The research agreement between the Ford Motor Company and MIT is larger than the sum of its financial commitment.
The multimillion Ford-MIT Alliance that began in 1997 now has linked MIT and Ford researchers on more than 80 research projects as diverse as voice recognition and possible energy sources. Three of these areas illustrate the breadth of Ford's interests and influence: climate research, a new CAD modeling concept and voice recognition for automobiles.
Global challenges like developing sustainable energy require broad approaches such as the more than 30 projects at the Laboratory for Energy and the Environment and affiliated programs. The globalization of environmental and security-related concerns plus the likelihood of regulation has given the transportation industry a vested interest in both competitive advantage and sustainable strategies like clean energy.
In May, a team led by Professor David Marks, director of the Laboratory for Energy and the Environment, briefed Ford executives on MIT's climate research and its implications for business, and on a new, large-scale initiative on near-term energy strategies. That conversation continues this fall, focusing on topics such as emissions trading, carbon sequestration, the infrastructure for a hydrogen economy, and energy security issues.
"Through MIT's many research approaches, we are helping Ford see the broader implications of growing energy use, environmental impacts, and technological and regulatory advances. The research product itself allows us to advance the debate about pathways to a sustainable future and the methods that will identify those paths," said Marks.
In the Computer Aided Design (CAD) Lab, Professor David Wallace, the lab's co-director, has been working since 1996 on a new modeling concept and software that allows product designers to share, integrate and experiment with one another's computational models easily. While the web allows users to navigate and contribute to a network of static documents, the Distributed Object-based Modeling Environment (DOME) allows users to dive into networks of dynamic computational simulations.
"We wanted to test the idea on real problems," said Wallace. "What we've been able to do with Ford is show that this type of approach allows you to build the models much faster and more cheaply, to evolve and change them, and to reuse them flexibly. It removes the barriers that prevent integrated modeling from being practical."
Most people have talked to cars--often not very politely. At the Computer Science and Artificial Intelligence Laboratory (CSAIL), however, people are building the tools that will allow drivers to speak to their cars and get a useful response.
James R. Glass, head of CSAIL's Spoken Language Systems group, is working with Ford to explore the use of voice recognition technologies in realistic operating environments. The problem is challenging for a number of reasons specific to the automobile. "The car environment presents an acoustic challenge because it can be extremely noisy," said Glass. "Because the driver's hands and eyes are busy, we try to reduce the driver's cognitive load. In addition, since the driver's location is relatively fixed, it is a good environment to explore audio-visual processing methods."
MIT students have spent the past two summers at Ford road testing next-generation voice recognition systems called "In Vehicle Conversational Interfaces." One thing they're working on now is an address-recognition system that will allow users to say a destination aloud, rather than having to enter it via keystrokes into an onboard navigation system. Future applications may make it possible simply to ask for Russian folk music or the Boston weather rather than taking your eyes off the road to scan the radio display on the dashboard.
A version of this article appeared in MIT Tech Talk on November 10, 2004 (download PDF).