[Solved] Learning Beam Codebooks in mmWave/THz MIMO System  

Hi Burhan,

I am glad that you reproduced the figure and thank you for pointing out the potential typo on the website.

Regarding your question for generating Fig. 7 (a), you simply need to run the simulation multiple times (i.e., what you have done to generate Fig. 7 (c) corresponds to only one point in Fig. 7 (a)). Specifically, each time before you run the codes, you should change the value of options['num_NNs'] in the main.py script to be the codebook size you want. As a result, each time you will get a codebook with its size specified by options['num_NNs'].

After acquiring these learned codebooks, you can test their performance on your user channel dataset. You can refer to Eq. (8) in the paper for the calculation. Along with DFT codebook and the EGC beamforming method, you can finally calculate the average beamforming gain achieved by each one of them and get Fig. 7 (a).

Let me know if this is not clear or you have any other questions.

Best,
Yu

Hi Burhan,

Thanks for your reply.

Yes, it will take a long time if you want to get codebooks of different sizes. I am really sorry for causing the inconvenience to you. To shorten the execution time (which of course also heavily depends on the computation platform you have), for instance, I was simply opening multiple sessions (i.e., whatever program you use to run the python scripts) to accelerate the simulation by that time (i.e., different processes execute the learning of different codebook sizes). Another way to make the time shorter could be by changing options['num_loop'] and train_opt['num_iter'] which basically control the number of iterations of the simulation. Of course, if you make these numbers very small then you will expect a certain level of degradation in the final performance, which is always the tradeoff in these stochastic algorithms. Or similarly, if you observe that the achieved gain is already good enough, then you can just stop the program without waiting till the end. Regarding the calculation of the performance, kindly find the following very easy example code:

M = 64; % M: number of antenna
N = 16; % N: codebook size
U = 100; % U: number of users

Theta = randn(M, N); % The learned phases
W = (1/sqrt(M))*exp(1j*Theta); % Beam codebook (output of the simulation)

H = randn(M, U) + 1j * randn(M, U); % User channel dataset

average_gain = mean(max(abs(W'*H).^2)); % Eq. (8) in the paper

You simply need to replace W with either your learned codebook or DFT codebook or EGC codebook (just to make sure that it has a shape of the number of antennas by the number of beams). And replace H with your user channel dataset (similarly, make sure that it has a shape of the number of antennas by the number of users). As I mentioned last time, Eq. (8) in the paper is exactly what I used to calculate such average beamforming gain performance if you want to know more details.

Good luck,
Yu