Interference Suppression using Impropriety and Cyclostationarity of Signals in Multi-User MIMO Systems

Abstract

In this thesis, we consider the interference suppression using the impro- priety and cyclostationarity of signals in multi-user multiple-input multiple- output (MU-MIMO) communication systems. First, we consider the minimum mean-squared error (MMSE) equalization for uplink MU-MIMO systems. In particular, the transmitters (TXs) adopt discrete-Fourier transform (DFT)-spread orthogonal frequency-division multi- plexing (OFDM) to send π/2-PAM or square-QAM symbols, possibly with the frequency-domain spectrum shaping (FDSS). It is well known that, in such cases with improper-complex symbols, widely-linear (WL) equalizers can im- prove performance over linear equalizers, but at the cost of sometimes dras- tically increased computational complexity. To overcome this, we exploit the spectral correlation in the received signal induced by the impropriety from π/2- PAM symbols and the cyclostationarity from the FDSS. The key differences from conventional high-complexity structures are to represent all the symbols as equivalent PAM symbols that are subsequently π/2-modulated, and to factor the effective channel into the π/2-modulator, the DFT-spreader, and a partial channel defined in the frequency domain. These representation and factoring enable us to design a low-complexity structure of the WL-MMSE equalizer that performs the signal processing mostly in the frequency domain and involves only sparse and structured matrices. Numerical results show that the proposed structure achieves significantly lower complexity than the conventional struc- tures and that it greatly outperforms the linear MMSE equalizer with a similar complexity. Second, we design a receiver (RX) for an MU-MIMO underlay system in the presence of a strong cyclostationary legacy signal. Each desired signal is chosen to be a single-carrier signal that exhibits cyclostationarity with the same cycle period as the legacy signal. If a conventional RX is used that does not take into account the effect of the analog-to-digital converter (ADC), then the strong interference from the legacy TX may cause quantization noise to dominate the desired signals from the underlay TXs and severely degrade the reception performance. To address this issue together with the cyclostationar- ity of the signals, we propose an ADC-aware RX that adopts an analog-digital hybrid combiner structure. The analog combiner first suppresses the interfer- ence power to a negligible level through the projection onto an approximate null space of the interference correlation matrix. Then, it enhances the total power of desired signals by applying the principal component analysis on the signal correlation matrix. After the output of the analog combiner is sampled and quantized at the ADCs, the digital combiner generates the linear MMSE estimates of the symbols transmitted by the underlay TXs. During the sym- bol estimation, it fully exploits the periodic spectral correlation induced by the cyclostationarity of signals. Simulation results show that our proposed RX ef- fectively maximizes signal-to-interference-plus-noise ratio and, under certain conditions, approaches the performance achievable in the absence of the legacy TX.