基于机器学习的无线通信信道估计方法
By constructing a three-dimensional propagation structure model and a sparse prior constraint matrix, combined with complex-valued phase-amplitude modulation and a deep network, the accuracy problem of channel estimation in low-pilot scenarios is solved, and the stability of channel recovery and system performance are improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NORTHEASTERN UNIV CHINA
- Filing Date
- 2026-05-12
- Publication Date
- 2026-07-17
AI Technical Summary
Existing channel estimation methods for millimeter-wave massive MIMO systems rely on high-density pilot grids or large-scale labeled sample sets, which leads to a decrease in channel estimation accuracy in low-pilot scenarios, resulting in erroneous sparse support set matching results and channel impulse response fitting bias.
A machine learning-based channel estimation method is constructed. By collecting pilot signals, antenna array geometric parameters, and multipath propagation parameters, a three-dimensional propagation structure model is built to generate a sparse support prior constraint matrix. The complex-valued phase-amplitude modulation matrix and complex domain deep expansion network are used to extract and estimate the sparse features of the channel. The beamforming weights are optimized by combining the multi-user utility function.
It significantly narrows the search range for sparse channel reconstruction, improves the stability and accuracy of channel recovery, and enhances system throughput and multi-user interference suppression performance.
Smart Images
Figure CN122179275B_ABST