mirror of
https://github.com/gnss-sdr/gnss-sdr
synced 2026-10-16 04:31:50 +00:00
joebre
5c24058459
load_f_error_dump/find_f_error_scans/plot_f_error_scan (utils/matlab/libs) read and visualize the per-bin CSV written by the tracking f_error_dump diagnostic, and plot_all_f_error_scans (utils/matlab) batches plot_f_error_scan over every scan in a dump, optionally saving PNGs. f_error_sim (utils/matlab/libs) is a self-contained Monte Carlo simulation of the same bin-scan/CN0-estimator logic, useful for sizing f_error_step_num/f_error_accumulation/f_error_doppler_step without a live capture.
60 lines
2.4 KiB
Matlab
60 lines
2.4 KiB
Matlab
% plot_f_error_scan.m
|
|
%
|
|
% Bar-plots one state-5 frequency-error-reduction scan (as loaded by
|
|
% load_f_error_dump), with the selected/winning Doppler marked. Optionally
|
|
% overlays the theoretical sinc(pi*Delta_f*Td)^2 envelope (normalized to the
|
|
% scan's own peak) for comparison, if the coherent integration time Td [s] is
|
|
% given -- e.g. Td=0.004 for Galileo E1, Td=0.001 for GPS L1 C/A.
|
|
%
|
|
% Usage:
|
|
% scans = load_f_error_dump();
|
|
% plot_f_error_scan(scans, 5); % plot scans(5)
|
|
% plot_f_error_scan(scans, 5, 0.004); % + theoretical Galileo E1 sinc^2 overlay
|
|
%
|
|
% See also: load_f_error_dump, find_f_error_scans
|
|
|
|
% -------------------------------------------------------------------------
|
|
%
|
|
% GNSS-SDR is a Global Navigation Satellite System software-defined receiver.
|
|
% This file is part of GNSS-SDR.
|
|
%
|
|
% SPDX-FileCopyrightText: Joel Brenner, 2026. joel.brenner(at)saphyrion.ch
|
|
% SPDX-License-Identifier: GPL-3.0-or-later
|
|
%
|
|
% -------------------------------------------------------------------------
|
|
|
|
function plot_f_error_scan(scans, idx, Td)
|
|
if idx < 1 || idx > numel(scans)
|
|
error('plot_f_error_scan: idx %d out of range (1..%d)', idx, numel(scans));
|
|
end
|
|
s = scans(idx);
|
|
|
|
figure();
|
|
bar(s.doppler_hz, s.power);
|
|
hold on;
|
|
|
|
if nargin >= 3 && ~isempty(Td)
|
|
% theoretical envelope centered on the selected (winning) Doppler, since that's
|
|
% this scan's best estimate of the true residual frequency
|
|
theta = pi * (s.doppler_hz - s.selected_doppler_hz) * Td;
|
|
sinc0 = @(x) (x == 0) * 1 + (x ~= 0) .* (sin(x + (x == 0)) ./ (x + (x == 0)));
|
|
envelope = (sinc0(theta) .^ 2) * max(s.power);
|
|
plot(s.doppler_hz, envelope, 'k--', 'LineWidth', 1.3);
|
|
end
|
|
|
|
yl = ylim();
|
|
plot([s.selected_doppler_hz, s.selected_doppler_hz], yl, 'r--', 'LineWidth', 1.5);
|
|
hold off;
|
|
|
|
xlabel('Doppler tested [Hz]');
|
|
ylabel('Accumulated correlation power');
|
|
if nargin >= 3 && ~isempty(Td)
|
|
legend('bin power', 'theoretical sinc^2 (centered on selected)', 'selected Doppler', 'Location', 'northeast');
|
|
else
|
|
legend('bin power', 'selected Doppler', 'Location', 'northeast');
|
|
end
|
|
title(sprintf('PRN%d ch%d scan %d: CN0=%.1f dB-Hz, selected=%.1f Hz (bin %d/%d)', ...
|
|
s.prn, s.channel, s.scan_id, s.cn0_dBHz, s.selected_doppler_hz, s.selected_bin, numel(s.doppler_hz) - 1));
|
|
grid on;
|
|
end
|