Multiuser MISO Broadcast Channels with Imperfect CSI: Discrete Signaling without SIC

Abstract

In this paper, we study the communication problem of multiuser multiple-input single-output (MISO) broadcast channels with imperfect channel state information (CSI) at the transmitter. Zero-forcing precoding based on the imperfect CSI is adopted so that the channel can be transformed into a Gaussian interference channel. We consider a practical setting where only discrete input signalings are employed and all the receivers adopt single-user treating-interference-as-noise (TIN) decoding, as opposed to rate-splitting and successive interference cancellation. Under this setting, we first use the deterministic model to approximate the original channel model and develop communication schemes to achieve the entire capacity region. By translating the results of the deterministic model back to the MISO model, we develop a systematic way to design discrete input signalings for the original problem. Our simulation results show that our scheme is capable of approaching the (outer bound of) capacity region of the Gaussian interference channel.