gr-digitalhf/lib/adaptive_dfe_impl.cc

456 lines
17 KiB
C++

/* -*- c++ -*- */
/*
* Copyright 2018 hcab14@mail.com.
*
* This is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3, or (at your option)
* any later version.
*
* This software is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this software; see the file COPYING. If not, write to
* the Free Software Foundation, Inc., 51 Franklin Street,
* Boston, MA 02110-1301, USA.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <boost/format.hpp>
#include <gnuradio/math.h>
#include <gnuradio/expj.h>
#include <gnuradio/io_signature.h>
#include <gnuradio/logger.h>
#include <volk/volk.h>
#include "adaptive_dfe_impl.h"
#include "lms.hpp"
#include "nlms.hpp"
#include "rls.hpp"
#include "kalman_exp.hpp"
#include "kalman_hsu.hpp"
namespace gr {
namespace digitalhf {
adaptive_dfe::sptr
adaptive_dfe::make(int sps, // samples per symbol
int nB, // number of forward FIR taps
int nF, // number of backward FIR taps
int nW, // number of feedback taps
float mu,
float alpha)
{
return gnuradio::get_initial_sptr
(new adaptive_dfe_impl(sps, nB, nF, nW, mu, alpha));
}
adaptive_dfe_impl::adaptive_dfe_impl(int sps, // samples per symbol
int nB, // number of forward FIR taps
int nF, // number of backward FIR taps
int nW, // number of feedback taps
float mu,
float alpha)
: gr::block("adaptive_dfe",
gr::io_signature::make(1, 1, sizeof(gr_complex)),
gr::io_signature::make2(2, 2,
sizeof(gr_complex),
sizeof(gr_complex)*(sps*(nF+nB)+1)))
, _sps(sps)
, _nB(nB*sps)
, _nF(nF*sps)
, _nW(nW)
, _nGuard(2*sps)
, _mu(mu)
, _alpha(alpha)
, _use_symbol_taps(true)
, _tmp()
, _taps_samples()
, _taps_symbols()
, _hist_symbols()
, _hist_symbol_index(0)
, _constellations()
, _npwr()
, _npwr_max_time_constant(10)
, _constellation_index()
, _symbols()
, _scramble()
, _scramble_xor()
, _descrambled_symbols()
, _symbol_counter(0)
, _save_soft_decisions(false)
, _vec_soft_decisions()
, _msg_ports{{"soft_dec", pmt::intern("soft_dec")},
{"frame_info", pmt::intern("frame_info")}}
, _msg_metadata(pmt::make_dict())
, _num_samples_since_filter_update(0)
, _rotated_samples()
, _rotator()
, _control_loop(2*M_PI/100, 5e-2, -5e-2)
, _state(WAIT_FOR_PREAMBLE)
, _filter_update()
{
GR_LOG_DECLARE_LOGPTR(d_logger);
GR_LOG_ASSIGN_LOGPTR(d_logger, "adaptive_dfe");
set_history(_nGuard+_nB+1);
message_port_register_out(_msg_ports["soft_dec"]);
pmt::pmt_t constellations_port = pmt::intern("constellations");
message_port_register_in(constellations_port);
set_msg_handler(constellations_port, boost::bind(&adaptive_dfe_impl::update_constellations, this, _1));
pmt::pmt_t frame_info_port = _msg_ports["frame_info"];
message_port_register_in(frame_info_port);
message_port_register_out(frame_info_port);
set_msg_handler(frame_info_port, boost::bind(&adaptive_dfe_impl::update_frame_info, this, _1));
}
adaptive_dfe_impl::~adaptive_dfe_impl()
{
_msg_metadata = pmt::PMT_NIL;
}
void
adaptive_dfe_impl::forecast(int noutput_items, gr_vector_int &ninput_items_required)
{
// [guard | nB | 1 | nF | guard ]
ninput_items_required[0] = _sps*noutput_items + 2*_nGuard + _nB + _nF + 1;
}
int
adaptive_dfe_impl::general_work(int noutput_items,
gr_vector_int &ninput_items,
gr_vector_const_void_star &input_items,
gr_vector_void_star &output_items)
{
gr::thread::scoped_lock lock(d_setlock);
gr_complex const* in = (gr_complex const *)input_items[0];
gr_complex *out_symb = (gr_complex *)output_items[0];
gr_complex *out_taps = (gr_complex *)output_items[1];
const int nin = ninput_items[0];
// GR_LOG_DEBUG(d_logger, str(boost::format("work: %d %d") % ninput_items[0] % (2*_nGuard + _nB + _nF + 1)));
assert(ninput_items[0] >= 2*_nGuard + _nB + _nF + 1);
if (ninput_items[0] < 2*_nGuard + _nB + _nF + 1)
return 0;
int const ninput = ninput_items[0] - _nGuard - _nF;
int nout = 0; // counter for produced output items
switch (_state) {
case WAIT_FOR_PREAMBLE: {
std::vector<tag_t> v;
get_tags_in_window(v, 0, history()-1, ninput, pmt::intern("preamble_start"));
if (v.empty()) {
consume(0, ninput - history()+1);
} else {
tag_t const& tag = v.front();
reset_filter();
_descrambled_symbols.clear();
publish_frame_info();
consume(0, tag.offset - nitems_read(0));
_state = WAIT_FOR_FRAME_INFO;
GR_LOG_DEBUG(d_logger, "got preamble tag > wait for frame info");
}
_filter_update->reset();
break;
} // WAIT_FOR_PREAMBLE
case WAIT_FOR_FRAME_INFO: {
//GR_LOG_DEBUG(d_logger, "WAIT_FOR_FRAME_INFO");
//update_frame_info(delete_head_blocking(_msg_ports["frame_info"]));
break;
} // WAIT_FOR_FRAME_INFO
case DO_FILTER: {
_rotated_samples.resize(ninput+_nF+1);
int ninput_processed = 0;
for (int i0=history()-1, i=i0; i<ninput && nout<noutput_items; i+=_sps, ninput_processed+=_sps) {
if (_symbol_counter == _symbols.size()) {
publish_frame_info();
publish_soft_dec();
_symbol_counter = 0;
int const shift = recenter_filter_taps();
if (shift != 0)
ninput_processed += shift;
_state = WAIT_FOR_FRAME_INFO;
break;
}
// rotate samples
if (i == i0) {
#ifdef USE_VOLK_ROTATOR
_rotator.rotateN(&_rotated_samples[0] + i - _nB,
in + i - _nB,
_nB+_nF+1);
#else
for (int j=0; j<_nB+_nF+1; ++j)
_rotated_samples[j + i-_nB] = _rotator.rotate(in[j + i-_nB]);
#endif
} else {
#ifdef USE_VOLK_ROTATOR
_rotator.rotateN(&_rotated_samples[0] + i + _nF+1 - _sps,
in + i + _nF+1 - _sps,
_sps);
#else
for (int j=0; j<_sps; ++j)
_rotated_samples[j + i+_nF+1-_sps] = _rotator.rotate(in[j + i+_nF+1-_sps]);
#endif
}
assert(i+_nF < nin && i-1-_nB >= 0);
out_symb[nout] = filter(_rotated_samples.data() + i - _nB,
_rotated_samples.data() + i + _nF+1);
std::memcpy(&out_taps[(_nB+_nF+1)*nout], _taps_samples.data(), (_nB+_nF+1)*sizeof(gr_complex));
++nout;
} // next sample
consume(0, ninput_processed);
break;
} // DO_FILTER
}
return nout;
}
bool adaptive_dfe_impl::start()
{
gr::thread::scoped_lock lock(d_setlock);
_tmp.resize(_nB+_nF+1);
_taps_samples.resize(_nB+_nF+1);
_last_taps_samples.resize(_nB+_nF+1);
_taps_symbols.resize(_nW);
_hist_symbols.resize(2*_nW);
reset_filter();
GR_LOG_DEBUG(d_logger,str(boost::format("adaptive_dfe_impl::start() nB=%d nF=%d mu=%f alpha=%f")
% _nB % _nF % _mu % _alpha));
_filter_update = lms::make(_mu);
// _filter_update = nlms::make(0.5f, 1.0f);
// _filter_update = rls::make(0.001f, 0.9f); // not stable
// _filter_update = kalman_exp::make(1.0f, 0.9f); // too slow
// _filter_update = kalman_hsu::make(0.02f, 0.1f); // not stable
return true;
}
bool adaptive_dfe_impl::stop()
{
gr::thread::scoped_lock lock(d_setlock);
GR_LOG_DEBUG(d_logger, "adaptive_dfe_impl::stop()");
_filter_update.reset();
return true;
}
gr_complex adaptive_dfe_impl::filter(gr_complex const* start, gr_complex const* end) {
assert(end-start == _nB + _nF + 1);
// (1) run the filter filter
gr_complex filter_output(0);
// (1a) taps_samples
volk_32fc_x2_dot_prod_32fc(&filter_output,
start,
_taps_samples.data(),
_nB+_nF+1);
// (1b) taps_symbols
gr_complex dot_symbols(0);
gr::digital::constellation_sptr constell = _constellations[_constellation_index];
_use_symbol_taps = (constell->bits_per_symbol() <= 3);
if (_use_symbol_taps) {
for (int l=0; l<_nW; ++l) {
assert(_hist_symbol_index+l < 2*_nW);
dot_symbols += _hist_symbols[_hist_symbol_index+l]*_taps_symbols[l];
}
filter_output += dot_symbols;
}
assert(_symbol_counter < _symbols.size());
gr_complex known_symbol = _symbols[_symbol_counter];
bool const is_known = std::abs(known_symbol) > 1e-5;
bool const update_taps = constell->bits_per_symbol() <= 3 || is_known;
// (2) unknown symbols (=data): compute soft decisions
if (not is_known) {
gr_complex const descrambled_filter_output = std::conj(_scramble[_symbol_counter]) * filter_output;
unsigned int const jc = constell->decision_maker(&descrambled_filter_output);
gr_complex descrambled_symbol = 0;
constell->map_to_points(jc, &descrambled_symbol);
if (_save_soft_decisions) {
float const err = std::abs(descrambled_filter_output - descrambled_symbol);
std::vector<float> const soft_dec = constell->calc_soft_dec
(descrambled_filter_output, _npwr[_constellation_index].filter(err));
for (int j=0, m=soft_dec.size(); j<m; ++j)
_vec_soft_decisions.push_back(soft_dec[j] * _scramble_xor[_symbol_counter][j]);
}
known_symbol = _scramble[_symbol_counter] * descrambled_symbol;
}
// (3) filter update
if (update_taps) {
_num_samples_since_filter_update += _sps;
// (3a) update of adaptive filter taps
gr_complex const err = known_symbol - filter_output;
if (std::abs(err)>0.7)
std::cout << "err= " << std::abs(err) << std::endl;
// taps_samples
gr_complex const* gain = _filter_update->update(start, end);
std::copy(_taps_samples.begin(), _taps_samples.end(), _last_taps_samples.begin());
volk_32fc_s32fc_multiply_32fc(&_tmp[0], gain, err, _nB+_nF+1);
volk_32fc_x2_add_32fc (&_taps_samples[0], &_taps_samples[0], &_tmp[0], _nB+_nF+1);
// taps_symbols
if (_use_symbol_taps) {
for (int j=0; j<_nW; ++j) {
assert(_hist_symbol_index+j < 2*_nW);
_taps_symbols[j] -= _mu*err*std::conj(_hist_symbols[_hist_symbol_index+j]) + _alpha*_taps_symbols[j];
}
_hist_symbols[_hist_symbol_index] = _hist_symbols[_hist_symbol_index + _nW] = known_symbol;
if (++_hist_symbol_index == _nW)
_hist_symbol_index = 0;
}
}
// (3b) control loop update for doppler correction using the adaptibve filter taps
if (update_taps) {
if (_symbol_counter != 0) { // a filter tap shift might have ocurred when _symbol_counter==0
gr_complex acc(0);
volk_32fc_x2_conjugate_dot_prod_32fc(&acc, &_taps_samples[0], &_last_taps_samples[0], _nB+_nF+1);
float const frequency_err = gr::fast_atan2f(acc)/(0+1*_num_samples_since_filter_update); // frequency error (rad/sample)
GR_LOG_DEBUG(d_logger, str(boost::format("frequency_err= %f %d") % frequency_err % _num_samples_since_filter_update));
_control_loop.advance_loop(frequency_err);
_control_loop.phase_wrap();
_control_loop.frequency_limit();
_rotator.set_phase_incr(gr_expj(_control_loop.get_frequency()));
GR_LOG_DEBUG(d_logger, str(boost::format("frequency_err= %f %d %f")
% (frequency_err/(2*M_PI)*12000.0)
% _num_samples_since_filter_update
% _control_loop.get_frequency()));
}
_num_samples_since_filter_update = 0;
}
// (4) save the descrambled symbol (-> frame_info)
_descrambled_symbols[_symbol_counter] = filter_output*std::conj(_scramble[_symbol_counter]);
return _descrambled_symbols[_symbol_counter++];
}
int
adaptive_dfe_impl::recenter_filter_taps() {
float sum_w=0, sum_wi=0;
for (int i=0; i<_nB+_nF+1; ++i) {
float const w = std::norm(_taps_samples[i]);
sum_w += w;
sum_wi += w*i;
}
// TODO
static float mp = _nB+1;
mp = 0.9*mp + 0.1*sum_wi/sum_w;
ssize_t const idx_max = ssize_t(0.5 + mp);
// GR_LOG_DEBUG(d_logger, str(boost::format("idx_max=%2d abs(tap_max)=%f") % idx_max % std::abs(_taps_samples[idx_max])));
if (idx_max-_nB-1 > +2*_sps) {
// maximum is right of the center tap
// -> shift taps to the left left
GR_LOG_DEBUG(d_logger, "shift left");
std::copy(_taps_samples.begin()+4*_sps, _taps_samples.begin()+_nB+_nF+1, _taps_samples.begin());
std::fill_n(_taps_samples.begin()+_nB+_nF+1-4*_sps, 4*_sps, gr_complex(0));
return +4*_sps;
}
if (idx_max-_nB-1 < -2*_sps) {
// maximum is left of the center tap
// -> shift taps to the right
GR_LOG_DEBUG(d_logger, "shift right");
std::copy_backward(_taps_samples.begin(), _taps_samples.begin()+_nB+_nF+1-4*_sps,
_taps_samples.begin()+_nB+_nF+1);
std::fill_n(_taps_samples.begin(), 4*_sps, gr_complex(0));
return -4*_sps;
}
return 0;
}
void adaptive_dfe_impl::reset_filter()
{
std::fill(_taps_samples.begin(), _taps_samples.end(), gr_complex(0));
std::fill(_last_taps_samples.begin(), _last_taps_samples.end(), gr_complex(0));
std::fill(_taps_symbols.begin(), _taps_symbols.end(), gr_complex(0));
std::fill(_hist_symbols.begin(), _hist_symbols.end(), gr_complex(0));
_taps_symbols[0] = 1;
_hist_symbol_index = 0;
_num_samples_since_filter_update = 0;
}
void adaptive_dfe_impl::publish_frame_info()
{
pmt::pmt_t data = pmt::make_dict();
GR_LOG_DEBUG(d_logger, str(boost::format("publish_frame_info %d == %d") % _descrambled_symbols.size() % _symbols.size()));
data = pmt::dict_add(data,
pmt::intern("symbols"),
pmt::init_c32vector(_descrambled_symbols.size(), _descrambled_symbols));
// for (int i=0; i<_vec_soft_decisions.size(); ++i)
// _vec_soft_decisions[i] = std::max(-1.0f, std::min(1.0f, _vec_soft_decisions[i]));
data = pmt::dict_add(data,
pmt::intern("soft_dec"), pmt::init_f32vector(_vec_soft_decisions.size(), _vec_soft_decisions));
message_port_pub(_msg_ports["frame_info"], data);
_descrambled_symbols.clear();
}
void adaptive_dfe_impl::publish_soft_dec()
{
if (_vec_soft_decisions.empty())
return;
message_port_pub(_msg_ports["soft_dec"],
pmt::cons(pmt::dict_add(_msg_metadata, pmt::intern("packet_len"), pmt::mp(_vec_soft_decisions.size())),
pmt::init_f32vector(_vec_soft_decisions.size(), _vec_soft_decisions)));
_vec_soft_decisions.clear();
}
void adaptive_dfe_impl::update_constellations(pmt::pmt_t data) {
int const n = pmt::length(data);
_constellations.resize(n);
_npwr.resize(n);
GR_LOG_DEBUG(d_logger, str(boost::format("update_constellations %s n=%d") % data % n));
unsigned int const rotational_symmetry = 0;
unsigned int const dimensionality = 1;
for (int i=0; i<n; ++i) {
pmt::pmt_t c = pmt::vector_ref(data, i);
int const idx = pmt::to_long(pmt::dict_ref(c, pmt::intern("idx"), pmt::from_long(-1)));
assert(idx>=0 && idx < n);
_constellations[idx] = gr::digital::constellation_calcdist::make
(pmt::c32vector_elements(pmt::dict_ref(c, pmt::intern("points"), pmt::PMT_NIL)),
pmt::s32vector_elements(pmt::dict_ref(c, pmt::intern("symbols"), pmt::PMT_NIL)),
rotational_symmetry, dimensionality);
_npwr[i].reset(_npwr_max_time_constant);
}
}
void adaptive_dfe_impl::update_frame_info(pmt::pmt_t data)
{
//GR_LOG_DEBUG(d_logger,str(boost::format("adaptive_dfe_impl::update_frame_info() %s") % data));
_symbols = pmt::c32vector_elements(pmt::dict_ref(data, pmt::intern("symb"), pmt::PMT_NIL));
_scramble = pmt::c32vector_elements(pmt::dict_ref(data, pmt::intern("scramble"), pmt::PMT_NIL));
_constellation_index = pmt::to_long(pmt::dict_ref(data, pmt::intern("constellation_idx"), pmt::PMT_NIL));
_save_soft_decisions = pmt::to_bool(pmt::dict_ref(data, pmt::intern("save_soft_dec"), pmt::PMT_F));
bool const do_continue = pmt::to_bool(pmt::dict_ref(data, pmt::intern("do_continue"), pmt::PMT_F));
// make table +-1
std::vector<std::uint8_t> const scr_xor = pmt::u8vector_elements(pmt::dict_ref(data, pmt::intern("scramble_xor"), pmt::PMT_NIL));
_scramble_xor.resize(scr_xor.size());
gr::digital::constellation_sptr constell = _constellations[_constellation_index];
for (int i=0, n=scr_xor.size(); i<n; ++i) {
for (int j=0, m=constell->bits_per_symbol(); j<m; ++j) {
_scramble_xor[i][j] = 1 - 2*bool(scr_xor[i] & (1<<(m-1-j)));
// GR_LOG_DEBUG(d_logger, str(boost::format("XOR %3d %3d %d") % i % j % _scramble_xor[i][j]));
}
}
assert(_symbols.size() == _scramble.size());
_descrambled_symbols.resize(_symbols.size());
_vec_soft_decisions.clear();
_symbol_counter = 0;
_state = (do_continue ? DO_FILTER : WAIT_FOR_PREAMBLE);
}
} /* namespace digitalhf */
} /* namespace gr */