Congestion Pricing - Links to Congestion Pricing Home
< Previous | Table of Contents | Next >

3.0 DATA ANALYSIS

For the traffic sensor data, the analysis will focus on comparing pre-deployment measures with post-deployment measures. For the Minnesota UPA evaluation, two post-deployment periods will be examined – after expansion of the existing HOV to HOT lanes, implementation of the PDSL, and implementation of some transit projects in September-to-December, 2009, and after implementation of the new HOT lanes in the I-35W/Crosstown Commons section and other project elements in October, 2010.

3.1 Data Cleaning/Error Checking

After downloading the detector data, a series of quality control checks and screening procedures will be applied to the data to identify suspect or invalid data values. The evaluation team will use the quality control checks recommended in NCHRP's Guide to Effective Freeway Performance Measurement. Suspect data will be “flagged” for possible exclusion from the performance measures calculations and reviewed by team members who will determine the final validity and future use of the data. A “bad days” list will be compiled that includes dates, times, and road section that have failed a quality control check. “Bad days” are excluded from valid weekdays on a section-by-section basis. Where possible, procedures will be used to estimate and replace missing and/or erroneous data.

3.2 Data Aggregation and Control for Atypical Conditions

In its rawest form, Mn/DOT RTMC detector data is stored by travel lane per detector in 30-second intervals. To compute the performance measures, the Battelle team will aggregate data, both spatially and temporally. In addition, the data will be assessed for the influence of incidents and other atypical traffic conditions.

Aggregation

  • Spatial Aggregation. Figure 3-1 illustrates the process that will be used to aggregate the data spatially. At the lowest level, Mn/DOT RTMC data is collected on a lane-by-lane basis. Speed and volume data from each detector will first be aggregated across all lanes in each direction to provide detector station values. Data from the HOT lanes and the PDSL will be kept separate from general-purpose freeway lanes. Volume data from lane detectors will be summed to provide a station volume, while speed data will be averaged across all lanes to provide an average detector station speed. Data from slower-moving auxiliary lanes on the freeway will be excluded from this aggregation.

    At the next level of aggregation, detector station data will be converted to link-level data. At this level, a “zone of influence” will be assigned for each detector station. This zone of influence will be equivalent to one-half the distance to the nearest upstream and downstream sensors. Link travel times will be computed by applying the average detector station speed over the zone of influence for each detector station. Vehicle volumes will be subtotaled and multiplied by link length to estimate vehicle-miles of travel (VMT) for each link.

    Graphic representation of a divided highway with six lanes total for four levels of data collection activity. At the lane-by-lane level, traffic sensors collect data in each lane at a nominal spacing of 0.5 mile. At the station level, summary statistics are computed across all lanes in a given direction. At the link level, point-based properties are extrapolated to roadway links with lengths from 0.5 to 3 miles. At the section level, link properties are summed into analysis sections with lengths from 5 to 10 miles.

    Figure 3-1. Methodology for Aggregating Freeway Detector Data for Estimating Directional Route Travel Times and VMT from Spot Speeds and Volumes

    The link-level data will be aggregated to the section and corridor levels. For determining section travel times, the “vehicle trajectory” approach, as opposed to the “snapshot” approach, will be used. The vehicle trajectory method of computing travel time attempts to more closely estimate the actual travel times experienced by motorists. The approach “traces” vehicles trips in time as they progress through a corridor. This is done by applying the link travel time corresponding to the precise time in which a vehicle will be using a link. For example, if it takes a vehicle two minutes to traverse a link at 7:00, then the link travel time starting at 7:02 would be used as the travel time for next downstream link. This process is continued for all the links that make up section or corridor. The “snapshot” approach sums all the link travel times at the same instance in time. The vehicle trajectory method is routinely used in post hoc evaluations while the snapshot method is traditionally used to computing travel times for display on traveler information systems.

  • Temporal Aggregation. In addition to aggregating the data spatially, the individual detector data will be aggregated temporally. Thirty-second detector data will be aggregated to 5-minute intervals. This means that each 30-second vehicle count will be summed to provide a total number of vehicles in the 5-minute interval, while speed and occupancy data will be averaged to provide an average speed and occupancy for the 5-minute interval.

Incidents and Other Atypical Traffic Conditions. The Minnesota UPA projects are focused primarily on addressing the effects of typical, or recurring, congestion on traffic operations; however, “atypical” (or non-recurring) conditions are also a major source of congestion in most metropolitan areas. Because incidents and other external events can be major sources of variations in travel time, the effects of these events can dramatically influence the results of the analysis. Therefore, the national evaluation will analyze the effects of the UPA projects on both recurring congestion and non-recurring congestion.

There are a number of conditions that may result in non-recurring congestion, including traffic incidents, work zones, weather, fluctuations in demand, special events, traffic control devices, and inadequate base capacity. The effects of the following non-recurring events will be examined in the evaluation.

  • Incidents. Incidents are a source of travel time variability in the I-35W corridor. An incident is any event that unexpectedly and temporarily reduces capacity of the freeway and roadways. Incidents can range in nature from debris on the roadway to stalled vehicles blocking travel lanes, to major collisions and crashes. Those periods when traffic flow is severely impeded by incidents will be flagged in the data and their impacts will be analyzed separately from those days when traffic operations are “normal.” Mn/DOT RMTC Operator logs and FIRST Dispatch Logs will be used as the primary source of information for when incident conditions exist in the I-35.
  • Inclement Weather. Because weather can have significant impact of freeway performance, data from the National Weather Service archives as well as RMTC operator logs will be used to identify those data were inclement weather occurred. To determine when winter weather conditions might have impacted traffic operations, Mn/DOT maintenance records will be examined to determine those days when snow plowing or sanding operations occurred. National Weather service records will be used to flag days that are impacted by weather conditions. These would include days reporting limited visibility due to fog, rain or snow, and days experiencing significant ice, rainfall, or snowfall events. Days identified as being potentially impacted by inclement weather will be flagged and data from those days will be analyzed separately.
  • Roadway Construction and Maintenance. Roadway construction and maintenance can also be as source of congestion. It will be important to also exclude days when performance of the roadway is severely impacted by roadway construction and maintenance from the “typical” day analysis and include these days in the incident data. Mn/DOT Maintenance Records and RTMC Operator logs will be used to identify when and where construction and maintenance activities on I-35W impacted operations.

3.3 Performance Measure Calculation Procedures

The input data and procedures for calculating the primary performance measures for the congestion analysis are described in this section. The measures presented include travel time, vehicle throughput, person throughput, VMT, person-miles of travel (PMT), the travel time index, the planning time index, the buffer index, hours of congestion, and the number of congested links.

In analyzing the effects of the UPA projects on traffic operations, the test plan focuses on the morning and afternoon peak periods. The peak period analysis is the time periods when the UPA projects are mostly likely to have a significant impact on reducing congestion. In producing their annual congestion report on the Metropolitan Freeway System, the RTMC defines the peak periods as 6:00 a.m. to 9:00 a.m. and 2:00 p.m. to 7:00 p.m. These same time periods will be used to define the peak periods for the UPA evaluation. This approach will highlight where projects had an impact on the duration of congestion, as well as the level of congestion.

Travel Time. Many of the analyses use performance measures that are derived from travel time, which is the amount of time required to traverse the entire length or a predefined section of a corridor or roadway. For the purpose of this evaluation, travel times will be determined using the detector data from Mn/DOT's RTMC Detection System. The input data required to compute segment and corridor level travel times are the 5-minute average link-level speeds for peak traffic periods (6:00-9:00 a.m., 2:00-7:00 p.m.) and the “zone of influences” link length for each detector stations.

Vehicle Throughput. Vehicle throughput is a measure of the number of vehicles that are serviced in one direction of a facility during the analysis period. Mn/DOT freeway detectors measure the vehicle throughput for each individual link, while vehicle throughput on the arterial will be measured at the screen lines. Total vehicle throughput is the sum of all types of vehicle traversing the roadway or screen line during the analysis period (peak hour, peak period and/or day).

Person Throughput. Person throughput, or person volume, is the total number of people serviced in the corridor during the analysis period. Person throughput is the average vehicle occupancy multiplied by the volumes of vehicles within the corridor.

Person Throughput (PT) = Vehicle volume (V) * Average Vehicle Occupancy (AVO)

For the purposes of the Minnesota UPA, the contribution of different strategies to overall person throughput is of interest. Person throughput will be computed for each travel mode and estimated using average vehicle occupancy rates for different types of vehicles. In this process, the contributions to total corridor person throughput is computed using the following equation:

PT Total = PT Transit (External) + PT TDM + PT Arterial + PT General Purpose Lanes + PT Tolled Lanes

Table 3-1 presents each of the elements used in computing the total person throughput in the I-35W corridor.

Table 3-1. Component of Computing Total Person Throughput in Analysis Corridor
PT Contribution Analytical Source Description
PT Transit (External) Transit Analysis PTTransit (External) is the relative contribution of the transit-based UPA projects on other freeways/areas. (Example: park-and-ride improvements on I-35W north of the UPA corridor)
PTTDM TDM Analysis PTTDM is the relative contribution of the telecommuting program on other freeways/areas.
PT Arterial = PT Non-freeway auto + PT Non-freeway transit Congestion Analysis PTNon-freeway auto. is the PT from vehicles on adjacent facilities in the I-35W corridor, excluding buses.
PT Arterial = PT Non-freeway auto + PT Non-freeway transit Transit Analysis PTNon-freeway transit is the number of persons serviced by transit routes on adjacent facilities. It is computed as the number of transit vehicle times the average occupancy rate per transit vehicle
PT General Purpose Lanes = PTGPL transit + PTGPL auto Transit Analysis PTGPL transit is the relative contribution of transit vehicle traveling on the freeway general purpose lanes. It is computed by multiplying the average occupancy per transit vehicle for any routes / partial routes that use the general purpose lanes for the I-35W corridor.
PT General Purpose Lanes = PTGPL transit + PTGPL auto Congestion Analysis P GPL auto is the relative contribution of autos on the general purpose lanes in the I-35W corridor.
PT Tolled Lanes = PTTL Transit + PTTL HOV2 + PTTL van + PTTL SOV Transit Analysis PTTL transit is the relative contribution of transit vehicle traveling in the HOT lanes and the PDSL. It is computed using the average occupancy per route for any routes / partial routes that use the HOT lanes and the PDSL in the I-35W corridor.
PT Tolled Lanes = PTTL Transit + PTTL HOV2 + PTTL van + PTTL SOV Tolling Analysis PTTL SOV is the PT from percentage of HOT lane and PDSL traffic that was SOV (% traffic X volume).
PT Tolled Lanes = PTTL Transit + PTTL HOV2 + PTTL van + PTTL SOV Tolling Analysis PT TL HOV2 is the PT from percentage of HOT lane and PDSL traffic that was 2-person carpools (% traffic X number of occupants X volume).
PT Tolled Lanes = PTTL Transit + PTTL HOV2 + PTTL van + PTTL SOV Tolling Analysis PT TL Van is the PT from percentage of HOT lane and PDSL traffic that was vanpools (% traffic X number of occupants X volume).

Volume counts and occupancy rates of transit vehicles on each of the facilities will be determined based on data from the transit test plan. The number of non-transit vehicles using the HOT lanes and PDSL and the I-35W general purpose freeway lanes will be computed by subtracting the number of transit vehicle using the HOT lane and PDSL and the general-purpose freeway lanes from the total vehicle count of the Mn/DOT RTMC detector. The number of non-transit vehicle using the arterial streets will be used subtracting the number of transit vehicles crossing the screen line locations.

VMT. VMT is a common measure of throughput. It is the product of the number of vehicles traveling over a length of roadway times the length of the segment of roadway. It is computed using the following equations:

Begin expression open parenthesis begin uppercase VMT end uppercase close parenthesis begin subscript 'Link' end subscript equals 'Volume' open parenthesis begin uppercase V end uppercase close parenthesis begin subscript 'Link' end subscript operator times 'Link Length' end expression. Begin expression begin uppercase VMT end uppercase begin subscript 'Total' end subscript equals summation sign begin uppercase VMT end uppercase begin subscript 'Link' end subscript end expression.

To compute VMT for the general-purpose freeway lanes and the HOT lane and PDSL, 5-minute vehicle counts from Mn/DOT RTMC detector systems will be used. VMT will be computed for each link. Similarly, screen-line count data will be multiplied by the segment length for which the volume counts are deemed to represent. Segment level and corridor level VMTs will be calculated by summing link-level VMTs from the freeway, HOT lanes, PDSL, and arterial links respectively. Peak hour, peak period, and daily VMT will be calculated using this approach.

When computing average index values across road sections or time periods, VMT is used to weight the calculations. VMT can also be used to weight the average travel time index calculation across all weekdays; however, in most cases, the differences are not significant (for weighted vs. non-weighted averages) when computing averages across weekdays.

Person-Miles of Travel (PMT). Similar to VMT, PMT is a measure of throughput and is the product of passenger throughput times the length of segment of roadway. The basic equation for computing PMT is as follows:

Begin expression open parenthesis begin uppercase PMT end uppercase close parenthesis begin subscript 'Link' end subscript equals 'Person Throughput' open parenthesis begin uppercase PT end uppercase close parenthesis begin subscript 'Link' end subscript operator times 'Link Length' end expression. Begin expression begin uppercase PMT end uppercase begin subscript 'Total' end subscript equals summation sign begin uppercase PMT end uppercase begin subscript 'Link' end subscript end expression.

As discussed above, passenger throughput is the product of the number of specific classes of vehicles (transit, HOV, SOV vehicles) traversing a length of roadway times the average number of occupants in each vehicle class. PMT is the product of person throughput from each detector station in the corridor by the link length. Segment- and corridor-level PMT is computed by summing all the link-level PMTs across all the links defined in the segment or corridor. Peak hour, peak period, and daily PMT will be calculated using this approach.

Travel Time Index. The Travel Time Index (TTI) is the extra time spent in traffic during peak traffic times as compared to light traffic times. In mathematical terms, it is the peak travel time divided by the free-flow travel time as in the following equation:

Begin expression begin uppercase TTI end uppercase equals begin fraction begin numerator open bracket begin fraction begin numerator 'Average Speed' end numerator over begin denominator 'Free-flow Speed' end denominator end fraction operator times begin uppercase VMT end uppercase close bracket begin subscript 'Freeway' end subscript operator plus open bracket begin fraction begin numerator 'Average Speed' end numerator over begin denominator 'Free-flow Speed' end denominator end fraction operator times begin uppercase VMT end uppercase close bracket begin subscript 'Arterial' end subscript end numerator over begin denominator begin uppercase VMT end uppercase begin subscript 'Freeway' end subscript operator plus begin uppercase VMT end uppercase begin subscript 'Arterial' end subscript end denominator end fraction end expression.

To compute the TTI, one first computes the free-flow travel speed for each section as the 85th percentile speed during the previous 3 months during weekday off peak times, which are defined as weekdays from 9:00 a.m. to 4:00 p.m. and 7:00 p.m. to 10:00 p.m. and weekend/ holiday times from 6:00 a.m. to 10:00 p.m. TTI is computed by dividing the free-flow traffic speed by the peak traffic speed. If the computed travel time index is less than 1.00, it will be rounded up to 1.00. The TTI for the I-35W corridor for the each day is computed by taking a VMT-weighted average of all travel time indices during peak times. Monthly average travel time index are computed by taking a simple (or VMT-weighted) average across all valid weekdays.

Planning Time Index. The Planning Time Index (PTI) is the extra time cushion needed during peak traffic periods to prevent being late. In mathematical terms, it is the near-worst case travel time (95th percentile) divided by the free-flow travel time.

Begin expression begin uppercase PTI end uppercase equals begin fraction open bracket begin numerator begin fraction begin numerator '95th Percentile Travel Time' end numerator over begin denominator 'Free-flow Speed' end denominator operator times begin uppercase VMT end uppercase close bracket end fraction begin subscript 'Freeway' end subscript operator plus open bracket begin numerator '95th Percentile Speed' end numerator over begin denominator 'Free-flow Speed' end denominator end fraction operator times begin uppercase VMT end uppercase close bracket begin subscript 'Arterial' end subscript end numerator over begin denominator begin uppercase VMT end uppercase begin subscript 'Freeway' end subscript operator plus begin uppercase VMT end uppercase begin subscript 'Arterial' end subscript end denominator end fraction end expression.

If using speeds in calculations, it is the free-flow traffic speed divided by the near-worst case traffic speed (5th percentile). All weekdays in the month are included in the calculation of 5th percentile speeds. The PTI is computed as the free-flow traffic speed divided by the 5thpercentile traffic speed for each freeway section and 5-minute peak time period. If the planning time index is less than 1.00, it will be rounded up to 1.00. The average planning time index for each freeway section is computed by taking VMT-weighted averages across all 5-minute index values.

Buffer Index. The Buffer Index (BI) is the extra time (buffer) that travelers in a corridor need to allow to ensure an on-time arrival for most trips. The BI is equivalent to the extra time travelers must add to their average travel time when planning trips. With continuous data, such as the Mn/DOT RTMC detector data, the index will be calculated for each road or transit route segment, and a weighted average will be calculated using vehicle-miles or, more desirably, person-miles of travel as the weighting factor. The BI can be calculated for each road segment or particular system element using the following equations:

Begin expression 'Buffer Index (BI)' begin subscript 'Link' end subscript equals open bracket begin fraction begin numerator '95th Percentile Travel Time' operator minus 'Average Travel Time' end numerator over begin denominator 'Average Travel Time' end denominator end fraction close bracket operator times 100% end expression. Begin expression 'BI' begin subscript 'Corridor' end subscript equals begin fraction begin numerator summation sign open parenthesis 'BI' begin subscript 'Link' end subscript operator times begin uppercase VMT end uppercase begin subscript 'Link' end subscript close parenthesis end numerator over begin denominator summation sign uppercase VMT end uppercase begin subscript 'Link' end subscript end denominator end expression.

Note that a weighted average for more than one roadway section could be computed using VMT or PMT on each roadway section. The measure would be explained as “a traveler should allow an extra BI percent travel time due to variations in the amount of congestion delay on that trip.”

Hours of Congestion. This measure is intended to measure the temporal extent of congestion. It is the average length of time each period and/or day that a particular facility is operating in a congested state. While each city may define when a facility is congested differently, the Minneapolis UPA national evaluation will use the threshold used by Mn/DOT in their Metropolitan Freeway System 2007 Congestion Report, which defines congestion when speeds on a freeway link drop below 45 mph.

Hours of congestion will be computing using the speed measures reported by Mn/DOT in the RMTC detector data. Examining the detector data, links will be flagged when the reported 5-minute average speed drops below 45 mph. The uncongested and congested links in the corridor will be counted by each 5-minute time period and the percentage of congested links will be computed. If the percentage of congested links in the corridor exceeds 20 percent, the corridor will be identified as being congested for that 5-minute time period and day. For each valid weekday in the month, the number of congested 5-minute periods will be used to determine the total duration of congestion, daily and during the peak period. Only time periods between 6:00 a.m. and 10:00 p.m. will be included in this analysis.

Percent of Congested Lane Miles. The percent of congested lane miles measures the spatial extent of congestion. This measure is the lane miles of roadway within the I-35W corridor that the average travel times are 30 percent longer than the unconstrained travel time. For each 5-minute time period during the analysis period, the actual travel time in each lane on a link will be compared to the unconstrained, or free-flow, travel time for that link. If the actual link travel time is 30 percent longer, then the lane on that link will be defined as congested. The performance measure, which is the percentage of congested roadways, is calculated as the ratio of congested lane miles to the total lane miles. Days where incidents have occurred, inclement weather, or maintenance and/or construction activities are present will be analyzed separately from typical workdays.

3.4 Time Period for Analysis

In addition to aggregating the data spatially, data and performance measures will be aggregated temporally. For the purpose of the Minnesota UPA national evaluation, performance measures will be computed for peak hours, the peak period, and daily.

Peak hour statistics provide an indication of corridor performance when congestion is at its worst. The evaluation will use two methods to define the peak hour. The first method is the traditional method of determining the peak hour by applying the Highway Capacity Manual's definition of peak hour, which is the one-hour period experiencing the highest hourly volume. Typically, this is the one-hour period when throughput is at its maximum, just before congestion forms; however, because the UPA projects are intended to lessen the impacts of congestion, the evaluation will also use an alternative definition to determine the peak hour. This definition is the one-hour period when travel speeds are at their lowest. This alternative definition defines when congestion is at its worst. Typically, the peak hour based on lowest speeds lags the peak hour based on highest volumes by 30 to 60 minutes. The evaluation team will compute performance statistics for both morning and afternoon peak hour using both definitions.

In addition to computing peak hour statistics, the national evaluation will also compute peak period performance statistics. The peak period is a time period that extends beyond the peak hour. It not only includes the peak hour but also incorporates the shoulders of the peak hour when traffic congestion is building and dissipating. As noted previously, Mn/DOT defines the morning peak period from 6:00 a.m. to 9:00 a.m. and the evening peak period from 2:00 p.m. to 7:00 p.m.

The national evaluation team will also compute daily performance measure statistics. Analyses using daily averages are often less useful with the Travel Time Index and the Buffer Index. Using 24-hour speeds for computing the Travel Time Index is not meaningful because the measure is meant to compare peak and off-peak travel conditions. Likewise, the Buffer Index is intended to be a measure of reliability during a peak period. Daily values wash out the impact of peak-periods with the longer off-peak periods. Where appropriate, daily statistics will consist of the time between 6:00 a.m. and 10 p.m. and will exclude late night/early morning conditions.

For the purposes of this evaluation, only data from non-holiday weekdays will be included in the analysis. Data from the weekends and from federal and state holidays will be excluded from the analysis as traffic conditions are not typical on these days. The data may also exhibit significant season variations, such as summer versus winter and school and university in sessions/out of session. While the national evaluation team does not envision needing to conduct separate analysis for different seasons, data will be examined to determine if significant seasonal variations exist that might influence the overall analysis.

3.5 Types of Analysis

After calculating the performance measures, several types of analysis procedures will be used to assess the effectiveness of the Minnesota UPA projects. The performance measure will be tracked from the pre- and post-deployment time periods. This trend analysis will assist in determining the following elements:

  • whether or not changes in performance were due to the UPA projects or to typical daily, monthly, and seasonal variations in traffic operations in the corridor;
  • whether or not the effects of the implementing the Minnesota UPA projects, pricing in particular, in the I-35W corridor were sustainable during the post-deployment period; and
  • whether or not changes in the performance measures were isolated to the I-35W corridor or if similar trends in performance were occurring in other corridors in the area.

In addition to analyzing trends in the performance data, traditional statistical analysis, such as comparison of means and analysis of variance, will be performed to determine whether or not changes in level of congestion differed statistically during the pre- and post-deployment time periods. Table 3-2 presents the null and alternative hypotheses of the statistical comparisons envisioned in this evaluation. All statistical comparisons will be performed using standard statistical procedures and conducted at a 95 percentile confidence level.

Table 3-2. Anticipated Statistical Comparisons of UPA Performance Measures
Analysis Performance Measure Null Hypothesis Alternate Hypothesis
Quality of Service Average Travel Time (begin overbar begin uppercase TT end uppercase end overbar) Begin expression begin overbar begin uppercase TT end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase TT end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase TT end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase TT end uppercase end overbar begin subscript 'Before' end subscript end expression.
Quality of Service Average Travel Time Index(begin overbar begin uppercase TTI end uppercase end overbar) Begin expression begin overbar begin uppercase TTI end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase TTI end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase TTI end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase TTI end uppercase end overbar begin subscript 'Before' end subscript end expression.
Reliability Average Planning Index (begin overbar begin uppercase PI end uppercase end overbar) Begin expression begin overbar begin uppercase PI end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase PI end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase PI end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase PI end uppercase end overbar begin subscript 'Before' end subscript end expression.
Reliability Average Buffer Index (begin overbar begin uppercase BI end uppercase end overbar) Begin expression begin overbar begin uppercase BI end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase BI end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase BI end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase BI end uppercase end overbar begin subscript 'Before' end subscript end expression.
Throughput Average Vehicle-Miles of Travel (begin overbar begin uppercase VMT end uppercase end overbar) Begin expression begin overbar begin uppercase VMT end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase VMT end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase VMT end uppercase end overbar begin subscript 'After' end subscript operator greater than begin overbar begin uppercase VMT end uppercase end overbar begin subscript 'Before' end subscript end expression.
Throughput Average Passenger-Miles of Travel (begin overbar begin uppercase PMT end uppercase end overbar) Begin expression begin overbar begin uppercase PMT end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase PMT end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase PMT end uppercase end overbar begin subscript 'After' end subscript operator greater than begin overbar begin uppercase PMT end uppercase end overbar begin subscript 'Before' end subscript end expression.
Extent of Congested Average # of Congested Hours (begin overbar begin uppercase CHR end uppercase end overbar) Begin expression begin overbar begin uppercase CHR end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase CHR end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase CHR end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase CHR end uppercase end overbar begin subscript 'Before' end subscript end expression.
Extent of Congested Average # of Congested Links (begin overbar begin uppercase CLK end uppercase end overbar) Begin expression begin overbar begin uppercase CLK end uppercase end overbar begin subscript 'After' end subscript equals begin overbar begin uppercase CLK end uppercase end overbar begin subscript 'Before' end subscript end expression. Begin expression begin overbar begin uppercase CLK end uppercase end overbar begin subscript 'After' end subscript operator less than begin overbar begin uppercase CLK end uppercase end overbar begin subscript 'Before' end subscript end expression.

The performance measures calculated using traffic data apply to many parts of the evaluation. Examples of the analysis to be conducted are presented below.

  • Change in travel times on I-35W. The change in travel time for different user groups and different time periods on I-35W will be analyzed pre- and post-deployment. Changes in trip travel times in the general-purpose freeway travel lanes and the HOT lanes will be examined, as will changes in the Travel Time Index. Travel time measures are used not only in the congestion analysis but also in the goods movement and cost benefits analyses.
  • Change in trip-time reliability on I-35W. The change in trip-time reliability for different user groups and different time periods on I-35W will be analyzed pre- and post-deployment. Changes in trip-time reliability in the general-purpose freeway lanes and HOT lanes will be examined pre- and post-deployment. The percentage change in variability, the change in the Buffer Index, and the change in the Planning Index will be examined.
  • Changes in the hours per day of congestion conditions on I-35W. The number of hours I-35W operated in congested conditions (travel speeds less than 45 mph) will be analyzed pre- and post-deployment. This analysis will examine the various links and segments on I-35W, as well as the extent of congested conditions during the day.
  • Change in the number of vehicles served. The change in the number of vehicles served on I-35W (vehicle throughput) will be examined pre- and post-deployment. Changes in vehicle throughput will be examined for the various links, as well as different times of the day.
  • Change in the number of persons served. The change in the number of persons served on I-35W (person throughput) will be evaluated pre- and post-deployment. Person throughput will be examined for the various links, as well as different times of the day.
  • Change in traffic congestion on surrounding facilities. To the extent allowed by available data, changes in traffic congestion on other facilities in the corridor will be evaluated.
  • Changes in VMT and PMT. Changes in VMT and PMT on I-35W pre- and post-deployment will be evaluated. VMT is a key measure in several of the analysis areas including environmental analysis, cost benefit analysis and congestion analysis.
  • Assessment of impact of traffic management strategies. The impact of UPA traffic management strategies, including speed harmonization, on traffic congestion on I-35W will be analyzed. This analysis will examine sensor and other traffic data when the speed harmonization strategies are being used, with the time periods they are not in operation. The number and duration of incidents when the active traffic management strategies are in operation and when they are not will also be examined.