Development of the Idaho Statewide Travel Demand Model Trip Matrices Using Cell Phone OD Data and Origin Destination Matrix Estimation

DATE: 
Friday, October 24, 2014, 12:00pm to 1:00pm PDT
SPEAKERS: 
Ben Stabler, Senior Supervising Planner, Parsons Brinckerhoff
COST: 
Free and open to the public
CREDIT: 
N/A

Watch video

View slides

Topic: Airsage cell phone data and its application in travel modeling
Summary: As part of the initial phase of development for the Idaho Statewide Travel Demand Model, Parsons Brinckerhoff developed a base year auto and truck trip matrix using AirSage cell phone OD data, a statewide network in Cube, traffic counts, and origin-destination matrix estimation (ODME) procedures. To begin, the 4000+ statewide zone system was aggregated into a 700 super zone system for collecting the cell phone OD data. Next, the cell phone data was collected for the month of September 2013 for the following market segments: Average weekday resident HBW, HBO, NHB, and visitor NHB trips. The cell phone trips were then disaggregated to zones using each zone’s share of super zone population and employment. These initial trip matrices were assigned to the daily statewide network using free flow travel time for route impedance and iteratively adjusted to minimize the difference between the estimated link volumes and traffic counts by user class.

This iterative trip matrix balancing procedure, also known as ODME, converged nicely by user class and facility type and produced reasonable flows. The resulting trip matrix trip length frequencies matched fairly well with the Census Journey to Work trip distances as well as the COMPASS (Boise area) MPO surveyed trip lengths. The goodness-of-fit between the synthesized trip lengths and the observed trip lengths was analyzed across multiple dimensions, including urban versus rural, by zone, and by trip purpose, and some differences were observed. The use of cell phone OD data, in conjunction with ODME, is an innovative and promising solution for developing statewide travel OD data with limited resources.