Fixed presentation and saving data

lostinafro · Mar 23, 2023 · e8b1d91 · e8b1d91
1 parent b6c79ba
commit e8b1d91
Showing 1 changed file with 15 additions and 15 deletions.
diff --git a/cb_bsw_vs_minimum_snr.py b/cb_bsw_vs_minimum_snr.py
@@ -3,13 +3,11 @@
 import matplotlib.pyplot as plt
 
 import scenario.common as cmn
-from environment import RisProtocolEnv, command_parser, NOISE_POWER_dBm, TX_POW_dBm, T, OUTPUT_DIR, TAU
+from environment import RisProtocolEnv, command_parser, NOISE_POWER_dBm, TX_POW_dBm, T, OUTPUT_DIR, TAU, BANDWIDTH, NUM_PILOTS
 import os
 
 from matplotlib import rc
 
-# rc('font', **{'family': 'sans serif', 'serif': ['Computer Modern']})
-# rc('text', usetex=True)
 
 ##################################################
 # Parameters
@@ -24,9 +22,6 @@
 # Get noise power
 noise_power = cmn.dbm2watt(NOISE_POWER_dBm)
 
-# Define number of pilots
-num_pilots = 1
-
 # For grid mesh
 num_users = int(1e4)
 
@@ -87,7 +82,7 @@
     h_eq_cb_fixed = (g_rb.conj()[np.newaxis, np.newaxis, :] * codebook_fixed[np.newaxis, :, :] * h_ur[:, np.newaxis, :]).sum(axis=-1)
 
     # Generate some noise
-    var = noise_power / num_pilots
+    var = noise_power / NUM_PILOTS
     bsw_noise_ = np.sqrt(var) * noise_
 
     # Compute the SNR of each user when using CB scheme
@@ -123,7 +118,7 @@
             index = np.argmax(mask)
 
             # Store results
-            se_flex[ms, uu] = np.log2(1 + minimum_snr)
+            se_flex[ms, uu] = np.log2(1 + minimum_snr) if snr_cb_flexi[uu, index] > minimum_snr else 0.
             n_configurations_flex[ms, uu] = index + 1
 
         # Go through all users (FIXED)
@@ -135,12 +130,15 @@
             if np.sum(mask) == 0:
                 continue
 
+            # Get the index of the highest snr satisfying the constraint
+            index = np.argmax(snr_cb_hat_fixed[uu] * mask)
+
             # Store results
-            se_fixed[ms, uu] = np.log2(1 + minimum_snr)
+            se_fixed[ms, uu] = np.log2(1 + minimum_snr) if snr_cb_fixed[uu, index] > minimum_snr else 0.
 
 
     ## Varying the frame duration
-    tau_plot = [8.75, 10., 20.]
+    tau_plot = [7.5, 15, 30]
 
     opt_kpi_flexi = np.zeros((len(TAU), len(setups)))
     opt_kpi_fixed = np.zeros((len(TAU), len(setups)))
@@ -156,7 +154,7 @@
             datafilename = 'data/opt_ce_vs_tau_' + setup + '.npz'
 
             # Load data
-            rate = np.squeeze(np.load(datafilename)['rate'][TAU == total_tau])
+            rate = np.squeeze(np.load(datafilename)['rate'][TAU == total_tau]) * BANDWIDTH / 1e6
 
             if total_tau in tau_plot:
                 axes[ss].plot(minimum_snr_range_dB, np.mean(rate) * np.ones_like(minimum_snr_range), linewidth=1.5, color='black', label='OPT-CE')
@@ -170,6 +168,7 @@
             else:
                 tau_setup = 3 * T
 
+            # Flexible structure
             # Pre-log term
             tau_alg = (2 * n_configurations_flex - 1) * T
             prelog_term = 1 - (tau_setup + tau_setup + tau_alg) / total_tau
@@ -178,18 +177,19 @@
             rate_cb_bsw = prelog_term * se_flex
             avg_rate_cb_bsw_flexi = np.mean(rate_cb_bsw, axis=-1)
             if total_tau in tau_plot:
-                axes[ss].plot(minimum_snr_range_dB, avg_rate_cb_bsw_flexi, linewidth=1.5, linestyle=':', label='CB-BSW: Flexible', color='red')
+                axes[ss].plot(minimum_snr_range_dB, avg_rate_cb_bsw_flexi * BANDWIDTH / 1e6, linewidth=1.5, linestyle=':', label='CB-BSW: Flexible', color='red')
 
             opt_kpi_flexi[tt, ss] = minimum_snr_range_dB[np.argmax(avg_rate_cb_bsw_flexi)]
 
+            # Fixed structure
             # Pre-log term
             tau_alg = num_configs * T
-            prelog_term = 1 - (tau_setup + tau_setup + tau_alg) / total_tau
+            prelog_term = np.maximum(1 - (tau_setup + tau_setup + tau_alg) / total_tau, 0)
 
             rate_cb_bsw = prelog_term * se_fixed
             avg_rate_cb_bsw_fixed = np.mean(rate_cb_bsw, axis=-1)
             if total_tau in tau_plot:
-                axes[ss].plot(minimum_snr_range_dB, avg_rate_cb_bsw_fixed, linewidth=1.5, linestyle='--', label='CB-BSW: Fixed', color='blue')
+                axes[ss].plot(minimum_snr_range_dB, avg_rate_cb_bsw_fixed * BANDWIDTH / 1e6, linewidth=1.5, linestyle='--', label='CB-BSW: Fixed', color='blue')
 
             opt_kpi_fixed[tt, ss] = minimum_snr_range_dB[np.argmax(avg_rate_cb_bsw_fixed)]
 
@@ -206,7 +206,7 @@
         # Plot for specific values
         if total_tau in tau_plot:
             cmn.printplot(fig, axes, render=render, filename=f'rate_vs_gamma0_{total_tau:.2f}', dirname=OUTPUT_DIR,
-                          labels=[r'$\gamma_0$ [dB]', r'$R$ [bit/Hz/s]', r'$R$ [bit/Hz/s]'], orientation='vertical')
+                          labels=[r'$\gamma_0$ [dB]', r'$R$ [Mbit/s]', r'$R$ [Mbit/s]'], orientation='vertical')
 
     # Save optimal threshold values
     opt_kpi_fixed = np.mean(opt_kpi_fixed, axis=-1)