UNIVERSITY PARK, Pa. — The development of millimeter wave communication technologies could improve the speed, latency and reliability of wireless communications significantly, according to Penn State researchers. Jing Yang, assistant professor of electrical engineering, received a four-year, $800,000 National Science Foundation grant to develop machine learning and large-scale optimization-based schemes to advance the technology.
“Millimeter wave spectrum offers 200 times more bandwidth than microwave frequencies, offering substantial promise toward the future of wireless networks,” Yang said, noting that the technology has not been widely adopted due to issues involving how the data propagates along its communication path.
The millimeter wave spectrum operates on a higher frequency than microwave communication and provides a larger bandwidth as well, which allows it to process more data. When operating on this frequency band, some disruptive applications can be accommodated, such as collaborative autonomous driving and augmented and virtual reality.
For example, raw collaborative sensor data sharing in autonomous driving can enable a number of safety and cloud-controlled driving applications. Similarly, tracking of farm equipment with high-definition sensor data from crops in precision agriculture, as well as the tracking robotic equipment in the mining industry can benefit people, processes and environmental conditions because of the high bandwidth requirement from the millimeter wave spectrum, according to Yang.
She said that one of the challenges of this research has been the sensitive nature of the millimeter wave spectrum’s high frequency. Cellphones, for example, operate on a lower frequency, and blocking them with a hand or using them while walking will not impact their wireless connection. The millimeter wave spectrum, however, is susceptible to blockage from obstacles, including users’ own bodies near the device.
“In a manufacturing facility, this is not too much of a problem because the objects are stationary,” Yang said. “The challenge appears when the devices are moving.”
Yang and two co-principal investigators — Mahanth Gowda, assistant professor of electrical engineering and computer science, and Mehrdad Mahdavi, Dorothy Quiggle Career Development Assistant Professor of Computer Science and Engineering — are addressing this issue by developing approaches to make the communication as fast, stable and continuously reliable as possible.
“We have to estimate which channel of communication is good — and do this very quickly,” Yang said. “Channel estimation has to be done at a very fast rate. By leveraging machine learning and optimization, we’re finding computationally efficient ways to estimate the channel quickly and enable seamless handovers and connectivity as devices move.”