You are here

Real-Time Predictive Transit

August 24, 2016

CEE Professor Haris Koutsopoulos is using real-time factors such as weather and events to create predictive models for public transit.

Source: News @ Northeastern

Haris N. Kout­sopoulos believes that our ability to pre­dict the future has the poten­tial to enable var­ious inno­va­tions in public transit. The short-​​term future, that is.

Is a subway sta­tion or train about to get over­crowded during rush hour? Is an impending storm about to wreak havoc on the system, or is a bus sta­tion about to be flooded with fans leaving a base­ball game or con­cert? Should I, as a com­muter, con­sider taking an alter­na­tive route home or leaving work a bit later?

These are the types of public transit ques­tions Kout­sopoulos, pro­fessor of civil and envi­ron­mental engi­neering at North­eastern, is focused on addressing by way of real-​​time pre­dic­tive analysis. In one project involving Trans­port for London—the body that over­sees London’s transit system, including the London Under­ground, which is one of the world’s busiest metro systems—he and Peyman Nour­salehi, one of his doc­toral stu­dents, are devel­oping real-​​time pre­dic­tive models that fore­cast subway transit activity 15 or 30 min­utes into the future. Such models, he explains, are based upon analyses of large swaths of auto­mated fare-​​collection data that can reveal past travel pat­terns as well as real-​​time fac­tors such as weather, events, and even­tu­ally, even social media chatter. Kout­sopoulos’ team has devel­oped an ini­tial pro­to­type of a pre­dic­tive model as well as an accom­pa­nying visu­al­iza­tion tool.

A lot of the work with this data has basi­cally looked at what hap­pened yes­terday, where pas­sen­gers enter and where they exit the system,” Kout­sopoulos says. “Now what we’re thinking about is what can we learn from the travel pat­terns in all this data and using what we learn from the past to make short-​​term pre­dic­tions about the future. It’s about being proac­tive, not reac­tive. For example, some­times in those sys­tems, if a sta­tion gets too crowded, the gates are closed and pas­sen­gers aren’t allowed to come in until the crowd sub­sides. This is reac­tive. But if you can pre­dict that demand will increase in the near future, maybe you can take action ear­lier and pre­vent the problem from becoming bigger later on.”

Now what we’re thinking about is what can we learn from the travel pat­terns in all this data and using what we learn from the past to make short-​​term pre­dic­tions about the future. It’s about being proac­tive, not reac­tive.
— Pro­fessor Haris Koutsopoulos

The goal of Kout­sopoulos’ research is to develop tools that help transit oper­a­tors opti­mally manage their sys­tems and com­muters make informed trip deci­sions. For example, he says, pre­dic­tive ana­lytics could be used for real-​​time gate man­age­ment to pre­vent over­crowding at major sta­tions or feed into a transit app that sends real-​​time pre­dic­tive infor­ma­tion to riders.

The London project is part of Kout­sopoulos’ work in the MIT-​​NEU Transit Lab, a col­lab­o­ra­tion between North­eastern and the Mass­a­chu­setts Insti­tute of Tech­nology. In addi­tion to Trans­port for London, Kout­sopoulos and his doc­toral stu­dents are studying how pas­sen­gers on the MTR in Hong Kong use the under­ground transit system with the goal of devel­oping strate­gies to alle­viate con­ges­tion in the main parts of the net­work. These strate­gies include helping oper­a­tors improve crowd man­age­ment and incen­tivizing riders to alter their travel patterns.

Kout­sopoulos and his stu­dents are also pur­suing sep­a­rate, but related, research focused on observing how dif­ferent riders use a transit system and then infer­ring under­lying traits about these travel behav­iors. By clus­tering riders into dif­ferent groups based on these behav­iors, he explains, you can better under­stand rid­er­ship patterns—and there­fore improve your pre­dic­tive models.

Haris Koutsopoulos, professor of civil and environmental engineering, stands at Northeastern station on the MBTA Green Line. Photo by Matthew Modoono/Northeastern University

Haris Kout­sopoulos, pro­fessor of civil and envi­ron­mental engi­neering, stands at North­eastern sta­tion on the MBTA Green Line. Photo by Matthew Modoono/​Northeastern University

A traffic sim­u­la­tion pioneer

His work rep­re­sents an example of the next-​​generation of how we think about transportation—using Big Data to make informed deci­sions about how, when, and where people move. Kout­sopoulos describes his research as being focused pri­marily on intel­li­gent trans­porta­tion sys­tems. As he puts it, “The idea is to use tech­nology to improve how well we use the capacity that is actu­ally avail­able in the system to min­i­mize inefficiencies.”

Kout­sopoulos has been a pio­neer in the field of traffic sim­u­la­tion mod­eling for more than 20 years, and ear­lier this year he was hon­ored with the Traffic Sim­u­la­tion Life­time Achieve­ment Award by the Trans­porta­tion Research Board.

Prior to joining North­eastern in 2014, Kout­sopoulos founded the iMo­bility lab at the KTH Royal Insti­tute of Tech­nology in Stock­holm, where he used real-​​time GPS data from taxis to develop traffic man­age­ment and pre­dic­tion tools for local authorities.

As a glob­ally renowned researcher in this field, Kout­sopoulos recently co-​​hosted an inter­na­tional con­fer­ence at North­eastern, called TransitData2016, where scholars and trans­porta­tion offi­cials world­wide con­vened to dis­cuss new research and advance­ments cen­tered on using data from auto­mated sources to improve plan­ning and oper­a­tions at public transit sys­tems. He noted that three trends he observed were more dis­cus­sion about data fusion—using data from a variety of sources to make trans­porta­tion man­age­ment, evi­dence based decisions—increased use of visu­al­iza­tion tools by researchers and prac­ti­tioners to better com­mu­ni­cate how trans­porta­tion sys­tems are func­tioning, and increased interest by agen­cies in data ware­houses and open data.

We view all the work we’re doing with transit agen­cies as building blocks that they can use to improve ser­vice, be more respon­sive, com­mu­ni­cate better, plan their sys­tems better, and overall be more com­pet­i­tive in pro­viding mobility options,” Kout­sopoulos says.