Machine Learning Based Hybrid Precoding for MmWave MIMO-OFDM with Dynamic Subarray

Yiwei Sun,Zhen Gao,Hua Wang,Di Wu

DOI: https://doi.org/10.48550/arXiv.1809.03378

2018-09-10

Abstract:Hybrid precoding design can be challenging for broadband millimeter-wave (mmWave) massive MIMO due to the frequency-flat analog precoder in radio frequency (RF). Prior broadband hybrid precoding work usually focuses on fully-connected array (FCA), while seldom considers the energy-efficient partially-connected subarray (PCS) including the fixed subarray (FS) and dynamic subarray (DS). Against this background, this paper proposes a machine learning based broadband hybrid precoding for mmWave massive MIMO with DS. Specifically, we first propose an optimal hybrid precoder based on principal component analysis (PCA) for the FS, whereby the frequency-flat RF precoder for each subarray is extracted from the principle component of the optimal frequency-selective precoders for fully-digital MIMO. Moreover, we extend the PCA-based hybrid precoding to DS, where a shared agglomerative hierarchical clustering (AHC) algorithm developed from machine learning is proposed to group the DS for improved spectral efficiency (SE). Finally, we investigate the energy efficiency (EE) of the proposed scheme for both passive and active antennas. Simulations have confirmed that the proposed scheme outperforms conventional schemes in both SE and EE.

Information Theory

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the hybrid precoding design problem of Partially - Connected Subarray (PCS) in broadband millimeter - wave (mmWave) massive Multiple - Input Multiple - Output (MIMO) systems. Specifically, the paper focuses on how to optimize the hybrid precoding scheme through machine - learning methods under the Dynamic Subarray (DS) structure to improve the Spectral Efficiency (SE) and Energy Efficiency (EE) of the system.

### Background and Motivation
In the fifth - generation mobile communication technology, millimeter - wave has become one of the key technologies because of its high data rate and large bandwidth. However, the millimeter - wave channel has a serious path - loss problem. Although the traditional all - digital precoding method can alleviate this problem, its power consumption is extremely high. Therefore, hybrid precoding has been proposed as a solution that can both achieve large array gain and reduce hardware cost and power consumption. Most of the existing broadband hybrid precoding research focuses on Fully - Connected Array (FCA), while there is less research on Partially - Connected Subarray (such as Fixed Subarray, FS and Dynamic Subarray, DS). Therefore, the hybrid precoding design for PCS in broadband channels is a topic worthy of exploration.

### Paper Contributions
1. **PCA - Based Hybrid Precoding for Fixed Subarray (FS)**:
- A fixed subarray hybrid precoding scheme based on Principal Component Analysis (PCA) is proposed. By extracting the frequency - flat radio - frequency precoder of each subarray from the optimal frequency - selective precoder of all - digital MIMO, the abstraction of low - dimensional signal space is achieved.
- The optimality of the proposed PCA - based hybrid precoder design is theoretically proven, and its performance is verified by simulation.

2. **Shared - AHC Algorithm for Dynamic Subarray (DS)**:
- A Shared Agglomerative Hierarchical Clustering (Shared - AHC) algorithm is proposed for antenna grouping of dynamic subarrays. By considering the cross - correlation between antennas, this algorithm effectively avoids the problem of local optimal solutions, thereby improving the spectral efficiency.
- By implementing the Shared - AHC algorithm, the spectral efficiency performance of PCS is further enhanced, making it adapt to the spatial characteristics of frequency - selective channels.

3. **Energy Efficiency Analysis**:
- The energy efficiency performance of the proposed scheme under passive antenna and active antenna architectures is analyzed. The results show that the dynamic subarray has significant advantages in energy efficiency, especially in practical applications that need to balance power consumption and spectral efficiency.

### Conclusion
The paper proposes a machine - learning - based hybrid precoding scheme for dynamic subarrays in broadband millimeter - wave MIMO systems. By introducing the PCA - based precoder design and the Shared - AHC antenna grouping algorithm, this scheme is superior to existing methods in terms of spectral efficiency and energy efficiency. In particular, the dynamic subarray performs well in practical applications and can maintain high spectral efficiency while reducing power consumption.