Mercurial > hg > ltpda
view m-toolbox/classes/@ao/whiten2D.m @ 52:daf4eab1a51e database-connection-manager tip
Fix. Default password should be [] not an empty string
author | Daniele Nicolodi <nicolodi@science.unitn.it> |
---|---|
date | Wed, 07 Dec 2011 17:29:47 +0100 |
parents | f0afece42f48 |
children |
line wrap: on
line source
% WHITEN2D whiten the noise for two cross correlated time series. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % % DESCRIPTION: whiten2D whitens cross-correlated time-series. Whitening % filters are constructed by a fitting procedure to the cross-spectrum % models provided. % Note: The function assumes that the input model corresponds to the % one-sided csd of the data to be whitened. % % ALGORITHM: % 1) Fit a set of partial fraction z-domain filters using % utils.math.psd2wf % 2) Convert to bank of mIIR filters % 3) Filter time-series in parallel % The filtering process is: % b(1) = Filt11(a(1)) + Filt12(a(2)) % b(2) = Filt21(a(1)) + Filt22(a(2)) % % CALL: b = whiten2D(a, pl) % returns whitened time-series AOs % [b1,b2] = whiten2D(a1, a2, pl) % [b1,b2,...,bn] = whiten2D(a1,a2,...,an, pl); % Note: Input AOs must come in couples. % Note: this method cannot be used as a modifier, the % call a.whiten2D(pl) is forbidden. % % INPUT: % % - a is a couple of two colored noise time-series AOs % % OUTPUT: % % - b is a couple of "whitened" time-series AOs. The whitening % filters used are stored in the objects procinfo field under % the parameters: % - b(1): 'Filt11' and 'Filt12' % - b(2): 'Filt21' and 'Filt22' % % <a href="matlab:utils.helper.displayMethodInfo('ao', 'whiten2D')">Parameters Description</a> % % VERSION: $Id: whiten2D.m,v 1.35 2011/04/08 08:56:16 hewitson Exp $ % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function varargout = whiten2D(varargin) % Check if this is a call for parameters if utils.helper.isinfocall(varargin{:}) varargout{1} = getInfo(varargin{3}); return end import utils.const.* utils.helper.msg(msg.PROC3, 'running %s/%s', mfilename('class'), mfilename); % Collect input variable names in_names = cell(size(varargin)); for ii = 1:nargin,in_names{ii} = inputname(ii);end % Collect all AOs and plists [as, ao_invars] = utils.helper.collect_objects(varargin(:), 'ao', in_names); pl = utils.helper.collect_objects(varargin(:), 'plist', in_names); % Decide on a deep copy or a modify bs = copy(as, nargout); cs = copy(as, nargout); inhists = [as.hist]; % combine plists pl = parse(pl, getDefaultPlist()); pl.getSetRandState(); % Extract necessary model parameters csd11 = find(pl, 'csd11'); csd12 = find(pl, 'csd12'); csd21 = find(pl, 'csd21'); csd22 = find(pl, 'csd22'); if nargout == 0 error('### noisegen2D cannot be used as a modifier. Please give an output variable.'); end % Check the number of input AO if numel(bs)==1 error('!!! One input AO! The input AOs must come in pairs!'); end % Checks that the input AOs come in pairs odc = 0; if rem(numel(as),2) warning('The input AOs must come in pairs! Skipping AO %s during calculation', ao_invars{end}); odc = 1; end % Loop over input AOs to check for non time series objects fsv = zeros(numel(bs),1); for jj=1:numel(bs) if ~isa(bs(jj).data, 'tsdata') error('!!! %s expects ao/tsdata objects. ', mfilename); end fsv(jj,1) = bs(jj).data.fs; % collecting sampling frequencies end % Check that input Aos have the same sampling frequency if any(diff(fsv)) error('!!! Sampling frequency must be the same for all input objects') end % ------------------- Coloring Noise % ---------------------------------- % 1) - Fitting the models to identify the innovation filters % Build input structure for psd2tf params = struct(); params.idtp = 1; params.Nmaxiter = find(pl, 'MaxIter'); params.minorder = find(pl, 'MinOrder'); params.maxorder = find(pl, 'MaxOrder'); params.spolesopt = find(pl, 'PoleType'); params.weightparam = find(pl, 'Weights'); params.usesym = find(pl, 'UseSym'); params.spy = find(pl, 'Disp'); params.keepvar = find(pl, 'keepvar'); % Tolerance for MSE Value lrscond = find(pl, 'FITTOL'); % give an error for strange values of lrscond if lrscond<0 error('!!! Negative values for FITTOL are not allowed !!!') end % handling data lrscond = -1*log10(lrscond); % give a warning for strange values of lrscond if lrscond<0 warning('You are searching for a MSE lower than %s', num2str(10^(-1*lrscond))) end params.lrscond = lrscond; % Tolerance for the MSE relative variation msevar = find(pl, 'MSEVARTOL'); % handling data msevar = -1*log10(msevar); % give a warning for strange values of msevar if msevar<0 warning('You are searching for MSE relative variation lower than %s', num2str(10^(-1*msevar))) end params.msevar = msevar; if isempty(params.msevar) params.ctp = 'chival'; else params.ctp = 'chivar'; end if(find(pl, 'plot')) params.plot = 1; else params.plot = 0; end params.fs = fsv(1,1); params.dterm = 0; % get data variance vars = [1 1]; if numel(bs)==2 for ii=1:numel(bs) b = bs(ii).y; v = var(b); vars(ii) = v; end end params.vars = vars; % call psd2wf ostruct = utils.math.psd2wf(csd11.y,csd12.y,csd21.y,csd22.y,csd11.x,params); % ---------------------------------- % 2) - Convert into MIIR filters fs = fsv(1,1); % --- filter 1 --- res = ostruct(1).res; poles = ostruct(1).poles; dterm = ostruct(1).dterm; % construct a struct array of miir filters vectors pfilts1 = []; for kk=1:numel(res) ft = miir(res(kk), [ 1 -poles(kk)], fs); pfilts1 = [pfilts1 ft]; end % pfilts1 = [pfilts1 miir(dterm, [1 0], fs)]; % --- filter 2 --- res = ostruct(2).res; poles = ostruct(2).poles; dterm = ostruct(2).dterm; % construct a struct array of miir filters vectors pfilts2 = []; for kk=1:numel(res) ft = miir(res(kk), [ 1 -poles(kk)], fs); pfilts2 = [pfilts2 ft]; end % pfilts2 = [pfilts2 miir(dterm, [1 0], fs)]; % --- filter 3 --- res = ostruct(3).res; poles = ostruct(3).poles; dterm = ostruct(3).dterm; % construct a struct array of miir filters vectors pfilts3 = []; for kk=1:numel(res) ft = miir(res(kk), [ 1 -poles(kk)], fs); pfilts3 = [pfilts3 ft]; end % pfilts3 = [pfilts3 miir(dterm, [1 0], fs)]; % --- filter 4 --- res = ostruct(4).res; poles = ostruct(4).poles; dterm = ostruct(4).dterm; % construct a struct array of miir filters vectors pfilts4 = []; for kk=1:numel(res) ft = miir(res(kk), [ 1 -poles(kk)], fs); pfilts4 = [pfilts4 ft]; end % pfilts4 = [pfilts4 miir(dterm, [1 0], fs)]; % ---------------------------------- % 3) Filtering data for jj = 1:2:numel(bs)-1-odc bs(jj) = filter(cs(jj), pfilts1) + filter(cs(jj+1), pfilts2); bs(jj+1) = filter(cs(jj), pfilts3) + filter(cs(jj+1), pfilts4); % ----------------------------------- % 4) Output data % name for this object bs(jj).name = sprintf('whiten2D(%s)_c1', [ao_invars{jj} ao_invars{jj+1}]); % Collect the filters into procinfo bs(jj).procinfo = combine(plist('Filt11', pfilts1,'Filt12', pfilts2),as(jj).procinfo); % add history bs(jj).addHistory(getInfo('None'), pl, [ao_invars(:)], [inhists(:)]); % name for this object bs(jj+1).name = sprintf('whiten2D(%s)_c2', [ao_invars{jj} ao_invars{jj+1}]); % Collect the filters into procinfo bs(jj+1).procinfo = combine(plist('Filt21', pfilts3,'Filt22', pfilts4),as(jj+1).procinfo); % add history bs(jj+1).addHistory(getInfo('None'), pl, [ao_invars(:)], [inhists(:)]); end % clear errors bs.clearErrors; % Set output if nargout == numel(bs) % List of outputs for ii = 1:numel(bs) varargout{ii} = bs.index(ii); end else % Single output varargout{1} = bs; end end %-------------------------------------------------------------------------- % Get Info Object %-------------------------------------------------------------------------- function ii = getInfo(varargin) if nargin == 1 && strcmpi(varargin{1}, 'None') sets = {}; pl = []; else sets = {'Default'}; pl = getDefaultPlist; end % Build info object ii = minfo(mfilename, 'ao', 'ltpda', utils.const.categories.sigproc, '$Id: whiten2D.m,v 1.35 2011/04/08 08:56:16 hewitson Exp $', sets, pl); ii.setArgsmin(2); ii.setOutmin(2); ii.setModifier(false); end %-------------------------------------------------------------------------- % Get Default Plist %-------------------------------------------------------------------------- function plout = getDefaultPlist() persistent pl; if exist('pl', 'var')==0 || isempty(pl) pl = buildplist(); end plout = pl; end function pl = buildplist() pl = plist(); % CSD11 p = param({'csd11', 'A frequency-series AO describing the model csd11'}, paramValue.EMPTY_DOUBLE); pl.append(p); % CSD12 p = param({'csd12', 'A frequency-series AO describing the model csd12'}, paramValue.EMPTY_DOUBLE); pl.append(p); % CSD21 p = param({'csd21', 'A frequency-series AO describing the model csd21'}, paramValue.EMPTY_DOUBLE); pl.append(p); % CSD22 p = param({'csd22', 'A frequency-series AO describing the model csd22'}, paramValue.EMPTY_DOUBLE); pl.append(p); % MaxIter p = param({'MaxIter', 'Maximum number of iterations in fit routine.'}, paramValue.DOUBLE_VALUE(30)); pl.append(p); % PoleType p = param({'PoleType', ['Choose the pole type for fitting:<ol>'... '<li>use real starting poles</li>' ... '<li>generates complex conjugate poles of the<br>'... 'type <tt>a.*exp(theta*pi*j)</tt><br>'... 'with <tt>theta = linspace(0,pi,N/2+1)</tt></li>'... '<li>generates complex conjugate poles of the type<br>'... '<tt>a.*exp(theta*pi*j)</tt><br>'... 'with <tt>theta = linspace(0,pi,N/2+2)</tt></li></ol>']}, {3, {1,2,3}, paramValue.SINGLE}); pl.append(p); % MinOrder p = param({'MinOrder', 'Minimum order to fit with.'}, paramValue.DOUBLE_VALUE(2)); pl.append(p); % MaxOrder p = param({'MaxOrder', 'Maximum order to fit with.'}, paramValue.DOUBLE_VALUE(25)); pl.append(p); % Weights p = param({'Weights', ['Choose weighting for the fit:<ol>'... '<li>equal weights for each point</li>'... '<li>weight with <tt>1/abs(model)</tt></li>'... '<li>weight with <tt>1/abs(model).^2</tt></li>'... '<li>weight with inverse of the square mean spread<br>'... 'of the model</li></ol>']}, paramValue.DOUBLE_VALUE(3)); pl.append(p); % Plot p = param({'Plot', 'Plot results of each fitting step.'}, paramValue.FALSE_TRUE); pl.append(p); % Disp p = param({'Disp', 'Display the progress of the fitting iteration.'}, paramValue.FALSE_TRUE); pl.append(p); % MSEVARTOL p = param({'MSEVARTOL', ['Mean Squared Error Variation - Check if the<br>'... 'realtive variation of the mean squared error is<br>'... 'smaller than the value specified. This<br>'... 'option is useful for finding the minimum of Chi-squared.']}, ... paramValue.DOUBLE_VALUE(1e-2)); pl.append(p); % FITTOL p = param({'FITTOL', ['Mean Squared Error Value - Check if the mean<br>'... 'squared error value is lower than the value<br>'... 'specified.']}, paramValue.DOUBLE_VALUE(1e-2)); pl.append(p); % UseSym p = param({'UseSym', ['Use symbolic calculation in eigen-decomposition.<ul>'... '<li>0 - perform double-precision calculation in the<br>'... 'eigendecomposition procedure to identify 2-Dim<br>'... 'systems and for poles stabilization</li>'... '<li>1 - uses symbolic math toolbox variable precision<br>'... 'arithmetic in the eigen-decomposition for 2-Dim<br>'... 'system identification and double-precison for<br>'... 'poles stabilization</li>'... '<li>2 - uses symbolic math toolbox variable precision<br>'... 'arithmetic in the eigen-decomposition for 2-Dim<br>'... 'system identification and for poles stabilization.']}, {1, {0,1,2}, paramValue.SINGLE}); pl.append(p); % Keep var p = param({'keepvar', '???'}, paramValue.TRUE_FALSE); p.val.setValIndex(2); pl.append(p); end % PARAMETERS: % % 'csd11' - a frequency-series AO describing the model csd11 % 'csd12' - a frequency-series AO describing the model csd12 % 'csd21' - a frequency-series AO describing the model csd21 % 'csd22' - a frequency-series AO describing the model csd22 % % 'MaxIter' - Maximum number of iterations in fit routine % [default: 30] % % 'PoleType' - Choose the pole type for fitting: % 1 - use real starting poles % 2 - generates complex conjugate poles of the % type a.*exp(theta*pi*j) % with theta = linspace(0,pi,N/2+1) % 3 - generates complex conjugate poles of the type % a.*exp(theta*pi*j) % with theta = linspace(0,pi,N/2+2) [default] % % 'MinOrder' - Minimum order to fit with. [default: 2] % % 'MaxOrder' - Maximum order to fit with. [default: 25] % % 'Weights' - choose weighting for the fit: [default: 2] % 1 - equal weights for each point % 2 - weight with 1/abs(model) % 3 - weight with 1/sqrt(abs(model)) % 4 - weight with inverse of the square mean spread % of the model % % 'Plot' - plot results of each fitting step. [default: false] % % 'Disp' - Display the progress of the fitting iteration. % [default: false] % % 'FITTOL' - Mean Squared Error Value - Check if the mean % squared error value is lower than the value % specified. [defalut: 1e-2] % % 'MSEVARTOL' - Mean Squared Error Variation - Check if the % realtive variation of the mean squared error is % smaller than the value specified. This % option is useful for finding the minimum of Chi % squared. [default: 1e-2]] % % 'UseSym' - Use symbolic calculation in eigendecomposition. % [default: 0] % 0 - perform double-precision calculation in the % eigendecomposition procedure to identify 2dim % systems and for poles stabilization % 1 - uses symbolic math toolbox variable precision % arithmetic in the eigendecomposition for 2dim % system identification and double-precison for % poles stabilization % 2 - uses symbolic math toolbox variable precision % arithmetic in the eigendecomposition for 2dim % system identification and for poles % stabilization