The iPhone may seem to be today’s ultimate wireless communications device, but it isn’t to Giuseppe Caire.
“It’s very primitive, very slow, because the underlying network is primitive. It’s really based on second generation wireless technology,” he says. “But it does have a great user interface.”
Caire, a professor in the Ming Hsieh Department of Electrical Engineering, is working to design systems for the next generation of wireless cellular systems.
Giuseppe Caire --Photo by Van Urfalian
The first generation of cell phones were analog devices. The second generation phones were digital and the wireless systems featured standards like GSM (Global System for Mobile communications), the basis for the iPhone and CDMA (Code Division Multiple Access).
“We’re in the third generation now, which is the transition from voice to data,” says Caire. “But the third generation is not really maintaining its promises in terms of data rates, coverage and services. Data rates are too low and the handsets are not there yet. We have to solve that problem for the fourth generation.”
The problem, says Caire, is in the bottleneck in the downlink. When your cell phone is connected to the Internet, the amount of information transmitted on the uplink (from the phone to the network) is relatively small compared to the amount of information sent on the downlink (from the network to the phone). You may send email or maybe credit card data to a server, but that is mostly text.
“But on the downlink you need much higher capacity to download multimedia, such as songs, movies, or maybe a Google map,” he says. Streaming multimedia content and downloading other large files has become the killer app for wireless systems.
In 2003, when he was still at the Eurecom Institute in Sophia-Antipolis, France, Caire and Shlomo Shamai of Technion, the Israel Institute of Technology, published a seminal paper aimed at the downlink problem. The paper, published in IEEE Transactions on Information Theory, outlined the problem and explored a solution.
“The solution is to use more antennas to send information down. To take advantage of the many antennae, you have to share time and frequency, but you are separate in space.”
Multiple antennae transmitting multiple beams of data at the same time to multiple wireless device users is straightforward and simple in theory, but in reality, the problem is a very difficult.
Timeline evolution of the main wireless technologies projected from 2006 to 2010 and beyond. For larger image, click here.
Modern cellular networks contain many layers and as the amount of information – all in the form of digital bits, or ones and zeros – transmitted in the network increases, the schemes used to code and modulate the information grow increasingly complex. The multiple antennae, which increase the number of channels, add to the complexity. And the reality is that things like buildings and trees get in the way, scatter signals and create errors.
Caire says the downlink problem “belongs to a class of problems called ‘non-degraded broadcast channels’ in information theory, for which there is no known general capacity formula.”
In the late 1940s, when Claude Shannon laid out a mathematical and theoretical basis for digital telecommunications, he noted that the capacity of any communications channel could be well defined. And as long as a transmission of digital information does not exceed the channel’s capacity, there exist ways to encode information such that, after decoding, there will be essentially no errors.
“Shannon, however, doesn’t tell you how to achieve capacity,” chuckles Caire. “We’re designing modern coding schemes, such as turbo codes and low-density parity-check codes that are able to perform close to channel capacity with very small bit error rates.”
Fourth-generation wireless systems will utilize orthogonal frequency division modulation (OFDM), in which data are broken into a large number of parallel channels. OFDM’s main advantage is that it is able to cope with severe channel conditions such as fading. Furthermore, orthogonal access in time-frequency (FDMA/TDMA) inside each cell is optimal in fading channels, even from the information theory viewpoint.
Caire is confident that engineers are well on the way to solving the downlink problem, and that robust networks will someday provide a rich, seamless connection to the Internet for wireless devices. The goal is to design a system where capacity can grow without bounds with the number of base station antennas, while keeping the user handsets very simple, and essentially independent on the number of base station antennas. If an operator needs more capacity in some cell, it has just to add more antennas, in a totally transparent way with respect to the users.
Caire came to USC in 2005 as a full professor, attracted by the Viterbi School’s long history in communications. He and Alan Willner, also a professor in the Ming Hsieh Department, are co-directors of the Viterbi School’s Communications Sciences Institute.
Caire, a native of Turin Italy, received a B.S. and Ph.D. degrees in electrical engineering from Politecnico di Torino in 1990 and 1994, respectively. In between, he earned an M.S. from Princeton University. He worked for the European Space Agency in the Netherlands, was an assistant professor of telecommunications at the Politecnico di Torino and an associate professor at the University of Parma before going to the Eurecom Institute in France.
“We only had telecommunications engineers there,” he says. “Here there is a big campus with many different scholars. I like this environment.”
He says that he and his wife, Isabella, who is an orthopedic surgeon at the Keck School of Medicine of USC, have grown to love Southern California, despite the fact that his last academic post was located on the French Riviera.
“The weather’s similar, traffic is better than it is in France, and gas is cheaper…as for the cheese and wine, well ....” he laughs.