This work offers a new concept for two-way wireless communication: rather than avoiding self-interference as in half-duplex, or combatting self-interference as in conventional full-duplex, nodes will re-use the received interfering radio-carrier waves to transfer information.
Funded by NSF CAREER #1620902
Team: B. Smida, S. Khaledian, S. Islam, J. Earl, S. Metz, and M. Tomehy
A new analog self-interference cancellation (SIC) technique for in-band full-duplex (IBFD) in single-antenna systems. We use an RF circulator to separate transmitted (Tx) and received (RX) signals. Instead of estimating the self-interference (SI) signal and subtracting it from the Rx signal, we use the inherent secondary SI signal at the circulator, reflected by the antenna, to cancel the primary SI signal leaked from the Tx port to the Rx port.
Funded by NSF CAREER #1620902
Team: S. Khaledian, F. Farzami, B. Smida and D. Erricolo
Single-antenna full-duplex communication technology has the potential to substantially increase spectral efficiency. However, limited propagation domain cancellation of single-antenna system results in a higher impact of receiver chain nonlinearities on the residual self-interference (SI) signal. In this paper, we offer a comprehensive SI model for single-antenna full-duplex systems based on direct-conversion transceiver structure considering nonlinearities of all the transceiver radio frequency (RF) components, in-phase/quadrature (IQ) imbalances, phase noise effect, and receiver noise figure. To validate our model, we also propose a more appropriate digital SI cancellation approach considering receiver chain RF and baseband nonlinearities. The proposed technique employs orthogonalization of the design matrix using QR decomposition to alleviate the estimation and cancellation error. Finally, through circuit-level waveform simulation, the performance of the digital cancellation strategy is investigated, which achieves 20 dB more cancellation compared to existing methods.
Funded by NSF CAREER #1620902
Team: M. A. Islam and B. Smida
Incorporating delay requires re-thinking strategies commonly used in the multi-user setting. Currently, many downlink systems orthogonalize the users as in for example Time Division Multiple Access (TDMA), or conceptually similar Frequency or Code Division multiple access (FDMA, CDMA) schemes. These schemes are simple and practically relevant, but are not in general capacity achieving. For the single antenna, non-fading Gaussian BC, it is known that the Non-orthogonal Multiple Access (NOMA) superposition coding scheme is capacity achieving. In our preliminary work, we asked whether these commonly used schemes, well understood in a capacity / first order sense, perform well in the non-asymptotic. We considered a two-user AWGN BC with full channel state information (CSI) at all nodes. In general, the longer the blocklength over which messages are coded, the lower the probability of error. As such, it is intuitive that a scheme such as the following concatenate-andcode protocol (CCP) and superposition coding may outperform TDMA.
Funded by NSF CISE # 1900911
Team: D. Tuninetti, B. Smida, D. Devroye, and H. Seferoglu.