Mercurial > hg > ltpda

comparison m-toolbox/classes/@ssm/PSD.m @ 0:f0afece42f48

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Import.
author Daniele Nicolodi <nicolodi@science.unitn.it>
date Wed, 23 Nov 2011 19:22:13 +0100
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:f0afece42f48
1 % PSD computes the output theoretical CPSD shape with given inputs.
2 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3 %
4 % DESCRIPTION: PSD computes the output theoretical CPSD shape with given inputs.
5 % Unlike CPSD, it returns individual contributions and takes
6 % input vectors of objects (instead of square matrices)
7 %
8 % CALL: [mat_outSum, mat_out] = PSD(sys, pl)
9 %
10 % INPUTS:
11 % sys - (array of) ssm object
12 %
13 % OUTPUTS:
14 % mat_outSummed - contains specified returned aos, noise is
15 % summed over all the specified input noises
16 % mat_out - contains specified returned aos
17 %
18 % <a href="matlab:utils.helper.displayMethodInfo('ssm', 'PSD')">Parameters Description</a>
19 %
20 % VERSION: $Id: PSD.m,v 1.15 2011/04/17 21:28:05 adrien Exp $
21 %
22 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23
24 function varargout = PSD(varargin)
25
26 %% starting initial checks
27
28 % use the caller is method flag
29 callerIsMethod = utils.helper.callerIsMethod;
30
31 % Check if this is a call for parameters
32 if utils.helper.isinfocall(varargin{:})
33 varargout{1} = getInfo(varargin{3});
34 return
35 end
36
37 utils.helper.msg(utils.const.msg.MNAME, ['running ', mfilename]);
38
39 % Collect input variable names
40 in_names = cell(size(varargin));
41 for ii = 1:nargin,in_names{ii} = inputname(ii);end
42
43 % Collect all SSMs and plists
44 [sys, ssm_invars, rest] = utils.helper.collect_objects(varargin(:), 'ssm', in_names);
45 [pl, invars2, rest] = utils.helper.collect_objects(rest(:), 'plist');
46 if ~isempty(rest)
47 pl = combine(pl, plist(rest{:}));
48 end
49 pl = combine(pl, getDefaultPlist());
50
51 %%% Internal call: Only one object + don't look for a plist
52 internal = strcmp(varargin{end}, 'internal');
53
54 %% begin function body
55
56 %% retrieve system infos
57
58 if numel(sys)~=1
59 error('noisespectrum needs exactly one ssm as an input')
60 end
61 if ~sys.isnumerical
62 error(['error because system ',sys.name,' is not numerical']);
63 end
64 if ~sys.isStable
65 error('input ssm is not stable!')
66 end
67 if sys.timestep==0
68 timestep = 1;
69 else
70 timestep = sys.timestep;
71 end
72 if ~internal
73 inhist = sys.hist;
74 end
75
76 %% modifying system's ordering
77 if find(pl, 'reorganize')
78 sys = reorganize(sys, pl, 'set', 'for psd', 'internal', 'internal');
79 end
80
81 %% collecting functions i/o data
82 aos_in = find(pl, 'aos');
83 PZ_in = find(pl, 'PZmodels');
84 cov_in = find(pl, 'variance');
85 cpsd_in = find(pl, 'PSD');
86 noise_mat = [cov_in ; cpsd_in/(timestep*2)];
87
88 %% getting system's i/o sizes
89 inputSizes = sys.inputsizes;
90 outputSizes = sys.outputsizes; %#ok<NASGU>
91
92 Naos_in = inputSizes(1);
93 Nnoise = inputSizes(2);
94 NPZmodels = inputSizes(3);
95
96 %% retrieving frequency vector
97 if isempty(Naos_in)==0
98 f1 = find(pl,'f1');
99 f2 = find(pl,'f2');
100 NFreqs = find(pl,'nf');
101 if isempty(f1) || isempty(f2)|| isempty(NFreqs)
102 error('### Please specify frequency vector a start and stop frequency .');
103 else
104 freqs = 10.^linspace(log10(f1), log10(f2), NFreqs);
105 end
106 else
107 freqs = aos_in(1).x;
108 end
109
110 %% checking frequency vector
111 for i=2:numel(aos_in)
112 if ~isequal(freqs,aos_in(i).x)
113 error('there exist different frequency vectors');
114 end
115 end
116
117 %% reshape pzmodels and aos for input cross-spectra
118 if numel(PZ_in)==NPZmodels
119 PZdata = zeros(NPZmodels,NFreqs);
120 for i=1:NPZmodels
121 a = resp(PZ_in(i), freqs);
122 PZdata(i,:) = reshape(a.y,[1,NFreqs]) ;
123 end
124 else
125 error('Wrong size for field PZ_in')
126 end
127
128 if numel(aos_in)==Naos_in
129 AOdata = zeros(Naos_in,NFreqs);
130 for i=1:Naos_in
131 AOdata(i,:) = reshape(aos_in(i).y,[1,NFreqs]) ;
132 end
133 else
134 error('Wrong size for field aos_in')
135 end
136
137 %% SSM Transfer function
138 [a, b, c, d, Ts, InputName, StateName, OutputName,...
139 inputvarunits, ssvarunits, outputvarunits] = double(sys); %#ok<ASGLU>
140 resps = ssm.doBode(a, b, c, d, 2*pi*freqs, Ts);
141 Noutputs = numel(OutputName);
142
143 %% power for each frequency with SVD computation
144 Result = zeros(Noutputs, Nnoise+Naos_in+NPZmodels, NFreqs);
145
146 for ff=1:NFreqs
147 for ii = 1:(Nnoise+Naos_in+NPZmodels)
148
149 AmpWhiteNoise = zeros(1,Nnoise);
150 AmpAO = zeros(1, Naos_in);
151 AmpPZ = zeros(1,NPZmodels);
152 if ii<Nnoise+1,
153 %% contribution from white noise
154 if noise_mat(ii)<0
155 error('input PSD is not positive!')
156 end
157 AmpWhiteNoise(ii) = noise_mat(ii)^0.5;
158 elseif ii<Nnoise+Naos_in+1
159 %% contribution from aos
160 i_input2 = ii-Nnoise;
161 if AOdata(i_input2,ff)<0
162 error('input PSD is not positive!')
163 end
164 AmpAO(i_input2) = AOdata(i_input2,ff)^0.5;
165 else
166 %% contribution from PZmodels
167 i_input2 = ii-Nnoise-Naos_in;
168 if PZdata(i_input2,ff)<0
169 error('input PSD is not positive!')
170 end
171 AmpPZ(i_input2) = PZdata(i_input2,ff)^0.5;
172 end
173 %% computing CPSD
174 Amp = diag([AmpAO; AmpWhiteNoise; AmpPZ]);
175 RespLoc = squeeze(resps(:,:,ff));
176 noise = RespLoc * Amp * (RespLoc*Amp)';
177 Result(:,ii,ff) = diag( real(noise) * (2*timestep) ); % 2 correction added here
178 end
179 end
180
181 %% saving in aos
182 if nargout ~= 1;
183 ao_out = ao.initObjectWithSize(Noutputs, Nnoise+Naos_in+NPZmodels);
184 end
185 ao_outSum = ao.initObjectWithSize(Noutputs, 1);
186
187 for oo=1:Noutputs
188 %% individual inputs
189 if nargout ~= 1;
190 for ii=1:(Nnoise+Naos_in+NPZmodels)
191 ao_out(oo,ii).setData(fsdata(freqs, squeeze(Result(oo,ii,:))));
192 ao_out(oo,ii).setName( ['PSD of ' , OutputName{oo} ' due to ' InputName{ii}]);
193 ao_out(oo,ii).setXunits('Hz');
194 ao_out(oo,ii).setYunits(outputvarunits(oo)^2/unit('Hz'));
195 ao_out(oo,ii).setDescription( ['PSD of ' , OutputName{oo} ' due to ' InputName{ii}]);
196 end
197 end
198
199 %% sum of all inputs
200 ao_outSum(oo,1).setData(fsdata(freqs, sum(squeeze(Result(oo,:,:)),1) ));
201 ao_outSum(oo,1).setName( ['PSD of ' , OutputName{oo} ' due to all contributions']);
202 ao_outSum(oo,1).setXunits('Hz');
203 ao_outSum(oo,1).setYunits(outputvarunits(oo)^2/unit('Hz'));
204 ao_outSum(oo,1).setDescription( ['PSD of ' , OutputName{oo} ' due to all contributions']);
205
206 end
207
208 %% construct output matrix object
209 if nargout ~= 1;
210 out = matrix(ao_out);
211 end
212 outSum = matrix(ao_outSum);
213 if callerIsMethod
214 % do nothing
215 else
216 myinfo = getInfo('None');
217 if nargout ~= 1;
218 out.addHistory(myinfo, pl , ssm_invars(1), inhist );
219 end
220 outSum.addHistory(myinfo, pl , ssm_invars(1), inhist );
221 end
222
223 %% Set output depending on nargout
224 if nargout == 1;
225 varargout = {outSum};
226 elseif nargout == 2;
227 varargout = {outSum out };
228 elseif nargout == 0;
229 iplot(ao_outSum, ao_out);
230 else
231 error('Wrong number of outputs')
232 end
233 end
234
235
236
237
238 %--------------------------------------------------------------------------
239 % Get Info Object
240 %--------------------------------------------------------------------------
241 function ii = getInfo(varargin)
242
243 if nargin == 1 && strcmpi(varargin{1}, 'None')
244 sets = {};
245 pl = [];
246 else
247 sets = {'Default'};
248 pl = getDefaultPlist;
249 end
250 % Build info object
251 ii = minfo(mfilename, 'ssm', 'ltpda', utils.const.categories.op, '$Id: PSD.m,v 1.15 2011/04/17 21:28:05 adrien Exp $', sets, pl);
252
253 end
254
255 %--------------------------------------------------------------------------
256 % Get Default Plist
257 %--------------------------------------------------------------------------
258 function pl = getDefaultPlist()
259 pl = ssm.getInfo('reorganize', 'for PSD').plists;
260 pl.remove('set');
261
262 p = param({'variance', 'The variance vector of this noise between input ports for the <i>time-discrete</i> noise model. '}, []);
263 pl.append(p);
264
265 p = param({'PSD', 'The one sided psd vector of the white noise between input ports.'}, []);
266 pl.append(p);
267
268 p = param({'aos', 'A vector of input PSD AOs, The spectrum of this noise between input ports for the <i>time-continuous</i> noise model.'}, ao.initObjectWithSize(1,0));
269 pl.append(p);
270
271 p = param({'PZmodels', 'vector of noise shape filters for the different corresponding inputs.'}, paramValue.DOUBLE_VALUE(zeros(0,1)));
272 pl.append(p);
273
274 p = param({'reorganize', 'When set to 0, this means the ssm does not need be modified to match the requested i/o. Faster but dangerous!'}, paramValue.TRUE_FALSE);
275 pl.append(p);
276
277 p = param({'f2', 'The maximum frequency. Default is Nyquist or 1Hz.'}, paramValue.EMPTY_DOUBLE);
278 pl.append(p);
279
280 p = param({'f1', 'The minimum frequency. Default is f2*1e-5.'}, paramValue.EMPTY_DOUBLE);
281 pl.append(p);
282
283 p = param({'nf', 'The number of frequency bins. Frequencies are scale logarithmically'}, paramValue.DOUBLE_VALUE(200));
284 pl.append(p);
285
286 end
287