Capacity Quest

When asked recently what kept him up at night, Dick Lynch, BAM CTO, answered capacity. Although the capacity issue may not be as severe in the United States as in countries such as Japan, it is always on carriers' minds, and solid methods to increase capacity and improve network efficiency are like gold.

University research labs can be a gold mine of cutting-edge information. A research team at Virginia Tech is working on improving location-system performance. Although the benefits to E-911 are obvious, commercial applications can help carriers improve network efficiency. A research team at Rutgers is looking at capacity as it relates to a user's location in a cell site. The projects head in two directions -- voice and data -- and lead to interesting applications.

Jeffrey Reed, associate director of the Mobile & Portable Radio Research Group (MPRG) at Virginia Tech, said that MPRG researchers have been looking at the methodology for classifying how good a location system is.

"The thrust of our research is ensuring the reliability of location," Reed said. "Given the physical limitations of base-station geometry, we've been looking at how good of a performance we could expect theoretically if everything were working perfectly."

LOCATION-BASED PRICINGLocation applications are reliant on the accuracy of the location system. Reed said researchers use hypothesis-testing procedures to tell them how well the location system is working for a particular area, as well as what sort of real errors are introduced by the imperfections of the location system. The procedures also tell how reliable that error figure is so the carrier can be assured that the location system is working correctly.

Reed said location-sensitive billing is one way to improve overall capacity because it drives people to use their phones in locations where they normally wouldn't use them. If you can get people in residential neighborhoods to pick up their wireless phones instead of their cordless phones, you're taking advantage of an area that is not being tapped, he said.

Location-sensitive billing helps increase the number of dollars a carrier gains per hertz because it helps overcome the problem of many cell sites idling while others are jammed, Reed said.

RELIABILITY ISSUESSo far the big problem with location-sensitive billing is its unreliability. If people are afraid of being charged at a high rate inside the border of the low-rate area, they are less likely to pick up their wireless phones at home.

"One aspect that we're looking at are location-finding algorithms," Reed said. "We're looking at coming up with angle-of-arrival determination techniques that are robust."

He said they're also evaluating locating CDMA signals using time difference of arrival (TDOA) and finding how you manage multiple-access interference and power control.

"It turns out that if you try to apply the classical TDOA algorithms to CDMA, your estimates get worse as you get closer to the base station, which is counter-intuitive," Reed said.

The base station controls the mobile unit power, and as the mobile unit approaches the base station, it tells the mobile to turn down the power because it doesn't need as much power if it's close. Even though the mobile may present a clean signal for the serving base station, the signal quality is degraded for the other base stations that need to pick up the signal to estimate the location.

Reed said that researchers have been considering measures to combat the problem at the base station such as smart antennas and relaxing the power control.

Another way that location technology can help the carriers is through better planning. For instance, if a carrier has a system that can monitor the call quality, and it notices the call quality is degraded, the carrier can geo-locate that user to help find holes in the coverage.

"It radically alters the way that carriers do network planning, and that will allow them to get by with fewer base stations," Reed said.

MANYTIME, MANYWHEREResearchers at Rutgers WIN-LAB are focusing on the location of users within a cell for data purposes.

Categorizing other-cell interference shows what kind of effect it has on capacity, said Narayan Mandayam, WINLAB research director for radio systems. Systems work well when users come closer to the base station.

"Particularly when you talk about data, the paradigm is that it should not be anytime, anywhere; it should be manytime, manywhere," Mandayam said. "This means that you can do data a lot of times and in a lot of places, but you don't have to do it all the time. For voice, anytime, anywhere makes sense. 'Here I am, and I would like to make a phone call.' But that's not true when you talk about data services."

To illustrate their theory, researchers at WINLAB are working on a project called Infostation, which is a new system concept to support data services. Data's much-higher rates call for more stringent requirements. Mandayam said typical users want to download their information and then move on.

If you want high data rates and if you want something that is generically non-voice, it makes sense to design systems where users operate only when they're close enough.

"If you're close enough, it makes 70% of your problems go away," Mandayam said.

The idea for the Infostation comes from the gas station concept, Mandayam said. When you need gas, you find a station, go there and fill up. Infostation uses the same basic concept, except users don't need to connect physically to a "station."

"If they're close enough, they will be able to take care of their data communication," Mandayam said.

THREE SCENARIOSThe researchers envision the Infostation in three scenarios. One could be a sit-down Infostation in an airport terminal. Road warriors waiting to board a plane could sit in the lounge and download their e-mail, a Web site and appointment information. Another example is to have the Infostation at the security X-ray point at the airport.

"Every time you board a plane, your laptop has to go through the security X-ray machine," Mandayam said. "Why not have some sort of wireless access port there so that when your laptop goes through the X-ray machine, it could download the information, or upload your e-mails?" The close proximity to the base station at the X-ray machine would allow users to download megabits of information easily.

"The third type of Infostation we envision is what I call a drive-through, similar to a drive-through ATM or fast-food restaurant," Mandayam said. When users drive through, they download information.

"What we are trying to motivate through this (research) is how folks look at data, and, depending on how users are located and distributed, you can handle it more cleanly," Mandayam said.

If you want data anytime, anywhere, it will mean designing the system inefficiently in order to meet the worst-case user's requirement -- somebody at the edge of the cell, Mandayam said. When you do this, you will be wasting resources.