Massachusetts Institute of Technology
Sign in | Create Account

Data-driven Traffic Modeling, Prediction, and Planning

03/15/2011 4:00 PM 3"270
Daniela Rus, MIT Electrical Engineering and Computer Science Associate Director of CSAIL;

Description: Some professors work primarily in labs, and others mainly at desks. Daniela Rus conducts her research on the bustling streets of Singapore, where she is helping to design a "future mobility project" whose goal is to "marry information technology with the transportation industry." This venture aims to improve urban passenger and freight transportation, addressing issues of gridlock and other traffic frustrations _ a giant step toward a more rational, sustainable travel system.

Rus's work, part of the Singapore"MIT Alliance for Research and Technology, involves piloting a "mobility on demand" transport network on a campus springing up to host foreign research groups. Part of this vision involves creating autonomous robot vehicles (golf carts with brains) that can sense their way with laser scanners and GPS to pick people up at one point, and drop them off at another. These robots must navigate their way through a densely populated human environment, avoiding collisions and timing departures and arrivals with precision. Fleets of lightweight vehicles figure in the dream of making cities greener by decreasing private vehicle use.

In addition, as part of this multi"phase project, Rus and her collaborators are collecting and analyzing vast amounts of data on current traffic patterns in Singapore, with the goal of "maximizing the experience in traffic." Accessing GPS information from Singapore's 16 thousand taxis, her group gathered data logged at one"minute intervals on taxi speed, location, and occupancy. "This rich data set for the entire country is a source of joy for me and my colleagues," says Rus. This information showed traffic spikes, and the most frequent origins and destinations _ "an interesting way of seeing what everyone in Singapore is up to."

With data gathered from roadbed detectors as well, Rus built up predictions about how taxi and general traffic moved from day to day, and then came up with algorithms for mapping "congestion aware routing," not just for single users, but for the entire city. Says Rus, "The system computes different paths according to different times of day," and can" help cars get from one point to another without getting stuck." Rus proudly relates that her algorithm works better "than the simple searches Google supports."

In the final phase of the project, says Rus, autonomous robot cars will be deployed carrying onboard navigation systems for sensing and predicting traffic patterns, picking up and dropping off riders at stations so as to minimize wait times, and generally moving where needed "in a congestion aware way."

About the Speaker(s): Daniela Rus is also the co"director of the CSAIL Center for Robotics, and an associate director of CSAIL. Her research interests include distributed robotics, mobile computing, and programmable matter. She has several research activities in environmental robotics. She is the recipient of an NSF Career award and an Alfred P. Sloan Foundation fellowship. She is a class of 2002 MacArthur Fellow, and a fellow of AAAI.

Previously, she was an assistant professor, associate professor, and professor in the Computer Science Department at Dartmouth. She holds a Ph.D. in computer science from Cornell University.

Host(s): School of Engineering, Transportation@MIT

Comments (0)

It looks like no one has posted a comment yet. You can be the first!

You need to log in, in order to post comments. If you don’t have an account yet, sign up now!

MIT World — special events and lectures

MIT World — special events and lectures

Category: Events | Updated 1 month ago

Created
December 19, 2011 12:11
Category
Tags
License
All Rights Reserved (What is this?)
Additional Files


Viewed
4518 times

More from MIT World — special events and lectures

What (if Anything) Should Be Done About Improving the System of Electing a President? (Part three)

What (if Anything) Should Be Done A...

Added 3 years ago | 01:00:00 | 3495 views

Worms, Life and Death: Cell Suicide in Development and Disease

Worms, Life and Death: Cell Suicide...

Added 3 years ago | 01:12:00 | 2749 views

Machine Learning of Language from Distributional Evidence

Machine Learning of Language from D...

Added 3 years ago | 00:40:50 | 2259 views

Bill Porter in Conversation with Howard Anderson

Bill Porter in Conversation with Ho...

Added 3 years ago | 00:53:46 | 3864 views

Energy: The Past Must Not Be Prologue

Energy: The Past Must Not Be Prologue

Added 3 years ago | 00:57:43 | 2907 views

The Story Is Mightier than the Data

The Story Is Mightier than the Data

Added 3 years ago | 00:27:14 | 1531 views