The Intelligent Transportation Systems
Vertically-Integrated Project (VIP) Team
Improve the transportation systems to, within, and from Georgia Tech
Distributed architectures, wireless networks, forecasting, dynamical systems, traffic modeling, simulation
Tech buses circulate on fixed routes according to a pre-set schedule. Such systems are notoriously subject to bus "bunching" or "platooning" in which small disruptions accumulate at an accelerating rate, so that headways become increasingly variable. Large headways tend to grow even larger and small headways tend to shrink still more. As a result, some riders must wait a long time for service, and then see one full bus followed by another that is nearly empty. The figure shows a real example of 3 Tech Trolleys platooning:
This project will build communications devices to be installed on buses. The devices will be used to coordinate buses so that headways tend to remain equal, so service is more predictable and no one has to wait too long for a bus.
Some tasks: (1) Read/interpret data streams from GPS devices already on buses. (2) Build model to predict time until next bus arrives to an arbitrary stop. The scheme will likely be similar to those used by advanced retailers to forecast customer demand based on weather, time of day, special events, etc. We have 5 years of data on which to develop/test forecasting schemes. (3) Design the architecture for a distributed control system, most likely based on Android phones, by which management can control the bus system. (4) Extend devices to manage fuel consumption, exhaust emissions, etc.
Professors John Bartholdi (ISyE), Russ Clark (CS), Jan Shi (ISyE)
PARTNERS AND SPONSORS:
David Williamson, GT Department of Parking and Transportation
ISyE, CS, CmpE, IC: Java programming, statistics and forecasting, business analytics, architecture of distributed software systems, user interface design.