A Decoupling Property in Low-Resolution MIMO-OFDM Systems and its Applications

We consider a multi-user multi-antenna system in which a base-station (BS) equipped with numerous antennas communicates simultaneously with multiple users over a frequency selective wideband channel using orthogonal frequency division multiplexing (OFDM). The BS employs low resolution Analog-to-Digital-Converters (ADCs) in the Uplink and low resolution Digital-to-Analog-Converters (DACs) in the downlink. Given infinite resolution (ideal) quantization it is well known that OFDM enables transforming the time-domain channel afflicted by inter-symbol interference into a bank of parallel channels in the frequency domain, such that the noise across those channels is uncorrelated. We refer to the latter property as the decoupling property. We show that under a mild assumption, this decoupling property can also hold for arbitrary quantization resolution and for both uplink and downlink. Our result implies that the optimal linear minimum mean squared error (MMSE) equalizer for a system employing coarse quantization with N subcarriers can be obtained with an O(N) complexity instead of the hitherto assumed O(N
³
) prohibitive complexity. We also show that this complexity can be further reduced via interpolation schemes devised using theory of Laurent polynomial matrices.