diff options
Diffstat (limited to 'dpd/src/Model.py')
| -rw-r--r-- | dpd/src/Model.py | 355 |
1 files changed, 355 insertions, 0 deletions
diff --git a/dpd/src/Model.py b/dpd/src/Model.py new file mode 100644 index 0000000..827027a --- /dev/null +++ b/dpd/src/Model.py @@ -0,0 +1,355 @@ +# -*- coding: utf-8 -*- +# +# DPD Calculation Engine, model implementation. +# +# http://www.opendigitalradio.org +# Licence: The MIT License, see notice at the end of this file + +import datetime +import os +import logging + +logging_path = os.path.dirname(logging.getLoggerClass().root.handlers[0].baseFilename) + +import numpy as np +import matplotlib.pyplot as plt +from sklearn import linear_model + +class Model: + """Calculates new coefficients using the measurement and the old + coefficients""" + + def __init__(self, + c, + SA, + MER, + coefs_am, + coefs_pm, + learning_rate_am=1., + learning_rate_pm=1., + plot=False): + self.c = c + self.SA = SA + self.MER = MER + + self.learning_rate_am = learning_rate_am + self.learning_rate_pm = learning_rate_pm + + self.coefs_am = coefs_am + self.coefs_am_history = [coefs_am, ] + self.mses_am = [] + self.errs_am = [] + + self.tx_mers = [] + self.rx_mers = [] + + self.coefs_pm = coefs_pm + self.coefs_pm_history = [coefs_pm, ] + self.errs_pm = [] + + self.plot = plot + + def sample_uniformly(self, tx_dpd, rx_received, n_bins=5): + """This function returns tx and rx samples in a way + that the tx amplitudes have an approximate uniform + distribution with respect to the tx_dpd amplitudes""" + mask = np.logical_and((np.abs(tx_dpd) > 0.01), (np.abs(rx_received) > 0.01)) + tx_dpd = tx_dpd[mask] + rx_received = rx_received[mask] + + txframe_aligned_abs = np.abs(tx_dpd) + ccdf_min = 0 + ccdf_max = np.max(txframe_aligned_abs) + tx_hist, ccdf_edges = np.histogram(txframe_aligned_abs, + bins=n_bins, + range=(ccdf_min, ccdf_max)) + n_choise = np.min(tx_hist) + tx_choice = np.zeros(n_choise * n_bins, dtype=np.complex64) + rx_choice = np.zeros(n_choise * n_bins, dtype=np.complex64) + + for idx, bin in enumerate(tx_hist): + indices = np.where((txframe_aligned_abs >= ccdf_edges[idx]) & + (txframe_aligned_abs <= ccdf_edges[idx + 1]))[0] + indices_choise = np.random.choice(indices, + n_choise, + replace=False) + rx_choice[idx * n_choise:(idx + 1) * n_choise] = \ + rx_received[indices_choise] + tx_choice[idx * n_choise:(idx + 1) * n_choise] = \ + tx_dpd[indices_choise] + + assert isinstance(rx_choice[0], np.complex64), \ + "rx_choice is not complex64 but {}".format(rx_choice[0].dtype) + assert isinstance(tx_choice[0], np.complex64), \ + "tx_choice is not complex64 but {}".format(tx_choice[0].dtype) + + return tx_choice, rx_choice + + def dpd_amplitude(self, sig, coefs=None): + if coefs is None: + coefs = self.coefs_am + assert isinstance(sig[0], np.complex64), "Sig is not complex64 but {}".format(sig[0].dtype) + sig_abs = np.abs(sig) + A_sig = np.vstack([np.ones(sig_abs.shape), + sig_abs ** 1, + sig_abs ** 2, + sig_abs ** 3, + sig_abs ** 4, + ]).T + sig_dpd = sig * np.sum(A_sig * coefs, axis=1) + return sig_dpd, A_sig + + def dpd_phase(self, sig, coefs=None): + if coefs is None: + coefs = self.coefs_pm + assert isinstance(sig[0], np.complex64), "Sig is not complex64 but {}".format(sig[0].dtype) + sig_abs = np.abs(sig) + A_phase = np.vstack([np.ones(sig_abs.shape), + sig_abs ** 1, + sig_abs ** 2, + sig_abs ** 3, + sig_abs ** 4, + ]).T + phase_diff_est = np.sum(A_phase * coefs, axis=1) + return phase_diff_est, A_phase + + def _next_am_coefficent(self, tx_choice, rx_choice): + """Calculate new coefficients for AM/AM correction""" + rx_dpd, rx_A = self.dpd_amplitude(rx_choice) + rx_dpd = rx_dpd * ( + np.median(np.abs(tx_choice)) / + np.median(np.abs(rx_dpd))) + err = np.abs(rx_dpd) - np.abs(tx_choice) + mse = np.mean(np.abs((rx_dpd - tx_choice) ** 2)) + self.mses_am.append(mse) + self.errs_am.append(np.mean(err**2)) + + reg = linear_model.Ridge(alpha=0.00001) + reg.fit(rx_A, err) + a_delta = reg.coef_ + new_coefs_am = self.coefs_am - self.learning_rate_am * a_delta + new_coefs_am = new_coefs_am * (self.coefs_am[0] / new_coefs_am[0]) + return new_coefs_am + + def _next_pm_coefficent(self, tx_choice, rx_choice): + """Calculate new coefficients for AM/PM correction + Assuming deviations smaller than pi/2""" + phase_diff_choice = np.angle( + (rx_choice * tx_choice.conjugate()) / + (np.abs(rx_choice) * np.abs(tx_choice)) + ) + plt.hist(phase_diff_choice) + plt.savefig('/tmp/hist_' + str(np.random.randint(0,1000)) + '.svg') + plt.clf() + phase_diff_est, phase_A = self.dpd_phase(rx_choice) + err_phase = phase_diff_est - phase_diff_choice + self.errs_pm.append(np.mean(np.abs(err_phase ** 2))) + + reg = linear_model.Ridge(alpha=0.00001) + reg.fit(phase_A, err_phase) + p_delta = reg.coef_ + new_coefs_pm = self.coefs_pm - self.learning_rate_pm * p_delta + + return new_coefs_pm, phase_diff_choice + + def get_next_coefs(self, tx_dpd, rx_received): + # Check data type + assert tx_dpd[0].dtype == np.complex64, \ + "tx_dpd is not complex64 but {}".format(tx_dpd[0].dtype) + assert rx_received[0].dtype == np.complex64, \ + "rx_received is not complex64 but {}".format(rx_received[0].dtype) + # Check if signals have same normalization + normalization_error = np.abs(np.median(np.abs(tx_dpd)) - + np.median(np.abs(rx_received))) / ( + np.median(np.abs(tx_dpd)) + np.median(np.abs(rx_received))) + assert normalization_error < 0.01, "Non normalized signals" + + tx_choice, rx_choice = self.sample_uniformly(tx_dpd, rx_received) + new_coefs_am = self._next_am_coefficent(tx_choice, rx_choice) + new_coefs_pm, phase_diff_choice = self._next_pm_coefficent(tx_choice, rx_choice) + + logging.debug('txframe: min {:.2f}, max {:.2f}, ' \ + 'median {:.2f}; rxframe: min {:.2f}, max {:.2f}, ' \ + 'median {:.2f}; new coefs_am {};' \ + 'new_coefs_pm {}'.format( + np.min(np.abs(tx_dpd)), + np.max(np.abs(tx_dpd)), + np.median(np.abs(tx_dpd)), + np.min(np.abs(rx_choice)), + np.max(np.abs(rx_choice)), + np.median(np.abs(rx_choice)), + new_coefs_am, + new_coefs_pm)) + + if logging.getLogger().getEffectiveLevel() == logging.DEBUG and self.plot: + off = self.SA.calc_offset(tx_dpd) + tx_mer = self.MER.calc_mer(tx_dpd[off:off + self.c.T_U]) + rx_mer = self.MER.calc_mer(rx_received[off:off + self.c.T_U], debug=True) + self.tx_mers.append(tx_mer) + self.rx_mers.append(rx_mer) + + if logging.getLogger().getEffectiveLevel() == logging.DEBUG and self.plot: + dt = datetime.datetime.now().isoformat() + fig_path = logging_path + "/" + dt + "_Model.svg" + + fig = plt.figure(figsize=(2 * 6, 2 * 6)) + + i_sub = 1 + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + ax.plot(np.abs(tx_dpd[:128]), + label="TX sent", + linestyle=":") + ax.plot(np.abs(rx_received[:128]), + label="RX received", + color="red") + ax.set_title("Synchronized Signals of Iteration {}" + .format(len(self.coefs_am_history))) + ax.set_xlabel("Samples") + ax.set_ylabel("Amplitude") + ax.text(0, 0, "TX (max {:01.3f}, mean {:01.3f}, " \ + "median {:01.3f})".format( + np.max(np.abs(tx_dpd)), + np.mean(np.abs(tx_dpd)), + np.median(np.abs(tx_dpd)) + ), size=8) + ax.legend(loc=4) + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + ccdf_min, ccdf_max = 0, 1 + tx_hist, ccdf_edges = np.histogram(np.abs(tx_dpd), + bins=60, + range=(ccdf_min, ccdf_max)) + tx_hist_normalized = tx_hist.astype(float) / np.sum(tx_hist) + ccdf = 1.0 - np.cumsum(tx_hist_normalized) + ax.semilogy(ccdf_edges[:-1], ccdf, label="CCDF") + ax.semilogy(ccdf_edges[:-1], + tx_hist_normalized, + label="Histogram", + drawstyle='steps') + ax.legend(loc=4) + ax.set_ylim(1e-5, 2) + ax.set_title("Complementary Cumulative Distribution Function") + ax.set_xlabel("TX Amplitude") + ax.set_ylabel("Ratio of Samples larger than x") + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + ax.semilogy(np.array(self.mses_am) + 1e-10, label="log(MSE)") + ax.semilogy(np.array(self.errs_am) + 1e-10, label="log(ERR)") + ax.legend(loc=4) + ax.set_title("MSE History") + ax.set_xlabel("Iterations") + ax.set_ylabel("MSE") + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + ax.plot(self.tx_mers, label="TX MER") + ax.plot(self.rx_mers, label="RX MER") + ax.legend(loc=4) + ax.set_title("MER History") + ax.set_xlabel("Iterations") + ax.set_ylabel("MER") + + ax = plt.subplot(4, 2, i_sub) + rx_range = np.linspace(0, 1, num=100, dtype=np.complex64) + rx_range_dpd = self.dpd_amplitude(rx_range)[0] + rx_range_dpd_new = self.dpd_amplitude(rx_range, new_coefs_am)[0] + i_sub += 1 + ax.scatter( + np.abs(tx_choice), + np.abs(rx_choice), + s=0.1) + ax.plot(rx_range_dpd / self.coefs_am[0], rx_range, linewidth=0.25, label="current") + ax.plot(rx_range_dpd_new / self.coefs_am[0], rx_range, linewidth=0.25, label="next") + ax.set_ylim(0, 1) + ax.set_xlim(0, 1) + ax.legend() + ax.set_title("Amplifier Characteristic") + ax.set_xlabel("TX Amplitude") + ax.set_ylabel("RX Amplitude") + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + coefs_am_history = np.array(self.coefs_am_history) + for idx, coef_hist in enumerate(coefs_am_history.T): + ax.plot(coef_hist, + label="Coef {}".format(idx), + linewidth=0.5) + ax.legend(loc=4) + ax.set_title("AM/AM Coefficient History") + ax.set_xlabel("Iterations") + ax.set_ylabel("Coefficient Value") + + phase_range = np.linspace(0, 1, num=100, dtype=np.complex64) + phase_range_dpd = self.dpd_phase(phase_range)[0] + phase_range_dpd_new = self.dpd_phase(phase_range, + coefs=new_coefs_pm)[0] + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + ax.scatter( + np.abs(tx_choice), + np.rad2deg(phase_diff_choice), + s=0.1) + ax.plot( + np.abs(phase_range), + np.rad2deg(phase_range_dpd), + linewidth=0.25, + label="current") + ax.plot( + np.abs(phase_range), + np.rad2deg(phase_range_dpd_new), + linewidth=0.25, + label="next") + ax.set_ylim(-60, 60) + ax.set_xlim(0, 1) + ax.legend() + ax.set_title("Amplifier Characteristic") + ax.set_xlabel("TX Amplitude") + ax.set_ylabel("Phase Difference") + + ax = plt.subplot(4, 2, i_sub) + i_sub += 1 + coefs_pm_history = np.array(self.coefs_pm_history) + for idx, coef_phase_hist in enumerate(coefs_pm_history.T): + ax.plot(coef_phase_hist, + label="Coef {}".format(idx), + linewidth=0.5) + ax.legend(loc=4) + ax.set_title("AM/PM Coefficient History") + ax.set_xlabel("Iterations") + ax.set_ylabel("Coefficient Value") + + fig.tight_layout() + fig.savefig(fig_path) + fig.clf() + + self.coefs_am = new_coefs_am + self.coefs_am_history.append(self.coefs_am) + self.coefs_pm = new_coefs_pm + self.coefs_pm_history.append(self.coefs_pm) + return self.coefs_am, self.coefs_pm + +# The MIT License (MIT) +# +# Copyright (c) 2017 Andreas Steger +# +# Permission is hereby granted, free of charge, to any person obtaining a copy +# of this software and associated documentation files (the "Software"), to deal +# in the Software without restriction, including without limitation the rights +# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +# copies of the Software, and to permit persons to whom the Software is +# furnished to do so, subject to the following conditions: +# +# The above copyright notice and this permission notice shall be included in all +# copies or substantial portions of the Software. +# +# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +# SOFTWARE. |