Metro Transit selects Cambridge Systematics for real-time bus predictions pilot
Cambridge Systematics (CS) was selected by Metro Transit in Minneapolis to participate in a pilot program to evaluate approaches and technologies to improve the agency’s bus arrival time predictions based on current information about vehicle location, traffic conditions, weather, and other relevant factors.
The CS team is building a real-time predictions solution for Metro Transit using the open-source software, TheTransitClock. The CS team includes Sean Óg Crudden and Simon Berrebi, who are independent leaders in maintaining and enhancing TheTransitClock.
“We’re excited to use open-source software to deliver a customized solution that combines historical and real-time information with Metro Transit-specific operating rules and practices,” says Sarah Anderson, Vice President for Cambridge Systematics. “Transportation agencies benefit from open-source software because it is cost-effective, there are no licensing fees or vendor lock-in, and the functionality is constantly being improved upon. It allows for meaningful partnerships and needed updates, which are not guaranteed when agencies purchase off-the-shelf software.”
To improve the accuracy of predicted arrival times, the CS team’s solution employs a Kalman Filter, an adaptive algorithm that recognizes unusual conditions, like abnormal traffic congestion and weather events, and can draw on the most relevant historical and real-time data to improve the quality of its arrival time predictions by more accurately estimating vehicle travel times and dwell times (periods when the bus is at a stop, loading and unloading passengers).
“The accuracy of real-time predictions is crucial to better serving the needs and expectations of transit passengers, and the Metro Transit service area is subject to highly variable travel conditions based on severe winter weather along with major events in the Twin Cities” says Paul Sorensen, Principal of Cambridge Systematics. “That’s why our team is using the most advanced techniques for improving arrival predictions during both typical and adverse travel conditions. When the weather is 10 degrees below zero, you don’t want to go wait at the bus stop until you know the bus is about to arrive.”
CS is conducting the TheTransitClock pilot through October, while working iteratively with Metro to assess accuracy and continuously improve performance. Following the pilot program, Metro expects to select a real-time solution to implement this winter.