Research | 2013
Prioritizing Access Management Implementation
Schultz, G., K. Bradley, and T. Boschert. Prioritizing Access Management Implementation, Transportation Research Record: Journal of the Transportation Research Board, No. 2092, Transportation Research Board of the National Academies, Washington, D.C., (2009, pp. 57-65)
Using principles of performance indices, this research develop a performance-index-based prioritization process for targeting arterial road segments that would benefit from access management, and suggesting appropriate access management techniques and treatments. A database was created of identifying features, roadway geometric and traffic characteristics, and crash history for 175 arterial highway segments. Stepwise linear regression was applied to the data collected to determine which characteristics of the roads were correlated with crash rate, crash severity, and collision type.
Statistical analyses showed that the lack of access management techniques (e.g., high access density, numerous signals per mile, and no medians) were positively correlated with increased crash rate and severity as well as certain collision types. Adjacent land use was also identified as being highly correlated with the safety of arterial roads as those road segments with adjacent commercial land use tended to have higher crash rates and severity scores.
A decision tree was developed to classify existing or future road segments into subcategories based on volume, signal spacing, adjacent land use, and other criteria. Various sources were consulted to determine the optimal way to partition the decision tree data into relatively homogenous groupings (roadways having similar characteristics and crash histories). These included the Classification and Regression Trees (CART) method developed by Breiman et al.(1984), the literature, the distribution of data collected, and so on. Appropriate countermeasures could then be determined for the given segments. Access management techniques were recommended for each subcategory based on correlations between access management techniques and crash severity score . Volume and signal spacing were found to be the two most important variables in determining crash severity score.
The six steps of the decision tree are summarized in the paper as follows:
Step 1: Obtain Data
The first step in the decision tree is to collect data for the road segment being analyzed, including the AADT, signals per mile, adjacent land use, and potential for future development.
Step 2: Classify by Volume
Next, the road segment being analyzed is categorized as having low, medium, or high volume corresponding to less than or equal to 15,000 vpd, greater than 15,000 vpd and less than or equal to 25,000 vpd, or greater than 25,000 vpd, respectively. Alternatively, a road segment could also be categorized by future expected volume in order to determine future needed access management treatments.
Step 3: Classify by Signals per Mile
Following classification by volume, the road segment is classified by signal spacing based on whether the segment has 2 or less signals per mile or greater than 2 signals per mile.
Step 4: Classify by Land Use
Depending on the classification of the road segment according to its volume and signals per mile, segments are further classified as having either adjacent commercial or residential land use.
Step 5: Other Classification
Based on the overall characteristics of the road, some road segments are classified according to the potential for future growth. Additionally, high volume arterials with greater than 2 signals per mile are classified according to median type (i.e., raised median or no raised median).
Step 6: Recommended Access Management Treatments
Access management treatments are recommended based on the classification from steps two through five, including limit access density, install raised median, future planning, and no recommendation.