PI: Sean Qian, Heinz College
Co-PI: Afsaneh Doryab, Robotics Institute
Automated vehicle location (AVL) systems and automatic passenger counters (APCs) have been widely adopted in the transit industry to collect high-resolution passenger and vehicle information. Though they hold great potentials to improve system performance, little effort has been made to fully utilize the data to understand travelers’ behavior and optimize transit operations. The essential idea of this research is to fully utilize the big data in public transit to provide travelers fine-grained customizable information regarding transit service performance (efficiency, reliability and quality), and to facilitate decision making for transit agencies.
We propose efficient algorithms to measure transit service performance using AVL and APC data in various perspectives, including bus travel time ratio to car travel time, excessive waiting time, crowding, bunching, park-and-ride information, and incidents. A full set of customizable performance measures in different levels of temporal and spatial granularity will be provided. Bunching, the most critical issue in bus transit service, can be detected by examining excessive waiting time and crowding factors of sequential bus trips. In addition, we explore the most promising performance measures related to the bus ridership and bunching, and optimize bus schedules to improve service performance and increase ridership. We propose to develop a Transit Service Performance Information and Optimization system (TranSEPIO), which embeds those algorithms of performance measures and data analytics for decision making of both travelers and agencies. TranSEPIO is a generic platform that can be deployed to meet the specific need of any transit agencies.
Our research and initial development of TranSEPIO will fully incorporate the rich data sets provided from Port Authority of Allegheny County (PACC). A prototype version of TranSEPIO for PACC is the expected outcome of this research.
Co-PI: Afsaneh Doryab, Robotics Institute
Automated vehicle location (AVL) systems and automatic passenger counters (APCs) have been widely adopted in the transit industry to collect high-resolution passenger and vehicle information. Though they hold great potentials to improve system performance, little effort has been made to fully utilize the data to understand travelers’ behavior and optimize transit operations. The essential idea of this research is to fully utilize the big data in public transit to provide travelers fine-grained customizable information regarding transit service performance (efficiency, reliability and quality), and to facilitate decision making for transit agencies.
We propose efficient algorithms to measure transit service performance using AVL and APC data in various perspectives, including bus travel time ratio to car travel time, excessive waiting time, crowding, bunching, park-and-ride information, and incidents. A full set of customizable performance measures in different levels of temporal and spatial granularity will be provided. Bunching, the most critical issue in bus transit service, can be detected by examining excessive waiting time and crowding factors of sequential bus trips. In addition, we explore the most promising performance measures related to the bus ridership and bunching, and optimize bus schedules to improve service performance and increase ridership. We propose to develop a Transit Service Performance Information and Optimization system (TranSEPIO), which embeds those algorithms of performance measures and data analytics for decision making of both travelers and agencies. TranSEPIO is a generic platform that can be deployed to meet the specific need of any transit agencies.
Our research and initial development of TranSEPIO will fully incorporate the rich data sets provided from Port Authority of Allegheny County (PACC). A prototype version of TranSEPIO for PACC is the expected outcome of this research.