58 lines
4.1 KiB
Matlab
58 lines
4.1 KiB
Matlab
function results_CLOE = run_CLOE(robot, benchmarkSettings, experimentDataStruct, optionsCLOE, progressBar)
|
|
|
|
%% Define result data structure:
|
|
|
|
if benchmarkSettings.displayProgression == true
|
|
waitbar(0, progressBar, sprintf('CLOE: First Iteration...'));
|
|
end
|
|
|
|
results_CLOE.benchmarkSettings = benchmarkSettings; % Data structure containing the benchmark settings
|
|
results_CLOE.options = optionsCLOE; % Options that are specific to the identification method.
|
|
results_CLOE.Betas = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint, numel(benchmarkSettings.Beta_obj)); % Identified parameters
|
|
results_CLOE.times = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint); % Computation times
|
|
results_CLOE.flag = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint); % Termination flag of the identification code
|
|
results_CLOE.iterations = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint); % Number of iterations
|
|
results_CLOE.lambdas_upper = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint, robot.paramVectorSize);
|
|
results_CLOE.lambdas_lower = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint, robot.paramVectorSize);
|
|
% results_CLOE.jacobians = zeros(benchmarkSettings.numberOfInitialEstimates, benchmarkSettings.numberOfExperimentsPerInitialPoint, (benchmarkSettings.nbSamples-2*benchmarkSettings.samplingBorder)*robot.nbDOF/benchmarkSettings.decimRate, robot.paramVectorSize);
|
|
results_CLOE.jacobian_sigma_upper = zeros(benchmarkSettings.numberOfInitialEstimates); % Upper singular value of the Jacobian matrix.
|
|
results_CLOE.jacobian_sigma_lower = zeros(benchmarkSettings.numberOfInitialEstimates); % Lower singular value of the Jacobian matrix.
|
|
|
|
%% Run the identification code:
|
|
|
|
for initEst = 1:benchmarkSettings.numberOfInitialEstimates % For each set of initial parameters
|
|
|
|
Beta_0 = benchmarkSettings.Initial_Beta(:,initEst);
|
|
|
|
for expNb = 1:benchmarkSettings.numberOfExperimentsPerInitialPoint % For each trajectory noise
|
|
tic
|
|
[Beta_CLOE, fval, output, exitflag, lambda, jacobian] = CLOE_identification(robot, benchmarkSettings, experimentDataStruct, expNb, Beta_0, optionsCLOE);
|
|
results_CLOE.times(initEst, expNb)=toc;
|
|
|
|
if optionsCLOE.verbose == true
|
|
fprintf('CLOE status: initial estimate %d, experiment %d \n', initEst, expNb);
|
|
fprintf('CLOE status: initial parameter error = %d\n', norm(Beta_0-benchmarkSettings.Beta_obj));
|
|
fprintf('CLOE status: estimated parameter error = %d\n', norm(Beta_CLOE-benchmarkSettings.Beta_obj));
|
|
disp('---------------------------------------------')
|
|
end
|
|
|
|
if benchmarkSettings.displayProgression == true
|
|
waitbar(((initEst-1)*benchmarkSettings.numberOfExperimentsPerInitialPoint+expNb)/((benchmarkSettings.numberOfInitialEstimates-1)*benchmarkSettings.numberOfExperimentsPerInitialPoint+benchmarkSettings.numberOfExperimentsPerInitialPoint), progressBar, sprintf('CLOE: %d%% done...', floor(100*((initEst-1)*benchmarkSettings.numberOfExperimentsPerInitialPoint+expNb)/((benchmarkSettings.numberOfInitialEstimates-1)*benchmarkSettings.numberOfExperimentsPerInitialPoint+benchmarkSettings.numberOfExperimentsPerInitialPoint))));
|
|
end
|
|
results_CLOE.flag(initEst,expNb)=exitflag;
|
|
results_CLOE.Betas(initEst,expNb,:)=Beta_CLOE;
|
|
results_CLOE.eps(initEst, expNb)=norm(Beta_CLOE-benchmarkSettings.Beta_obj);
|
|
results_CLOE.iteration(initEst,expNb)=output;
|
|
results_CLOE.lambdas_upper(initEst, expNb, :) = lambda.upper;
|
|
results_CLOE.lambdas_lower(initEst, expNb, :) = lambda.lower;
|
|
% results_CLOE.jacobians(initEst, expNb, :, :) = jacobian; % Requires too much memory
|
|
|
|
s = svd(jacobian);
|
|
results_CLOE.jacobian_sigma_upper(initEst, expNb) = max(s);
|
|
results_CLOE.jacobian_sigma_lower(initEst, expNb) = min(s);
|
|
|
|
end
|
|
end
|
|
end
|
|
|