A NEURAL NETWORK APPROACH TO BEAM SELECTION AND POWER OPTIMIZATION IN MM WAVE MASSIVE MIMO
Massive MIMO (multiple-input multiple-output) systems are important for millimeter-wave (mm-Wave) communication. These systems connect base stations (BS) with user equipment (UE) for faster and more efficient data transfer. However, there are challenges in achieving this performance. Beam selection...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Institute of Mechanics of Continua and Mathematical Sciences
2025-04-01
|
| Series: | Journal of Mechanics of Continua and Mathematical Sciences |
| Subjects: | |
| Online Access: | https://jmcms.s3.amazonaws.com/wp-content/uploads/2025/04/24061001/jmcms-2504040-A-Neural-Network-Approach-AK-2.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Massive MIMO (multiple-input multiple-output) systems are important for
millimeter-wave (mm-Wave) communication. These systems connect base stations (BS) with user equipment (UE) for faster and more efficient data transfer. However, there are challenges in achieving this performance. Beam selection and power control between the BS and UE must be done accurately. This is difficult because the channel state information (CSI) is not always available. Without proper information, selecting the best beam or power level results in inefficiency. To solve this, we propose a deep learning-based framework. This system uses a beam-steering technique to estimate signal strength. The goal is to choose the best beam for the user and the right power level for data transmission. The framework minimizes power usage when the user gets the required data rate. It works even when CSI is unknown. Missing data is another common problem in these systems. Some beams or signal information not be available . To handle this, we use a machine learning model called LSTM (Long Short-Term Memory). LSTM processes time-based data to predict the missing values. Using this,
the system still selects the best beam. To verify this approach, we used a dataset called Deep MIMO. This dataset is based on realistic simulations of wireless channels. Tests showed that the proposed framework works better than existing methods. The system performed well even without full CSI and handled missing data effectively. The paper offers a solution to improve communication in mm-Wave systems. It uses advanced deep-learning models to address challenges like beam selection, power control, and missing data. The proposed method is efficient and accurate, outperforming other strategies. |
|---|---|
| ISSN: | 0973-8975 2454-7190 |