The University of Massachusetts Amherst
University of Massachusetts Amherst

Search Google Appliance


Automatic horizontal curve identification and measurement method using gps data

TitleAutomatic horizontal curve identification and measurement method using gps data
Publication TypeJournal Article
Year of Publication2015
AuthorsAi C, Tsai Y
JournalJournal of Transportation Engineering
Date Published02/2015

Horizontal curves play a critical role in roadway safety by providing a smooth transition between tangent sections. Because radii of horizontal curves are one of the most fundamental elements in roadway geometry design, transportation agencies, e.g., state DOTs, need to measure them to support network-level safety analysis. However, the traditional methods that are commonly used by transportation agencies, e.g., plan sheet reading method and chord-offset method, are time consuming, labor intensive, and inaccurate. Although some semiautomatic and automatic methods have been developed using global positioning system (GPS) data and/or geographic information system (GIS) functions in recent years, these methods are not yet ready to be practically used in a network-level analysis because they either require intensive manual intervention or lack of the capability in automatically identifying complex curves. This study is aimed to develop a new method using widely available GPS data that can automatically identify all types of horizontal curves and measure the corresponding curve radii, including the most challenging spiral curve. The simulation test using 385 synthetic horizontal curves shows that the proposed method can correctly identify 90.1% of the tested curves and can accurately classify 87.3% of the detected curves types. The field test shows that the proposed method can correctly identify all of the 25 tested curves and can accurately measure the corresponding radii. The results from an experimental test clearly demonstrate the accuracy and effectiveness of the proposed method. A case study on Interstate 285 demonstrates that proposed method is a promising method for transportation agencies to achieve reliable and efficient network-level analysis.