From LOS to VMT, VHT and beyond through data fusion: application to integrated corridor management.
PI: Alexandre Bayen, UC Berkeley
Abstract: This work supports an implementation of some of the objectives of SB-743 for operations and planning. First, the project will investigate the current gaps in the traffic data sources commonly used in the computation of standard metrics such as Level of Service (LOS). It will investigate the data requirements necessary for the computation of metrics suggested by SB-743, in particular Vehicle Miles Traveled (VMT) and Vehicle Hours Traveled (VHT). Second, it will propose two algorithms to integrate new sources of data in the computations of these metrics. The first algorithm aims to improve the computations of VHT in arterial networks by integrating available GPS data from existing infrastructure. The second algorithm will then investigate the use of cell tower phone data, i.e., call data records (CDR), to solve the route inference problem, through compressed sensing and convex optimization based least square optimization. The project will suggest how to use the enhanced information to produce more reliable VMT and VHT estimates. The contributions will be illustrated by leveraging the Connected Corridors I-210 Pilot in the San Gabriel Valley, Los Angeles, California. We will show that the resulting computations present significant improvement over current approaches, and can support the aspects of SB-743 that require this data.
Quarterly Progress Reports
Click on the links below to download any of the quarterly progress reports already finished.