SUMMARY

This project aimed to develop rational methods for assigning short-duration traffic volume counts to adjustment factor groups for estimating annual average daily traffic (AADT). The research was concerned with all roadway functional classes and traffic volumes and involved addressing a series of frequently asked questions. The first phase of the project involved reviewing the literature, conducting a national survey of state agencies, and identifying methods for further consideration. In the second phase of the research, 19 assignment methods were developed and validated using continuous count station (CCS) data from 45 states – three years per state. More than 10,000 CCS-year combinations were used in this validation. The methods were divided into three broad groups: (a) existing methods, (b) improved methods, and (c) new methods. The table below briefly describes all methods.

To validate Methods 18 and 19, probe data were obtained from three states – one vendor per state. The main findings from the validation of these methods are summarized below.

Assignment Methods’ Effectiveness

Cluster analysis produces more homogeneous groups, more precise group factors, and slightly more accurate AADT estimates than the other methods. Annualizing counts using probe-based adjustment factors (M18) is a promising method that can potentially produce more accurate AADT estimates than traditional methods as long as the penetration rate of the probe data is sufficient and the correlations between probe-based and actual adjustment factors are strong (>0.85). Further, some traditional methods such as FC_RU (M3) tend to produce slightly less accurate AADT estimates than cluster analysis, but the development of CCS groups and the assignment of counts to groups are straightforward.

Important Assignment Attributes

The most important attributes used in DTs (M14) and SVMs (M15) are hourly factors by day of year, followed by hourly factors developed by month and day of week. The most important attributes used to develop clusters are the 84 monthly day-of-week factors, followed by the 12 monthly factors. Attributes with comparatively lower importance include census variables and traffic volumes.

Functional Classification vs. Volume Factor Groups

The results of these methods vary considerably from one state to another. When CCSs from all functional classes were taken into consideration, all three methods—functional classification (M2), FC_RU (M3), five volume groups (M4), and 10 volume groups (M5)—resulted in similar AADT accuracy across all states. FC_RU (M3) and FC (M2) performed slightly better than volume groups (M4 and M5) in lower functional classes, FC6 and FC7. Dividing five volume groups (M4) into 10 volume groups (M5) did not improve AADT accuracy.

AADT Accuracy of Cluster Analysis

Based on the average results across all states, cluster analysis tends to produce significantly more homogeneous clusters than the traditional grouping methods, though the improvement in AADT accuracy is not as pronounced. For instance, the overall within-group variability of seven clusters (M10) is around 6.4 percent, while that of the seven FCs (M2) is much higher (11.8 percent). The mean absolute percent error (MAPE) of cluster analysis (M10) was 6.7 percent, whereas the corresponding error from FC (M2) was higher, 9.3 percent. However, the results vary from one state to another. For example, in some states, clustering improved AADT accuracy over traditional methods by more than 60.0 percent, whereas in other states, no improvement was observed. The decision to use cluster analysis over traditional methods should be made by practitioners based on state-specific results, goals, and available resources.

AADT Accuracy of Combined Assignment Methods

The combination of two or more methods can have varying effects on AADT accuracy. It may slightly increase the AADT accuracy or slightly reduce it but mitigate individual method limitations. These effects depend on the methods to be combined. For example, factor groups developed by FC_RU (M3) yielded slightly lower AADT estimation errors, around 0.5 percent, than developing groups only by FC (M2). In contrast, combining cluster analysis (M10) with DTs (M14) tends to reduce the AADT accuracy by 1.0–2.0 percent, but facilitates the assignment process, making it data driven. When FC is combined with cluster analysis, the latter tends to become slightly less effective, but the clusters are better defined.

Ease of Defining Clusters from a Practical/Assignment Standpoint

Some clustering methods, such as the k-prototypes partitioning algorithm and hierarchical clustering, can handle both continuous and categorical variables, which, to some extent, assist in improving the definition of clusters by incorporating some assignment characteristics. Clustering may be a good starting point for identifying similar patterns and determining factor groups. Ultimately, analysts have to manually review and refine the produced clusters.

Important Clustering Attributes

The most important attributes used to develop clusters are the 84 monthly day-of-week factors, followed by the 12 monthly factors. Other variables that can be used to better define the produced clusters include the FC, area type, land use, and geographical coordinates of the CCSs.

Other Data-Driven Assignment Methods

This project applied and validated DTs and SVMs and found that the latter are slightly more effective than DTs but have a more complex structure. Assigning counts to clusters using DTs or SVMs resulted in more accurate AADT estimates than assigning them to individual CCSs within the same FC.

AADT Accuracy of Counts Factored Using Factors from Higher Functional Classes

For FC6 and FC7, the average AADT estimation errors of weekday counts annualized using adjustment factors from FC5 (M6) or from 5R and 5U (M7) exceeded 13 percent. In comparison, other traditional methods such as FC (M2) yielded more accurate AADT estimates (MAPE = 11.8 percent) than M6 and M7. For rural local roads (7R), counts annualized using factors from 6R (M9) resulted in MAPEs of 11.3 percent, which were slightly higher than those (10.6 percent) obtained from FC (M2).

Importance of Factoring Counts

Among all methods examined in this project, not factoring counts (M1) yields the least accurate AADT estimates. Simple traditional methods such as FC (M2) and FC_RU (M3) are expected to improve the AADT accuracy of weekday counts by approximately 26 percent and 30 percent, respectively. Cluster analysis combined with SVMs (M15) improved the AADT accuracy of weekday and weekend counts by more than 37 percent and 50 percent, respectively. Likewise, annualizing counts using probe-based factors (M18) that are strongly correlated with actual adjustment factors can considerably increase the AADT accuracy of unfactored counts (M1).

Factoring Counts on Low-Volume Roads when no CCSs are Available

Agencies should ideally operate CCSs on low-volume roads. If CCSs are not available on local roads (FC7), one alternative is to use group adjustment factors from 6R and 6U to annualize counts taken on 7R and 7U, respectively. Another option is to use volume factor groups. A promising option is to use segment-specific probe-based adjustment factors provided that the latter are strongly correlated with actual adjustment factors.

Probe-Based Temporal Adjustment and Axle-Correction Factors

Probe data can be a viable data source for calculating temporal adjustment factors, but the effectiveness of the latter depends on the (a) penetration rate of the probe data, (b) representativeness of the entire population, (c) correlations with actual adjustment factors, and (d) the types of adjustment factors being used. As of the publication date of this report, it is more difficult to develop accurate axle-corrections factors due to the limited availability of probe data for specific vehicle classes. Though exceptions exist, in general, the higher the penetration rate of the probe data, the stronger the correlation between probe and actual factors, which, in turn, results in more accurate AADT estimates developed from counts annualized using probe factors.

Accuracy of Probe-Based AADT Estimates

The following table shows the average penetration rates, correlations, and AADT estimation errors produced using probe data from three different vendors and states (M18).

These results hold true only for the three states and the probe datasets that were analyzed in this project. Different results may be obtained using data from different vendors.

Probe-Based Adjustment Factors for Lower Roadway Functional Classes

Probe-based factors can be used to factor counts on FC6 and FC7 as long as they are strongly correlated with actual adjustment factors. The correlation analysis performed using Vendor A’s seven annual day-of-week factors revealed that the correlations increased from higher to lower functional classes; however, a larger sample size is needed to draw more robust conclusions for Vendor A’s data. The main advantages of probe-based factors over other grouping and assignment methods include a) the ability to develop one or more sets of probe-based adjustment factors separately for every roadway segment of the entire network; b) the elimination of the need to create factor groups; and c) the elimination of the need to assign counts to factor groups.

Penetration Rate and Accuracy of Probe-Based Factors

In general, higher penetration rates tend to more effectively capture temporal changes in traffic volumes, leading to stronger correlations between probe-based factors and actual adjustment factors. This in turn increases the accuracy of probe-based AADT estimates developed from annualized counts. However, many factors may affect the penetration rates, the correlations, and thus the AADT accuracy, such as the types of probe data being used and their representativeness of the entire population. If the probe data only capture a specific subset of the traffic or are biased toward certain conditions or a specific area, they may not be fully representative of the entire population over time, leading to discrepancies between estimated and actual AADT values.

Another challenge that can lead to poor correlations and AADT accuracy is vendors obtaining and adding new raw probe data from new sources to increase the penetration rate of their raw probe data and the accuracy of the data products that they develop. This ultimately causes sudden changes in the temporal traffic patterns captured from raw probe data, particularly if the dates on which new data sources are added are unknown or are not taken into consideration when probe-based adjustment factors are developed. Something similar can happen when existing sources of probe data are discontinued or abandoned by a vendor.