University of Minnesota
University Relations
http://www.umn.edu/urelate
612-624-6868
myU OneStop


Go to unit's home.

Home | DTI | 2008–09 funded proposals | Henry Liu, Chen-Fu Liao, John Levin, Janet Hopper, Gary Nyberg, Kevin Sederstrom

Initiatives in Digital Technology: 2008–09 Funded Proposals

Henry Liu, Chen-Fu Liao, John Levin, Janet Hopper, Gary Nyberg, Kevin Sederstrom

Mining Bus Automatic Vehicle Location (AVL) and Automatic Passenger Count (APC) Database for Intelligent Transit Applications

An Automatic Vehicle Location (AVL) system was previously installed on every bus in the Metro Transit fleet to monitor vehicle location and track its schedule adherence. In addition to improving the efficiency of transit operations, AVL data can also be used to provide real-time transit travel time information or schedule planning. An Automatic Passenger Count (APC) system was installed on approximately 10% of current Metro Transit buses to collect ridership data for route planning, schedule frequency analysis and quality of service evaluation. Significant amounts of AVL and APC data are collected by Metro Transit on a daily basis. Examining the AVL or APC data individually provides a good understanding of operational bottlenecks and service needs. However, it does not offer a system wide perspective on how well the transit network is running. These data can be further utilized and integrated with arterial traffic data for intelligent transit applications, such as real-time transit operation analysis. In addition to these systems, each bus is also equipped with a registering farebox and a smartcard (GoTo Card) reader that represent two more independent sources of ridership data.

Collected data are like separate puzzle pieces. We are proposing a database model and data mining/fusing algorithm to pull these puzzle pieces together for systematic evaluation and analysis. The database model, by generating performance measures, will allow a transit manager or operator to identify problematic locations in the network and deploy appropriate strategies to maintain the overall system performance. We also would like to utilize the proposed model to improve the data quality though data mining and fusion from additional data sources, when they become available. The proposed model can potentially incorporate the raw or pre-processed datasets from Metro Transit with arterial traffic data along a bus route to investigate and better understand the causes of bus delay and other operational challenges. The proposed database model can later be integrated with the arterial travel time estimation model developed by the PI for various intelligent transit applications.