Bayesian Joint Channel-and-Data Estimation for Quantized OFDM Over Doubly Selective Channels

Abstract

This paper investigates the design of orthogonal frequency division multiplexing (OFDM) receiver, which operates over time-frequency doubly selective (DS) channels and uses a low-resolution analog-to-digital converter (ADC) to quantize the received signal. The time-varying multipaths in DS channel and the nonlinear quantization distortion caused by low-resolution ADC, destroy the orthogonality among OFDM subcarriers, incurring severe inter-carrier interference (ICI). Therefore, joint channel-and-data estimation is preferred, since in this way one can effectively tackle the ICI and make the best of the quantized signal in each OFDM symbol. In this paper, we elaborately develop the joint estimation scheme by casting the problem into the Bayesian inference framework. The proposed scheme not only takes into account of the nonlinear quantization in a probabilistic manner, but also exploits the channel sparsity enabled by the basis expansion model (BEM) approximation to DS channels. Furthermore, to benchmark the proposed scheme, we derive the asymptotic theoretical bounds in the large-system regime. Simulation results demonstrate that the proposed scheme can achieve the performance close to the theoretical bounds, thereby verifying the effectiveness of the proposed scheme.