chenpengseu / DNN-DANM

Simulation Tips

1. Train the network

1. main_generate_data_train.m generate the data for traning. The system parameters are saved in the file param.mat, the training data is saved in the file data_set.mat, and the valid data is saved in the file valid_set.mat.
2. train.py train the network, the loss results are saved in the file loss.npz, and the trained network is saved in the file net.pkl。

2. Test the performance

2.1 RMSE-SNR performance

1. main_generate_data_SNR.m generate the data to test the RMSE-SNR performance, and the data is saved in the file data_SNR.mat.
2. main_SNR.py is used to get the denoised signal, which is saved in the file denoise_signal.mat.
3. main_SNR_RMSE.m: show the performance of different algorithms. The performance of different methods is obtained as
   • get_SNR_RMSE_performance('fft'): fft method
   • get_SNR_RMSE_performance('omp'): omp method
   • get_SNR_RMSE_performance('fft-denoise'): fft method with the proposed network
   • get_SNR_RMSE_performance('omp-denosie'): omp method with the proposed network

2.2 RMSE-SNR performance (change RIS amp errors)

1. main_generate_data_SNR_amp.m generate the data to test the RMSE-SNR performance, and the data is saved in the file data_SNR.mat.
2. main_SNR.py is used to get the denoised signal, which is saved in the file denoise_signal.mat.
3. main_SNR_RMSE_amp.m: show the performance of different algorithms. The performance of different methods is obtained as
   • get_SNR_RMSE_performance('fft'): fft method
   • get_SNR_RMSE_performance('omp'): omp method
   • get_SNR_RMSE_performance('fft-denoise'): fft method with the proposed network
   • get_SNR_RMSE_performance('omp-denosie'): omp method with the proposed network

2.3 RMSE-SNR performance (change RIS phase errors)

1. main_generate_data_SNR_phase.m generate the data to test the RMSE-SNR performance, and the data is saved in the file data_SNR.mat.
2. main_SNR.py is used to get the denoised signal, which is saved in the file denoise_signal.mat.
3. main_SNR_RMSE_phase.m: show the performance of different algorithms. The performance of different methods is obtained as
   • get_SNR_RMSE_performance('fft'): fft method
   • get_SNR_RMSE_performance('omp'): omp method
   • get_SNR_RMSE_performance('fft-denoise'): fft method with the proposed network
   • get_SNR_RMSE_performance('omp-denosie'): omp method with the proposed network

2.3 RMSE-SNR performance (change mutual coupling)

1. main_generate_data_SNR_mc.m generate the data to test the RMSE-SNR performance, and the data is saved in the file data_SNR.mat.
2. main_SNR.py is used to get the denoised signal, which is saved in the file denoise_signal.mat.
3. main_SNR_RMSE_mc.m: show the performance of different algorithms. The performance of different methods is obtained as
   • get_SNR_RMSE_performance('fft'): fft method
   • get_SNR_RMSE_performance('omp'): omp method
   • get_SNR_RMSE_performance('fft-denoise'): fft method with the proposed network
   • get_SNR_RMSE_performance('omp-denosie'): omp method with the proposed network