Full-Duplex Millimeter Wave MIMO Channel Estimation: A Neural Network Approach

TL;DR

This work tackles full-duplex mmWave MIMO channel estimation by introducing NN-based estimators for both SI and UE channels, aiming to reduce pilot overhead while coping with interference and nonideal ADC effects. It demonstrates that shallow CNNs can achieve NMSE comparable to or better than traditional LS/MMSE estimators under various pilot dimensions and hardware distortions, and it shows that NN-based RX-TX channel mapping effectively predicts downlink channels in separate antenna configurations. The study also investigates distribution shift with DeepMIMO data, revealing that simpler networks generalize better in such conditions, and provides a practical complexity comparison highlighting the computational advantages of NN-based and LS methods over MMSE. Overall, the results suggest NN-based channel estimation as a promising approach to enable efficient, high-performance full-duplex mmWave systems with reduced pilot overhead and hardware constraints.

Abstract

Millimeter wave (mmWave) multiple-input-multi-output (MIMO) is now a reality with great potential for further improvement. We study full-duplex transmissions as an effective way to improve mmWave MIMO systems. Compared to half-duplex systems, full-duplex transmissions may offer higher data rates and lower latency. However, full-duplex transmission is hindered by self-interference (SI) at the receive antennas, and SI channel estimation becomes a crucial step to make the full-duplex systems feasible. In this paper, we address the problem of channel estimation in full-duplex mmWave MIMO systems using neural networks (NNs). Our approach involves sharing pilot resources between user equipments (UEs) and transmit antennas at the base station (BS), aiming to reduce the pilot overhead in full-duplex systems and to achieve a comparable level to that of a half-duplex system. Additionally, in the case of separate antenna configurations in a full-duplex BS, providing channel estimates of transmit antenna (TX) arrays to the downlink UEs poses another challenge, as the TX arrays are not capable of receiving pilot signals. To address this, we employ an NN to map the channel from the downlink UEs to the receive antenna (RX) arrays to the channel from the TX arrays to the downlink UEs. We further elaborate on how NNs perform the estimation with different architectures, (e.g., different numbers of hidden layers), the introduction of non-linear distortion (e.g., with a 1-bit analog-to-digital converter (ADC)), and different channel conditions (e.g., low-correlated and high-correlated channels). Our work provides novel insights into NN-based channel estimators.

Figures (18)

• Figure 1: System model with a full-duplex BS serving $K_{\text{u}}$ single-antenna UEs in the uplink and $K_{\text{d}}$ single-antenna UEs in the downlink channel.
• Figure 2: Different pilot dimensions. $T$ is the frame dimension during one coherence block and $\tau$ is the pilot dimension for estimating SI and UE channels.
• Figure 3: The CNN architecture employed for SI and UE channel estimation. we apply $n$ ($n = 0, 1, 2, 10$) convolutional hidden layer(s) each with a $3\times 3$ window size and 64 convolutional kernels.
• Figure 4: Phase shift between the RX and TX arrays in the separate antenna configuration in a full-duplex BS.
• Figure 5: NMSESI vs SNRSI with a varying number of hidden layers.