Innovation increases wireless data rates

University engineers have invented a method of characterizing and managing the multiple channel paths generated when antenna arrays are used at a wireless transmitter and/or receiver.

The technology, developed by electrical and computer engineering assistant professor Akbar Sayeed, electrical and computer engineering professor Barry VanVeen and graduate student Eko Onggosanusi, effectively partitions the overall channel into a series of nonoverlapping subchannels.

VanVeen says it’s as if a large pipe were filled with a number of smaller pipes of varying sizes. Each smaller pipe has an individual capacity and characteristic that can be paired with the needs of the signal to be sent. Most importantly, the subchannels do not interfere with one another.

Consider the nine pairings possible in an array of three transmission antennas and three receiver antennas. The inherent distortion between any pair of antennas is different from every other pair. The distortion is often due to reflections off structures located between the receiver and transmitter, which cause different components of the signal to arrive at different times. Varying types and levels of distortion mean each channel will have varying transmission capabilities.

The UW–Madison team’s invention allows all of these factors to be considered in choosing over which subchannel(s) information should be sent. By weighing the options, information can be sent using the least transmission power to get the best error performance at the other end or to maximize data rate.

“In reality, it’s not that I’m going to use any one antenna on either side,” says VanVeen. “I use all three to form a new synthetic transmit antenna that optimally couples into the distortion of the channel by weighting thephysical antennas differently. At the physical antennas of the receiver, I then take a weighted combination of the signals to optimally capture the energy in the transmitted signal. So there are actually infinite combinations that one could achieve with different weightings. Because we’ve identified the optimal sets of weightings that result in noninterfering subchannels, we can send multiple messages in parallel and therefore increase the overall data rate while using minimum transmit power.”