demo_trajGMM01.m 8.79 KB
Newer Older
Sylvain Calinon's avatar
Sylvain Calinon committed
1
function demo_trajGMM01
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
% Trajectory synthesis with a GMM with dynamic features (trajectory GMM).
%
% Writing code takes time. Polishing it and making it available to others takes longer! 
% If some parts of the code were useful for your research of for a better understanding 
% of the algorithms, please reward the authors by citing the related publications, 
% and consider making your own research available in this way.
%
% @article{Calinon15,
%   author="Calinon, S.",
%   title="A Tutorial on Task-Parameterized Movement Learning and Retrieval",
%   journal="Intelligent Service Robotics",
%   year="2015"
% }
% 
% Copyright (c) 2015 Idiap Research Institute, http://idiap.ch/
% Written by Sylvain Calinon, http://calinon.ch/
% 
% This file is part of PbDlib, http://www.idiap.ch/software/pbdlib/
% 
% PbDlib is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License version 3 as
% published by the Free Software Foundation.
% 
% PbDlib 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 PbDlib. If not, see <http://www.gnu.org/licenses/>.

Sylvain Calinon's avatar
Sylvain Calinon committed
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

addpath('./m_fcts/');

%% Parameters
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
model.nbStates = 5; %Number of components in the GMM
model.nbVarPos = 2; %Dimension of position data (here: x1,x2)
model.nbDeriv = 3; %Number of static&dynamic features (D=2 for [x,dx], D=3 for [x,dx,ddx], etc.)
model.nbVar = model.nbVarPos * model.nbDeriv; %Dimension of state vector
model.dt = 0.01; %Time step (without rescaling, large values such as 1 has the advantage of creating clusers based on position information)
nbSamples = 4; %Number of demonstrations
nbData = 100; %Number of datapoints in a trajectory

[PHI,PHI1] = constructPHI(model,nbData,nbSamples); %Construct PHI operator (big sparse matrix)


%% Load dataset
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% %Load different types of movements
% load('data/dataTrajGMM01.mat');
% %load('data/dataTrajGMM02.mat'); %periodic motion
% %load('data/dataTrajGMM03.mat'); %motion with options
% %load('data/dataTrajGMM04.mat'); %partial demonstrations
% 
% %Resampling of dataset
% nbData0 = size(Data,2)/nbSamples; %Original number of datapoints in a trajectory
% DataTmp = [];
% for n=1:nbSamples
% 	DataTmp = [DataTmp spline(1:nbData0, Data(2:model.nbVarPos+1,(n-1)*nbData0+1:n*nbData0), linspace(1,nbData0,nbData))]; 
% end
% Data = DataTmp;

65 66
%Load handwriting movements
load('data/2Dletters/S.mat');
Sylvain Calinon's avatar
Sylvain Calinon committed
67 68 69 70 71 72 73 74
Data=[];
for n=1:nbSamples
	s(n).Data = spline(1:size(demos{n}.pos,2), demos{n}.pos, linspace(1,size(demos{n}.pos,2),nbData)); %Resampling
	Data = [Data s(n).Data]; 
end

%Re-arrange data in vector form
x = reshape(Data, model.nbVarPos*nbData*nbSamples, 1) * 1E2; %Scale data to avoid numerical computation problem
75
zeta = PHI*x; %y is for example [x1(1), x2(1), x1d(1), x2d(1), x1(2), x2(2), x1d(2), x2d(2), ...]
Sylvain Calinon's avatar
Sylvain Calinon committed
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
Data = reshape(zeta, model.nbVarPos*model.nbDeriv, nbData*nbSamples); %Include derivatives in Data


%% Parameters estimation
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% %Initialization with kmeans
% model = init_GMM_kmeans(Data, model);

%Initialization with homogeneous time intervals
model = init_GMM_timeBased([repmat(1:nbData,1,nbSamples); Data], model);
model.Mu = model.Mu(2:end,:);
model.Sigma = model.Sigma(2:end,2:end,:);

%Refinement of parameters
[model, GAMMA2] = EM_GMM(Data, model);
%[model, GAMMA2] = EM_MFA(Data, model);
%[model, GAMMA2] = EM_MPPCA(Data, model);
%[model, GAMMA2] = EM_HDGMM(Data, model);

%Precomputation of inverses (optional)
for i=1:model.nbStates
	model.invSigma(:,:,i) = inv(model.Sigma(:,:,i));
end


%% Reproduction
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
for n=1:1 %nbSamples
	%Compute best path for the n-th demonstration
	[~,r(n).q] = max(GAMMA2(:,(n-1)*nbData+1:n*nbData),[],1); %works also for nbStates=1
	
107
	%Create single Gaussian N(MuQ,SigmaQ) based on optimal state sequence q
Sylvain Calinon's avatar
Sylvain Calinon committed
108 109 110 111 112 113 114 115 116 117 118 119
	MuQ = reshape(model.Mu(:,r(n).q), model.nbVar*nbData, 1); 
	%MuQ = zeros(model.nbVar*nbData,1);
	SigmaQ = zeros(model.nbVar*nbData);
	for t=1:nbData
		id = (t-1)*model.nbVar+1:t*model.nbVar;
		%MuQ(id) = model.Mu(:,r(n).q(t)); 
		SigmaQ(id,id) = model.Sigma(:,:,r(n).q(t));
	end
	%SigmaQ can alternatively be computed with:
	%SigmaQ = (kron(ones(nbData,1), eye(model.nbVar)) * reshape(model.Sigma(:,:,r(n).q), model.nbVar, model.nbVar*nbData)) .* kron(eye(nbData), ones(model.nbVar));


120 121 122 123
	%Least squares computation method 1 (using lscov Matlab function)
	%%%%%%%%%%%%%%%%%%%
	[zeta,~,mse,S] = lscov(PHI1, MuQ, SigmaQ, 'chol'); %Retrieval of data with weighted least squares solution
	r(n).Data = reshape(zeta, model.nbVarPos, nbData); %Reshape data for plotting
Sylvain Calinon's avatar
Sylvain Calinon committed
124 125

	
126
% 	%Least squares computation method 2 (most readable but not optimized)
Sylvain Calinon's avatar
Sylvain Calinon committed
127 128 129 130 131 132 133
% 	%%%%%%%%%%%%%%%%%%%
% 	PHIinvSigmaQ = PHI1' / SigmaQ;
% 	Rq = PHIinvSigmaQ * PHI1;
% 	rq = PHIinvSigmaQ * MuQ;
% 	zeta = Rq \ rq; %Can also be computed with c = lscov(Rq, rq)
% 	size(zeta)
% 	r(n).Data = reshape(zeta, model.nbVarPos, nbData); %Reshape data for plotting
134
% 	%Covariance Matrix computation of ordinary least squares estimate
Sylvain Calinon's avatar
Sylvain Calinon committed
135 136 137 138
% 	mse =  (MuQ'*inv(SigmaQ)*MuQ - rq'*inv(Rq)*rq) ./ ((model.nbVar-model.nbVarPos)*nbData);
% 	S = inv(Rq) * mse; 
	

139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
% 	%Least squares computation method 3 (efficient computation using Cholesky and QR decompositions, inspired by lscov code)
% 	%%%%%%%%%%%%%%%%%%%
% 	T = chol(SigmaQ)'; %SigmaQ=T*T'
% 	TA = T \ PHI1;
% 	TMuQ = T \ MuQ;
% 	[Q, R, perm] = qr(TA,0); %PHI1(:,perm)=Q*R
% 	z = Q' * TMuQ;
% 	zeta = zeros(nbData*model.nbVarPos,1);
% 	zeta(perm,:) = R \ z; %zeta=(TA'*TA)\(TA'*TMuQ)
% 	r(n).Data = reshape(zeta, model.nbVarPos, nbData); %Reshape data for plotting
% 	%Covariance Matrix computation of ordinary least squares estimate
% 	err = TMuQ - Q*z;
% 	mse = err'*err ./ (model.nbVar*nbData - model.nbVarPos*nbData);
% 	Rinv = R \ eye(model.nbVarPos*nbData);
% 	S(perm,perm) = Rinv*Rinv' .* mse; 
Sylvain Calinon's avatar
Sylvain Calinon committed
154 155
	
	
156
	%Rebuild covariance by reshaping S
Sylvain Calinon's avatar
Sylvain Calinon committed
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
	for t=1:nbData
		id = (t-1)*model.nbVarPos+1:t*model.nbVarPos;
		r(n).expSigma(:,:,t) = S(id,id) * nbData;
	end
	
end %nbSamples


%% Plot timeline
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
figure('position',[10 10 600 600]);
for m=1:model.nbVarPos
	limAxes = [1, nbData, min(Data(m,:))-1E0, max(Data(m,:))+1E0];
	subplot(model.nbVarPos,1,m); hold on;
	for n=1:1 %nbSamples
		msh=[]; x0=[];
		for t=1:nbData-1
			if size(msh,2)==0
				msh(:,1) = [t; model.Mu(m,r(n).q(t))];
			end
			if t==nbData-1 || r(n).q(t+1)~=r(n).q(t)
				msh(:,2) = [t+1; model.Mu(m,r(n).q(t))];
				sTmp = model.Sigma(m,m,r(n).q(t))^.5;
				msh2 = [msh(:,1)+[0;sTmp], msh(:,2)+[0;sTmp], msh(:,2)-[0;sTmp], msh(:,1)-[0;sTmp], msh(:,1)+[0;sTmp]];
				patch(msh2(1,:), msh2(2,:), [.85 .85 .85],'edgecolor',[.7 .7 .7]);
				plot(msh(1,:), msh(2,:), '-','linewidth',3,'color',[.7 .7 .7]);
				plot([msh(1,1) msh(1,1)], limAxes(3:4), ':','linewidth',1,'color',[.7 .7 .7]);
				x0 = [x0 msh];
				msh=[];
			end
		end
		msh = [1:nbData, nbData:-1:1; r(n).Data(m,:)-squeeze(r(n).expSigma(m,m,:).^.5)'*1, fliplr(r(n).Data(m,:)+squeeze(r(n).expSigma(m,m,:).^.5)'*1)];
		patch(msh(1,:), msh(2,:), [1 .4 .4],'edgecolor',[1 .2 .2],'edgealpha',.8,'facealpha',.5);
	end
	for n=1:nbSamples
		plot(1:nbData, Data(m,(n-1)*nbData+1:n*nbData), '-','lineWidth',1,'color',[.2 .2 .2]);
	end
	for n=1:1
		plot(1:nbData, r(n).Data(m,:), '-','lineWidth',2.5,'color',[.8 0 0]);
	end
	
	set(gca,'xtick',[],'ytick',[]);
	xlabel('$t$','interpreter','latex','fontsize',18);
	ylabel(['$x_' num2str(m) '$'],'interpreter','latex','fontsize',18);
	axis(limAxes);
end


%% Plot 2D
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
if model.nbVarPos>1
	figure('position',[620 10 600 600]); hold on;
	for n=1:1 %nbSamples
		plotGMM(r(n).Data([1,2],:), r(n).expSigma([1,2],[1,2],:), [1 .2 .2],.2);
	end
	plotGMM(model.Mu([1,2],:), model.Sigma([1,2],[1,2],:), [.5 .5 .5],.8);
	for n=1:nbSamples
		plot(Data(1,(n-1)*nbData+1:n*nbData), Data(2,(n-1)*nbData+1:n*nbData), '-','lineWidth',1,'color',[.2 .2 .2]); %-0.2+0.8*(n-1)/(nbSamples-1)
	end
	for n=1:1
		plot(r(n).Data(1,:), r(n).Data(2,:), '-','lineWidth',2.5,'color',[.8 0 0]);
	end
	set(gca,'xtick',[],'ytick',[]); axis equal; axis square;
	xlabel(['$x_1$'],'interpreter','latex','fontsize',18);
	ylabel(['$x_2$'],'interpreter','latex','fontsize',18);
end

%print('-dpng','graphs/demo_trajGMM01.png');
%pause;
%close all;