Estimation of Location-Dependent Fundamental Diagram Based on Connected Vehicle Data and Sparse Modeling

Abstract

Fundamental diagram (FD) is a relationship that expresses the basic characteristics of traffic flow, such as traffic capacity, etc. Location-Dependent FD is the FD of each segment when a road is divided into segments with mutually different characteristics. Therefore, if the Location-Dependent FD can be properly estimated, the location and capacity of bottlenecks that cause traffic congestion can be determined. However, it is not easy to estimate the FD of a road by dividing it into appropriate segments based on actual traffic data. In this study, a method for estimating FD by dividing a road into segments with different characteristics was developed by applying sparse modeling using Connected Vehicle (CV) data (so-called " Probe Vehicle Data"), which have information on arbitrary points on the road, as input. The properties of the proposed method are validated by simulation-based and real-world-data-based experiment, and the estimation accuracy of the proposed method and future issues are shown.