utils.py

from tensorflow.python.keras import *
import tensorflow as tf
import scipy.io as sio

# ---------------------
#  Global Parameters
# ---------------------
Nt = 64  # the number of antennas
P = 1   # the normalized transmit power


# ---------------------
#  Functions
# ---------------------

# transfer the phase to complex-valued analog beamformer
def trans_Vrf(temp):
    v_real = tf.cos(temp)
    v_imag = tf.sin(temp)
    vrf = tf.cast(tf.complex(v_real, v_imag), tf.complex64)
    return vrf


# For the simplification of implementation based on Keras, we use a lambda layer to compute the rate
# Thus, the output of the model is actually the loss.
def Rate_func(temp):
    h, v, SNR_input = temp
    hv = backend.batch_dot(
        tf.cast(h, tf.complex64), tf.transpose(a=v, perm=[1, 0]))
    rate = tf.math.log(tf.cast(1 + SNR_input / Nt * tf.pow(tf.abs(hv), 2), tf.float32)) / tf.math.log(2.0)
    return -rate


# load the saved .mat files generated by Matlab.
def mat_load(path):
    print('loading data...')
    # load the perfect csi
    h = sio.loadmat(path + '/pcsi.mat')['pcsi']
    # load the estimated csi
    h_est = sio.loadmat(path + '/ecsi.mat')['ecsi']
    print('loading complete')
    print('The shape of CSI is: ', h_est.shape)
    return h, h_est