Next generation (xG) wireless communications impose requirements on the data rate, spectral efficiency, and latency (among others) that are higher than those for today's systems by several orders of magnitude. Despite the many different perceptions of xG that exist, I envision that next generation wireless communications built on emergent wireless paradigms will involve unprecedented interactions between the MAC layer and the PHY layer. These transformative wireless paradigms will also necessarily be fueled by circuit design innovations, due to the extreme system requirements.
As an example, full-duplex (FD) wireless is an emergent wireless communication paradigm that can greatly improve wireless network performance but is also fraught with fundamental challenges in the design of integrated radios. FD operation involves simultaneous transmission and reception at the same frequency, resulting in the tremendous transmitter self-interference at the receiver input. This self-interference can be a billion times more powerful than the desired signal to be received.
In this talk, I will present a system design methodology that breaks the boundaries between the traditional functional domains of the radio through novel circuit design. First, I will introduce a noise-cancelling, self-interference-cancelling receiver that uses a co-design between a self-interference canceller and a noise-cancelling receiver to enable antenna interfaces with low transmitter-to-receiver isolation. Next, I will introduce frequency-domain equalization at radio frequencies (RF), a technique that leverages signal processing concepts traditionally implemented in the digital signal-processing block to achieve wideband self-interference cancellation (SIC). After that, I will talk about a joint optimization across the antenna, analog, and digital domains that achieves 85dB overall SIC, enabling a practical FD link. At the end of the talk, I will discuss my collaborative work with network theorists on power allocation algorithms and rate-gain characterization for OFDMA-based FD wireless networks. I will also talk about my collaboration with researcher from the field of MEMS on realizing tunable frequency-division-duplexing RF systems.