المفاهيم الأساسية
The author proposes a learning-based active sensing framework to optimize transmit precoder and receive combiner matrices in massive MIMO systems, outperforming traditional methods by leveraging neural networks for efficient channel estimation and beamforming.
الملخص
The content discusses the development of an active sensing framework for optimizing transmit precoder and receive combiner matrices in massive MIMO systems. The proposed approach utilizes neural networks to enhance channel estimation, outperforming traditional methods in various scenarios such as Rayleigh fading channels and hybrid MIMO systems.
Key points include:
- Introduction of active sensing for optimizing transmit precoder and receive combiner matrices in massive MIMO systems.
- Proposal of a learning-based framework using neural networks to improve channel estimation efficiency.
- Comparison with traditional methods like LMMSE+SVD, power iteration method, and summed power method.
- Performance evaluation under Rayleigh fading and ray-tracing based channel models.
- Extension of the active sensing framework to hybrid MIMO systems.
The results demonstrate the superiority of the proposed active sensing approach over conventional methods across different SNRs and channel models.
الإحصائيات
The system operates at a carrier frequency of 3.5 GHz.
Transmit power is set at 20 dBm with noise power at -114 dBm during the pilot phase.
اقتباسات
"The proposed active sensing method outperforms other benchmarks significantly across different SNRs."
"The neural network generalizes well for different numbers of ping-pong rounds."