A new wireless chip can perform a feat that could prove quite useful for the next generation of wireless technology: transmitting and receiving signals on the same frequency, at the same time with the help of a single antenna. This approach instantly doubles the data capacity of existing technology though is not yet capable of power levels necessary to operate on traditional mobile networks.
Last year, Harish Krishnaswamy, an electrical engineer at Columbia University demonstrated the ability to transmit and receive signals on the same frequency using two antennas in a full duplex radio that he built. Now, Negar Reiskarimian, a PhD student under Krishnaswamy, has embedded this technology on a chip that could eventually be used in smartphones and tablets. This time, the transmitter and receiver share a single antenna.
Devices such as smartphones and tablets typically exchange signals over at least two antennas—one for the transmitter and one for the receiver. These signals are usually coordinated in one of two ways: time-division duplex, in which a transmitter and receiver take turns broadcasting on the same frequency, and frequency-division duplex, in which the transmitter and receiver broadcast on separate frequencies at the same time.
Compared to the traditional models, the new full duplex radio chip is more efficient. “You’re not wasting time or frequency,” Krishnaswamy says. Such conservation is especially important as smartphones use more data, and companies search for ways to free up frequencies. Krishnaswamy says his lab is already working with several chip manufacturers to refine the concept.
To achieve its efficiency, the new chip had to circumvent a longstanding principle called Lorentz Reciprocity, in which electromagnetic waves are thought to move along the same paths when traveling both backward and forward.
In the past, electrical engineers have bypassed reciprocity by designing elements called circulators built of magnetic materials. By applying a magnetic field, an engineer can disrupt reciprocity by permitting waves to flow only forward and not backward, which allows for the simultaneous transmission of two signals.
But circulators built in this manner are often expensive and too bulky to insert into a smartphone. Plus, the magnetic fields they use would disrupt other functions if ever placed within an electronic device. Instead, these types of circulators have most often been used for military purposes (in fact, Krishnaswamy‘s latest research was funded by DARPA).
To overcome that limitation, Reiskarimian implanted silicon transistors on the face of a CMOS chip in an arrangement that reroutes signals as they are captured by both the transmitter and the receiver in order to avoid interference. “You essentially want the signals to kind of circulate in a clockwise sense,” Krishnaswamy says.
It also helped to use an echo-cancelling receiver that the lab also pioneered. This receiver solves the classic problem that transmitted signals tend to "echo" back into a receiver when a full duplex radio is in operation. This echo can be billions of times stronger than any external signal that a receiver needs to process. The echo-cancelling receiver cuts through this noise by learning what the transmitted signal was and subtracting that out of the signal that the receiver processes.
He likens the final result to enabling two people to both talk and listen to one another at the same time. “You can double data capacity right down to the physical hardware,” he says. If integrated throughout an entire network, he thinks this technique could potentially reduce delays in data transmission.
For now, the new chip does not have a high enough broadcasting power level to connect to a mobile network. It’s in the neighborhood of 10 to 100 milliwatts, which is about where a Wi-Fi network typically starts, but mobile operates at higher levels. There are a few ways that Krishnaswamy is already planning to try to bolster the power level, such as by rearranging the components of the chip or choosing different hardware to build it.
Editor’s note: This post was corrected on April 15 to reflect power level in milliwatts in the final paragraph instead of megawatts.