A multi-user detection method and system based on BGIG-SBL
By introducing the Bernoulli Gaussian Inverse Gamma-Sparse Bayesian Learning (BGIG-SBL) method, the parameter update is updated using variational Bayesian inference, which solves the problem of the failure to effectively utilize the sparsity of sparse signals in the existing technology and improves the performance of multi-user detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- FUZHOU UNIV
- Filing Date
- 2023-07-14
- Publication Date
- 2026-07-24
AI Technical Summary
Existing multi-user detection algorithms based on sparse Bayesian learning fail to effectively utilize the sparsity of active user indices corresponding to sparse signals on Gaussian inverse gamma prior models, resulting in insufficient detection performance.
We employ the Bernoulli Gaussian Inverse Gamma-Sparse Bayesian Learning (BGIG-SBL) method, which learns the sparsity of the active user index by introducing binary vectors from the Bernoulli prior model and uses variational Bayesian inference to update parameters to improve the sparsity of the reconstructed signal.
It improves the performance of multi-user detection, especially under different signal-to-noise ratio conditions, the BER performance is about 1 dB to 35% better than traditional methods.
Smart Images

Figure CN116886243B_ABST
Abstract
Citation Information
Patent Citations
Hybrid Gaussian spectrum sensing method based on block sparsity characteristics
CN108880713A
Authorization-free joint active user and data detection method
CN116032317A