EECS energy logo News & Events EECSenergy Classes EECSenergy Faculty MITEI Links

EECSenergy home Energy generation & storage > Thermoelectrics > Solar Photovoltaics > Thermophotovoltaics > Turbine projects > Micro Energy Harvesters > Storage Energy Efficiency > Lighting> Energy use: architectural > Energy use: electronic devices > Electricity Transmission > Electric Machines > Transportation Energy Reports & Resources EECSenergy Teaching


CarTel | Autonomous passenger vehicles

Hari Balakrishnan
Sam Madden

Traffic congestion wastes billions of gallons of fuel every year in the United States alone: a 2005 study from the Texas Transportation Institute estimated that more than five times as much gas is wasted in traffic congestion today compared to twenty years ago in urban areas in the US.

We are developing a distributed mobile sensor network and telematics system called CarTel that serves as a platform for a range of transportation and civil infrastructure applications, including those that can allow users to avoid congested roads. CarTel uses a network of sensor-equipped cars (e.g., taxi fleets, private cars, etc.) that collect information about road conditions. This information is delivered using opportunistic wireless connections (vehicular WiFi) to servers that process and analyze this information. CarTel is currently deployed on a 27-node limo fleet in the Boston/Cambridge area.

An important CarTel application is MyRoute, developed in collaboration with PhD student Sejoon Lim and EECS professors Daniela Rus and David Gifford (both members of the Computer Science and Artificial Intelligence Lab, CSAIL, at MIT). MyRoute uses historical traffic delay statistics obtained from the CarTel cars and real-time information to provide users with good routes that take congestion, delay, and desired deadlines into account. For example, using a class of practical stochastic shortest paths algorithms and models of observed delays on different road segments, MyRoute can tell a user what route to take and when to leave in order to reach a desitination by a given deadline at any time of the day. Our group of PhD students and post-doctoral researchers (Bret Hull, Jakob Eriksson, lewis Girod, and Ryan Newton) is also using CarTel to develop a road surface monitoring application called P2 (the Pothole Patrol) using vibration sensors on cars. For more information:

fig. 13. Screenshot of the CarTel MyRoute application interface.

Autonomous passenger vehicles
Seth Teller

We are engaged in a long-term research program to develop safe, autonomous (i.e., self-driving) passenger cars. The widespread availability of such cars would bring a number of societal benefits. First, safety: with sensors that exceed human abilities, and computers that beat human reaction times, cars can become much safer, avoiding most accidents -- eliminating some of the millions of automotive and pedestrian fatalities that occur annually worldwide.

Second, energy: by driving more smoothly in traffic, and eventually coordinating with other cars to traverse merges and intersections without stopping, cars can become much more fuel-efficient. Third, productivity: with humans no longer forced to use most of their brain capacity simply to survive their daily commutes, they will be free to work (or rest, read, etc.) while in their cars.

Images of the Darpa Urban Grand Challenge Team MIT in preparation for the Fall, 2007 urban challenge event held in California.

Prof. Teller working on autonomous vehicle for Darpa Urban Grand Challenge 2007
Seth Teller riding--no hands--in the autonomous DGC vehicle, spring 07
DGC vehicle under preparation at MIT, spring 2007
EECS Home Page | Site Map | Search | About this page | Comments and inquiries welcome
MIT Dept. of Electrical Engineering and Computer Science MIT Dept. of Electrical Engineering and Computer Science MIT homepage nano EECS EECSenergy bio EECS