compsim4.m

阵列信号处理的工具箱

function [sigvar, tgtSig] = compsim4(antenna, waveform, noSamp, tgtType, tgtParam, noiseType, noiseParam, clutterType, clutterParam)

%COMPSIM4 Simulates received antenna signals in passive mode.
%
%--------
%Synopsis:
%  compsim4(antenna, waveform, noSamp, tgtType, tgtParam, noiseType,
%    noiseParam)
%  [sigvar, tgtSig] = compsim4(antenna, waveform, noSamp, tgtType,
%    tgtParam, noiseType, noiseParam)
%
%  compsim4(antenna, lambda, noSamp, tgtType, tgtParam, noiseType,
%    noiseParam)
%  [sigvar, tgtSig] = compsim4(antenna, lambda, noSamp, tgtType,
%    tgtParam, noiseType, noiseParam)
%
%Output and Input:
%  sigvar (RxRadarSigT): Simulated radar signal.
%  tgtSig (CxMatrixT): The simulated target signal in a
%    (noOfTargets x noOfSnapshots) matrix.
%  antenna (AntDefT): Antenna definition for the receiver.
%  waveform (WaveformT): Waveform, including wavelength,
%    number of range bins, number of pulses, modulation within a pulse etc.
%  lambda (RealScalarT): The wavelength (lambda) [m].
%  noSamp (IntScalarT): Number of snapshots.
%  tgtType (StringT): Target type, se below.
%  tgtParam: Type and contents depends on tgtType, see below.
%  noiseType (StringT): Noise type, se below.
%  noiseParam: Type and contents depends on noiseType, see below.
%  clutterType (StringT): Clutter type, se below.
%  clutterParam: Type and contents depends on clutterType, see below.
%
%Description
%  Simulates received antenna signals in passive mode (listen only).
%  The target and noise signals are summed to form the output signal.
%  Possible target signals are random signals with a complex gaussian
%  distribution and an arbitrary source-to-source correlation matrix
%  or constant signals with Doppler shifts or no target signal at all.
%  Possible noise signals are complex gaussian noise with arbitrary
%  spatial correlation matrix or no noise signal at all.
%  There is no range information, i.e. there is only one range bin.
%
%  In this function the term "signal-to-noise ratio" means the ratio of signal
%  power to noise power for each targets signal, each channel and each time
%  snapshot. The signal-to-noise ratio is not allways easy to determine.
%  If tgtType = 'const' or 'rndnw', the received target signals are
%  created with a signal-to-noise ratio related to a one-variance noise
%  using the SNR input parameter. If tgtType = 'rndn' then the power of
%  the targets signals are given by the source correlation matrix tgtCorrMat.
%  In this case the SNR input parameter is not used. Then, if noiseType
%  = rndnw, noise with variance one is added. If noiseType = rndn, a noise
%  is added, whose power is given by the noise  correlation matrix noiseParam.
%
%  If the noise power is 1 and the signal power is the value of the input
%  parameter SNR, the formula X=A*B gives the expected signal-to-noise ratio
%  (=SNR) only if the the absolute value of all elements in A is 1. This
%  may be true for an ULA with isotropical elements but it is not true in
%  general.
%
%  This complicated description means that the resulting signal-to-noise
%  ratio not allways is the same as the SNR input  parameter. Only when
%  tgtType = 'const' or 'rndnw' and noiseType = rndnw and absolute value
%  of all elements in A is 1, the SNR input parameter is equal to the
%  signal-to-noise ratio.
%
%  The targets signals generated by this function corresponds to Swerling
%  cases 1,3 and 5 ([2] p. 95-99, [3] p. 65-67) in radar theory. Maybe
%  this is applicable only for tgtType = 'const' (constant target signals
%  with doppler shift) and not for random signals.
%  Swerling cases 1,3 and 5 means that the amplitude of the signal is
%  unchanged between adjacent snapshots (pulses) but can change between
%  different trials (CPI:s, coherent processing intervalls). Currently
%  this function only can generate signals for one trial. %  Swerling
%  case 1 means that the amplitude exhibits large fluctuations
%  between different CPI:s. Swerling case 3 means that the amplitude
%  exhibits small fluctuations between different CPI:s. Swerling case 5
%  (also called Swerling case 0 or Marcum case) means that the amplitude
%  does not fluctuate between different CPI:s. It is possible to simulate
%  Swerling cases 1,3 and 5 with this function by calling it once for each
%  CPI and letting yhe SNR input parameter be a random variable with suitable
%  probability distristribution, see [2,3].
%
%--------
%Target signals
%  tgtType = 'const': Constant source signals.
%    tgtParam Matrix = [theta, phi, SNR, alpha, dalpha, dist]
%      (Name) (Data Type)  (Description)
%      theta   (RealVectorT): Target directions (radians).
%      phi (RealVectorT): Target directions (radians). Unused.
%      SNR (RealVectorT): Target SNR:s measured per channel and snapshot
%        after the receiver and A/D-converter. Se "Description" above for
%        description of signal-to-noise ratio.
%      alpha (RealVectorT): Target start phases.
%      dalpha (RealVectorT): Target phase shifts (doppler shifts).
%      dist (RealVectorT): Target distances.
%      The parameter tgtCorrMat is unused.
%  tgtType ='rndnw'  : Complex gaussian random source signals. Uncorrelated
%    between targets and in time.
%    tgtParam Matrix = [theta, phi, SNR, alpha, dalpha, dist]
%      Parameters theta, phi, SNR, dalpha, dist as for 'const'.
%      The other parameters are unused.
%  tgtType ='rndn': Complex gaussian random source signals.
%    tgtParam Matrix = [theta, phi, SNR, alpha, dalpha, dist, tgtCorrMat]
%      tgtCorrMat (CxMatrixT): Source correlation matrix.
%      Parameters theta, phi, dalpha, dist as for 'const'.
%      The other parameters are unused. Note especially that SNR is not used
%        but tgtCorrMat.
%  tgtType ='notgt': No target signals.
%
%--------
%Noise signals
%  noiseType = 'rndnw': Complex gaussian noise. Uncorrelated between channels
%    and in time.
%  noiseType = 'rndn': Complex gaussian noise.
%    noiseParam (CxMatrixT): Noise correlation matrix.
%  noiseType = 'nonoise': No noise.
%
%--------
%Notations:
%  Data type names are shown in parentheses and they start with a capital
%  letter and end with a capital T. Data type definitions can be found in [1]
%  or by "help dbtdata".
%  [D] = This parameter can be omitted and then a default value is used.
%  When the [D]-input parameter is not the last used in the call, it must be
%  given the value [], i.e. an empty matrix.
%  ... = There can be more parameters. They are explained under respective
%  metod or choice.
%
%Examples:
%Algoritm:
%Software Quality
%  (About what is done to ascertain software quality. What tests are done.)
%
%Known Bugs:
%  Not all possible input parameters of the function "spastemat" is used;
%  pointing directions of subarray and antenna errors are not specified.
%
%References:
%  [1]: Bj鰎klund S.: "DBT, A MATLAB Toolbox for Radar Signal Processing.
%    Reference Guide", FOA-D--9x-00xxx-408--SE, To be published.
%  [2]: Kingsley S., Quegan S.: "Understanding Radar Systems",
%     McGraw-Hill 1992. ISBN 0-07-707426-2.
%  [3]: Sume A.: "Kursmaterial i radarteori", FOA 1995, FOA-D--95-00108-3.3--SE.
%
%See Also:
%  compsim5, expsig1, ecorrm, simcherr, doaspc1, doapar1, dbthelp, dbtex1

%   *  DBT, A Matlab Toolbox for Radar Signal Processing  *
% (c) FOA 1994-2000. See the file dbtright.m for copyright notice.
%
%  Start        : 951220 Svante Bj鰎klund (svabj).
%  Latest change: $Date: 2000/10/16 15:20:12 $ $Author: svabj $.
%  $Revision: 1.11 $
% *****************************************************************************

% Sample time also as input parameter!?! Waveform or PRI also ?

%  This function only manage PRI = T (= Sampling interval). This means that
%  there is only one range bin. <- ???

%  noTrials (IntScalarT): Number of trials. Not implemented.


% ----------------------------------------------------------------------- %
% Pick out fields from input parameters.
% ----------------------------------------------------------------------- %

chkdtype(antenna, 'AntDefT')
chkdtype(noSamp, 'IntScalarT')
chkdtype(tgtType, 'StringT')
chkdtype(noiseType, 'StringT')

noElem =      antenna.noElem;

if (isstruct(waveform))
  lambda = waveform.wavelength;
else
  % Allows a simplified call.
  lambda = waveform;
end%if

T =      noSamp;

M =      size(tgtParam,1); % Number of targets.
theta =  tgtParam(:,1);    % Target directions (radians).
%phi =      tgtParam(:,2);    % Target directions (radians).
SNR =       tgtParam(:,3);    % SNR for each target.
alpha =  tgtParam(:,4);    % Phase at the first antenna
               %  element for each target.
dalpha =    tgtParam(:,5);    % Phase increment for each
               %  target (doppler shift).
dist =      tgtParam(:,6);
if (size(tgtParam,2) > 6)
  tgtCorrMat =    tgtParam(:,7:(7-1+size(tgtParam,1)));
end


% ----------------------------------------------------------------------- %
% Return the current seed of normal generator and set it to different values
% each time.
% ----------------------------------------------------------------------- %

randn('seed',sum(100*clock));
rand('seed',sum(100*clock));


% ----------------------------------------------------------------------- %
% Simulate the signals. Preliminary testings.
% ----------------------------------------------------------------------- %

if (nargin > 7)
  error('DBT-Error: Clutter is not implemented.')
end%if


% ----------------------------------------------------------------------- %
% Simulate the TARGET signals.
% ----------------------------------------------------------------------- %
% B(m,n) 鋜 en matris med amplitud och fas hos m錶signalerna vid ena 鋘den
% av gruppantennen.
% m 鋜 m錶index och n tidsindex. (m 鋜 radindex och n kolumnindex i matrisen)
% ----------------------------------------------------------------------- %

% Bilda en m錶signalmatris.
M = size(theta,1);   % Antal m錶.
if (strcmp(tgtType,'const'))
  % The noise power is assumed to be 1. The power of each targets signal
  % is therefore equal to the value given by the input parameter SNR.
  % This means that the amplitude of the signals should be 10.^(SNR./20).
  B= (10.^(SNR./20).*exp(j*alpha))*ones(1,T);
elseif (strcmp(tgtType,'rndnw'))
  % The noise power is assumed to be 1. The power of each targets signal
  % is therefore equal to the value given by the input parameter SNR.
  sigmab = sqrt(0.5)*(10.^(SNR./20))*ones(1,T); % Source standard deviation.
  B = sigmab.*(randn(M,T) + j*randn(M,T));
elseif (strcmp(tgtType,'rndn')) % This part is from the FOA SASP course 96.
  if (size(tgtCorrMat,2) < size(tgtParam,1))
    error('DBT-Error: Missing source correlation matrix.')
  end
  [U,S,V] = svd(tgtCorrMat);
  B=(1/sqrt(2))*U*sqrt(S)*(randn(M,T) + j*randn(M,T));
elseif (strcmp(tgtType,'notgt'))
  B=0;
else
  error('DBT-Error: The source model does not exist.')
end%if

% Ordna konstant fasskift mellan tidssamplena.
% Simulates dopples shift due to movement.

B= B.*exp(j*dalpha*(0:(T-1)));


% ----------------------------------------------------------------------- %
% Simulate the NOISE-signals.
% ----------------------------------------------------------------------- %
% Nt(m,n) inneh錶ler komplext brus. m 鋜 antennelementindex. n 鋜 tidsindex.
% ----------------------------------------------------------------------- %

if (strcmp(noiseType,'rndnw'))
  sigman = sqrt(0.5);
    % Noise standard deviation for the real part and for the imaginary
    % part. The total power of the noise is 0.5 + 0.5 (real + imaginary) = 1.
  Nt  = sigman*(randn(noElem,T) + j*randn(noElem,T));
elseif (strcmp(noiseType,'rndn'))
  M = size(theta,1);
  % Compute the square root of the source covariance matrix with SVD
  [U,SIG,V] = svd(noiseParam);
  % Generate i.i.d random variables with variance two.
  n = randn(noElem,T) + i*randn(noElem,T);
  % Color the i.i.d random variables with the square root of the
  % noise covariance matrix.
  Nt = 1/sqrt(2)*U*sqrt(SIG)*n;
elseif (strcmp(noiseType,'nonoise'))
  Nt=0;
else
  error('DBT-Error: The noise model does not exist.')
end%if


% ----------------------------------------------------------------------- %
% Compose the simulated signals.
% ----------------------------------------------------------------------- %
% Mottagna signaler fr錸 antennelementen.
% X(m,n) inneh錶ler mottagna signaler fr錸 antennelementen.
% m 鋜 antennelementindex. n 鋜 tidsindex.
% ----------------------------------------------------------------------- %

A = spastemat(antenna, theta', lambda, dist);
 % Not all input parameters of spastemat is used.
 % Pointing directions of subarray and antenna errors are not specified.
 % This is an error.
X=A*B;
  % If the noise power is 1 and the signal power is the value of the input
  % parameter SNR, the formula X=A*B gives the expected signal-to-noise ratio
  % (=SNR) only if the the absolute value of all elements in A is 1. This
  % may be true for an ULA with isotropical elements but it is not true in
  % general.
if (~all(all(abs(abs(A) - 1) < eps)))
  disp('Warning: In compsim4, the absolute value of some elements in A is not 1.');
  disp('This means that the signal-to-noise ratio may not be the expected.');
end

X=X+Nt;


% ----------------------------------------------------------------------- %
% Return data.
% ----------------------------------------------------------------------- %

signals = zerosm([size(X,2) 1 size(X,1) 1 1]); % Only one trial is implemented.
signals = setm(signals, X.', ':',1,':',1,1);

if (~(isstruct(waveform)))
  waveform = defwave(lambda,1,noSamp,[],[]);
end%if

new = 1;
if (new)
  sigvar = RxRadarSigT(signals, {antenna, waveform});
else
  sigvar.dataType = 'RxRadarSigT';
  sigvar.signals = signals;

  sigvar.antenna = antenna;
  sigvar.waveform = waveform;
end%if

tgtSig = B;


% End Of File