King County Metro

Developing a Modeling System Designed to Predict Metro’s Bus Arrival Time

This student team worked to develop statistical models for travel time/arrival prediction on one or multiple bus routes for King County Metro (KCM) and also to implement the developed models on the subject routes. On-time-performance is the key performance indicator for a transit route. A reliable and efficient, low-latency prediction model will help to achieve better predictability for travel times along transit routes, which will provide much better customer satisfaction for KCM transit users. A predictable travel time/arrival prediction model will also help customers to confidently plan their transit trips that may connect to other modes of transportation (such as rail, water or micro-transit services). In addition, better implementation of the model will help the controllers in Transit Control Center to plan reduced idling times, which in essence may help to reduce air pollutants in the greater Seattle area and improve environmental sustainability. Outcomes this student team worked towards included: 1) a statistical module (includes mathematical formulas, parameter values, and Python programs), 2) implementation guidelines and test results of the model on the subject routes. The deliverables the student team worked toward were meant to include all the design and development documentation, test cases, and an implementation manual.