SmarTrAC: A Smartphone Solution for Travel and Activity Capturing

Principal Investigator:

Yingling Fan, Assistant Professor, Humphrey School of Public Affairs


Project Summary:

Although the use of mobile phones to collect travel behavior data has increased rapidly, the existing smartphone applications have not been designed to allow much interaction between sensor data and user input data. By bringing together automatic sensing, surveying, and statistical machine learning seamlessly, SmarTrAC offers the best of both worlds -- addressing the limitations of both passive sensing and recall surveys. It yields travel data of a breadth and depth not available by using either smartphone sensory data or user input data alone. It allows sensor data to interact with user input data so that the two data sources can calibrate with each other. Such interaction between sensor data and user input data allowed by SmarTrAC forms feedback loops to perfect the sensor data processing procedure and the process of capturing user input in SmarTrAC over time. SmarTrAC has passed both laboratory tests and field tests among real-world smartphone users. Tests confirmed that SmarTrAC has a reasonable battery consumption rate (with room for improvement), a moderate data storage/transmission requirement, a high accuracy in classifying episodes as activities vs. trip, a high accuracy in identifying travel modes for trips, and as a medium-high accuracy in classifying activity type for activities.


Project Details:

  • Start date: 06/2013
  • Project Status: Completed
  • Research Area: Planning and Economy
  • Topics: Planning