Title:
Exploiting power-amplifier non-linearity in relay networks for over-the-air computation to enhance confidentiality
Abstract:
We consider wireless networks with relays that amplify and forward signals. Most works on wireless communications systems assume that power amplifiers used are linear, or treat their non-linearity as an impediment that needs to be mitigated. In this work, we exploit the non-linearity of power amplifiers in relay networks. In particular, a relay network with non-linear power amplifiers has a structure that has many similarities to an artificial neural network (ANN), where the amplifier non-linearities act like activation functions of artificial neurons. Using this approach, techniques used to train deep neural networks can be used to optimize weights and biases applied at the relays to perform over-the-air, non-linear computations. Using this approach, we propose a novel method to enhance confidentiality in relay networks where a single transmitter transmits data to multiple receivers by multiplexing their data into the same symbol. We show that the non-linear relays can be used to minimize the information leakage to unintended receivers in such a scenario, highlighting the potential for exploiting power amplifier non-linearity in wireless relay networks to perform over-the-air computation.
Speaker:
Siddhartan Govindasamy
Professor of Engineering, Boston College
Bio:
Siddhartan Govindasamy is a Professor of Engineering at Boston College. He received the bachelor’s, master’s, and Ph.D. degrees from Massachusetts Institute of Technology (MIT), in 1999, 2000, and 2008, respectively. From 2000 to 2003, he was a DSP Engineer and later a Senior DSP Engineer at Aware Inc., where he worked on developing broadband modem technology. From 2008 to 2020, he was an Assistant Professor and later an Associate Professor of electrical and computer engineering at the Olin College of Engineering, Needham, MA, USA where he conducted research on large multiple-input–multiple-output (MIMO) systems and optical wireless communications. He joined the Department of Engineering, Boston College as a founding faculty member of their Human-Centered Engineering program in Fall 2020. He is the co-author of the textbook Adaptive Wireless Communications: MIMO Channels and Networks (Cambridge University Press, 2013) and his research interests include radio frequency and optical wireless communications, signal processing and frugal innovation.
Host: Professor Alex Wyglinski