From left, Sheila Hemami, assistant professor of electrical engineering, Venugopal Veeravalli, assistant professor of electrical engineering, Stephen Wicker, associate professor of electrical engineering, and Zygmunt Haas, associate professor of electrical engineering, in the Wireless and Multimedia Laboratory in Rhodes Hall. Charles Harrington/University Photography
By Larry Bernard
You are on the road and your boss needs those figures immediately -- complete with pie charts and illustrations. She pages you, and you send them electronically. But wait -- you forgot to tell your husband that it's his turn to car pool the kids, so you reach for your personal communicator. By this time you are lost, so you flip on the global navigator.
Far-fetched? Hardly. The technology already is in use. But the wireless networks and resource management that make such tasks seem easy are extremely complex. How does the digital signal from the cellular phone know which receiver tower to use? How does the network know where to find your particular phone? How does e-mail know where to find you, no matter where you access it? How can digital data and images be compressed and transmitted reliably so that anyone can receive the quality image they want?
Researchers in Cornell's Wireless and Multimedia Laboratory, established last fall in the School of Electrical Engineering in the College of Engineering and located in Rhodes Hall, are working on these and other questions for the next generation of digital, multimedia communications for consumers, the military and industry. Their work, funded by the National Science Foundation, the U.S. departments of Defense and of Energy, AT&T, Lucent Technologies and others, is exactly the kind of basic research, the scientists say, that can benefit society in a myriad of ways.
"In the future, we expect millions of people will have wireless devices. It will be very cheap and will keep them connected to a network," said Zygmunt Haas, associate professor of electrical engineering. "Our lab is addressing the multimedia problem -- how do you send voice, data, video and images? Each has a particular challenge. The goal is to provide integrated, ubiquitous, multimedia access to information and services by mobile and stationary users; anywhere, anytime, with anyone and in any format."
An example of where such technology would be needed, he said, would be for a soldier in a battlefield who wants to see a map of targets or possible escape routes. "You want to allow reliable visual communication to the user in a hostile environment, like a battlefield," Haas said. "How you make that possible is the focus of the research."
Haas and his students also are investigating how to improve the quality of digital signals. Currently, a cellular phone signal is received at a single receiver. But outputs of multiple receivers could be combined so that the best possible signal is achieved, Haas said. This requires no new infrastructure, using existing receivers.
"If you're connected to more than one base-station, we can do intelligent processing of the signals," he said. Haas and his research group already have achieved improvements of a couple of orders of magnitude in bit error rates using this system, called "multiply-detected macrodiversity." They are investigating such technology for pagers as well.
Personal communicators, wireless data wide-area and local-area networks, wireless second-generation phone services, some cordless phones, new mobile satellite service and pagers -- all make use of digital signals. The reception of these signals can be improved through such research.
Sheila Hemami, assistant professor of electrical engineering, would like to pave the information superhighway with images and video, making visual communications as ubiquitous and reliable as voice communications. "We've been limited to speech communications since Bell invented the telephone," she said. "If I'm taking a vacation in Puerto Rico, I would like to see a video clip that shows how far the hotel is from the beach, that shows all the different rooms -- not just a brochure."
Her main research project is to develop image and video compression techniques that allow all users to access all data, regardless of the quality of their Internet connection, computing power, bandwidth or other limiting factors. The project, "Visual Communications for Heterogeneous Networks," is funded by the U.S. Department of Energy.
"For visual communications, there are real challenges," Hemami said. "Not only are there multiple bandwidths reaching customers, but the quality of service for each connection type may vary. Varying network capacities, differences in viewing devices and a broad spectrum of user needs require coding techniques that span a larger range of quality and bandwidth."
One way of solving the problem of different quality for different uses is to develop what Hemami calls "scalable compression." Currently, JPEG (for images) and MPEG (for video) are the standard compression formats. But these require separately coding the data for transmission over each communication bandwidth. With scalable compression, the image is available at all bandwidths, and the user can download just what is necessary for their particular use. Hemami and her colleagues are developing the algorithms required for that and for reconstructing images that had packets of data lost during transmission.
Similarly, "voice communications can tolerate a slight delay and some amount of transmission errors. But in data transmission, errors are a big deal. We work on such protocols that make it possible to provide the required quality of service for any particular traffic type," Haas added. "You have to have something at both ends of the link to make it reliable."
Another key element to reliable visual communications is network adaptivity. For example, when the network gets really busy, transmission should not just cut off. Instead Hemami is working on algorithms that compress video to adapt to the real-time bandwidth and packet handling of the network.
"By adopting scalability, reconstructability and network adaptivity, you get many benefits," Hemami said, "including an increased user base, extended collaborations, increased resource usage and ease of development for information servers."
Stephen Wicker, associate professor of electrical engineering, researches the technology that makes personal communications systems possible. "Future personal communications systems will carry voice, data and video, allow global roaming, and will run in multiple environments, such as home, office, rural or urban settings, all from a single terminal," Wicker said. "In trying to bring these applications together on one platform, you create some very interesting engineering problems."
Wicker, at Cornell since July 1996, focuses his research on error control, data link protocols and medium access. He spent nine years on the faculty at the Georgia Institute of Technology and has served as a telecommunications systems consultant for industry and governments in North America, Europe and Asia.
Venugopal Veeravalli, assistant professor of electrical engineering, who joined Cornell in August, is another of the founding members of the wireless multimedia lab. "There is not a clear-cut theoretical foundation yet in the field of wireless communications," he said, "and people in academia can provide that."
For example, look at telephone line modems. "Many people thought the information limit was 2,400 bits per second. That was the baud rate for a long time in industry. But people familiar with information theory knew the limit was much higher than that. Now many people have 28,800 baud rate modems on their desks. So it's equally important to establish theoretical limits for wireless systems," he said.
Veeravalli studies the new technology standard for cellular communications, called CDMA, or code division multiple access. Although the industry has accepted it and is preparing systems based on it, technical problems remain. "It's a much more complex technology, but in the end, it will result in systems that can support more users, with higher signal quality," Veeravalli said.
He also investigates the optimal allocation of resources for wireless systems, or how best to deal with limited bandwidth, the proliferation of cellular base stations and limited battery power of the mobile units.
"Wireless systems introduce a host of new research problems that don't exist in traditional point-to-point communication systems," Veeravalli said. "In order to solve these problems, we can draw on ideas from information theory, detection and estimation theory, signal processing and control theory, all of which are strong at Cornell."
Such research is not insignificant: Cellular antenna towers are established at a rate of more than 5,000 a year and can be 300 feet tall. Cell phone subscribers have soared from less than 500,000 a decade ago to more than 38 million today. Communities are becoming concerned and are enacting zoning laws to limit or camouflage these towers.
Other Cornell faculty members working on technologies for wireless multimedia systems include professors of electrical engineering Toby Berger, Thomas Parks and C. Richard Johnson Jr., and associate professors of electrical engineering Chris Heegard and Richard Compton.