Deep Learning Model-Based Channel Estimation for THz Band Massive MIMO with RF Impairments

Pulok Tarafder,Imtiaz Ahmed,Danda B. Rawat,Ramesh Annavajjala,Kumar Vijay Mishra

2024-09-25

Abstract:THz band enabled large scale massive MIMO (M-MIMO) is considered as a key enabler for the 6G technology, given its enormous bandwidth and for its low latency connectivity. In the large-scale M-MIMO configuration, enlarged array aperture and small wavelengths of THz results in an amalgamation of both far field and near field paths, which makes tasks such as channel estimation for THz M-MIMO highly challenging. Moreover, at the THz transceiver, radio frequency (RF) impairments such as phase noise (PN) of the analog devices also leads to degradation in channel estimation performance. Classical estimators as well as traditional deep learning (DL) based algorithms struggle to maintain their robustness when performing for large scale antenna arrays i.e., M-MIMO, and when RF impairments are considered for practical usage. To effectively address this issue, it is crucial to utilize a neural network (NN) that has the ability to study the behaviors of the channel and RF impairment correlations, such as a recurrent neural network (RNN). The RF impairments act as sequential noise data which is subsequently incorporated with the channel data, leading to choose a specific type of RNN known as bidirectional long short-term memory (BiLSTM) which is followed by gated recurrent units (GRU) to process the sequential data. Simulation results demonstrate that our proposed model outperforms other benchmark approaches at various signal-to-noise ratio (SNR) levels.

Information Theory,Signal Processing

What problem does this paper attempt to address?

This paper attempts to solve the channel estimation problem in terahertz (THz) band ultra - large - scale multiple - input multiple - output (UM - MIMO) systems, especially in the presence of radio - frequency (RF) impairments such as phase noise. Specifically, the paper focuses on the following key issues:

1. **Channel characteristics at high frequencies**: The THz band has the advantages of huge bandwidth and low - latency connection, but at the same time, it also faces the problem of high attenuation, which makes channel estimation very challenging.

2. **Far - field and near - field channel models**: In UM - MIMO systems, due to the increase in antenna array size and the decrease in wavelength, far - field and near - field paths exist simultaneously, which further increases the complexity of channel estimation.

3. **The impact of RF impairments**: In THz transceivers, RF impairments (such as phase noise) can lead to a decline in channel estimation performance. Traditional channel estimation methods perform poorly when dealing with large - scale antenna arrays and RF impairments in practical applications.

4. **Application of deep - learning methods**: In order to effectively solve the above problems, the paper proposes a deep - learning model based on bidirectional long - short - term memory network (BiLSTM) and gated recurrent unit (GRU) for channel estimation. These models can capture the correlation between the channel and RF impairments, thereby improving the accuracy of estimation.

The paper verifies through simulation results that the proposed model outperforms other benchmark methods at different signal - to - noise ratio (SNR) levels, especially in the cases of low SNR and high phase - noise variance.