has been observed relating to traffic impact assessments. Forecasting for weather and usage forecasting for power, natural gas, and water are routinely published. The accuracy of these forecasts is rarely affected by public reaction to the forecasts. Traffic forecasts frequently result in a "self-defeating prognosis" [Bartsch, Matthias. "Unsnarling the Autobahn: How to Forecast Traffic Jams Days in Advance." Spiegel Online, August 13, 2007. http://www .spiegel.de/international/germany/0,1518,druck-499901,00.html (accessed May 5, 2008) .]. Forecasting congestion in a specific area often results in people avoiding the area (given the choice) and the shift in traffic may be enough that the congestion does not materialize. The shift may also be enough that congestion does occur in another location where no congestion was forecast.
While this is good, in the respect that many ICM strategies are based on changing the public behavior, the public may stop responding to ICM strategies if they do not believe that the forecasts are accurate enough to be trusted.
