A Study on the Classification of Road Type by Mixture Model

Road classification system is the first step for determining the road function and design standards. Currently, roads are classified by various indices such as road location and function. In this study, we classify road using various traffic indices as well as to identify traffic characteristics for each type of road. To accomplish the objectives, mixture model was applied for classifying road and analyzing traffic characteristics using traffic data that observed at permanent traffic count stations. A total of 8 variables were applied: annual average daily traffic(AADT), coefficient, heavy vehicle proportion, day volume proportion, peak hour volume proportion, sunday coefficient, vacation coefficient, and coefficient of variation(COV). A total of 350 permanent traffic count points were categorized into three groups : Group I (Urban road), Group II (Rural road), and Group III (Recreational road). AADT were 30,000 for urban, 16,000 for rural, and 5,000 for recreational road. Group III was typical recreational road showing higher average daily traffic volume during Sunday and vacational periods. Group I showed AM peak and PM peak, while group II and group III did not show AM peak and PM peak.