Matlab code (Matlab2014a) of ML-BiGAMP
This repository provides the implementation of multi-layer bilinear generalized approximate message passing (ML-BiGAMP), which can be found in arXiv: https://arxiv.org/pdf/2007.00436.pdf
This repository includes 7 experiments (please see its branches).
The 1st experiment (EXP1) verified the consistency of ML-BiGAMP and its SE in 2-layer model.
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP1)
The 2nd experiment (EXP2) provided SNR versus BER of pilot-only, JCD, and Perfect-CSI scheme.
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP2)
The 3rd experiment (EXP3) shown the influence of Kp/Kd on the JCD scheme.
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP3)
The 4th experiment (EXP4) compared ML-BiGAMP with ML-Mat-VAMP in 3-layer model.
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP4)
The 5th experiment (EXP5) gave the application of ML-BiGAMP in compressive matrix completion
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP5)
The 6th experiment (EXP6) shown the consistency of ML-BiGAMP and its SE in 1-layer model with Bernoulli Gaussian prior.
(https://github.com/QiuyunZou/ML-BiGAMP/tree/MLBiGAMP-EXP6)
The 7th experiment (EXP7) shown the consistency of ML-BiGAMP and its SE in 1-layer model with QPSK symbol.