script.m

% optim
% script.m
% Runs a single optimization using optim.m

% This is an example of how to use optim.m. 
% It runs an optimization using walk data from Gait_Data_Sub1

clear all

% define the movement data to be tracked: Excel filename and sheet nasme
problem.gaitdata = {'Gait_Data_Sub1.xls' 'Slow Walk'};

% Uncomment one of these structure elements so the program will use either
% a random initial guess, or results from the prevoius optimization. Only
% use result.mat if the program converged and the results were okay.
problem.initialguess = 'random';
%problem.initialguess = 'resultX.mat';

% general optimization settings
problem.MaxIterations 	= 50000;
problem.print = 1;
problem.solver = 'IPOPT';		% use this solver

problem.N = 30;					% use this number of nodes (optim will reduce this to number of gait samples, if needed

% lower and upper bounds on some of the variables we are optimizing (set lower=upper if variable should not be optimized)
problem.L_u1 = -1.00;       problem.U_u1 = 1.0;        % Lower/Upper bounds for pump signal
problem.L_u2 = 0.001;		problem.U_u2 = 1.0; 		% Lower/Upper bounds for valve U2 (uncomment this line to keep valve 1 closed)
problem.L_u3 = 0.001;		problem.U_u3 = 1.0; 		% Lower/Upper bounds for valve U3 
problem.L_k  = 0.1;			problem.U_k  = 100;			% spring stiffness in MPa/ml
problem.L_CPA  = 0;	    	problem.U_CPA  = 10;        % compliance of parallel accumulator ml/MPa

% typical ankle k = 400 n/rad

% weights for cost function
problem.w1 				= 10.0;			% weight for angle tracking term
problem.w2				= 1.0;			% weight for moment tracking term
problem.w3 				= 0.1;			% weight for valve 1 control accelerations
problem.w4 				= 0.1;			% weight for valve 2 control accelerations
problem.w5              = 0.05;         % weight for the pump controller 
problem.wreg 			= 0;			% weight for regularization (smoothness) term
problem.Wdisc 			= 0.0;			% weighting to encourage discrete control levels
problem.Ndisc 			= 2.0;			% number of discrete control levels (dont touch)
problem.phisd = 5*pi/180;				% we want to be within x deg when tracking
problem.Msd = 5;						% we want to be within x Nm when tracking

% model parameters for the hydraulic hardware
problem.G 			= 0.137;		% units: cm^-3
problem.C1max		= 30;			% Cmax for valve 1
problem.C2max		= 30;			% Cmax for valve 2
problem.B1 			= 0.01;			% viscous drag coefficient for valve 1 & its tubing
problem.B2 			= 0.01;			% viscous drag coefficient for valve 2 & its tubing
problem.B3          = 0.01;         % viscous drag coefficient for pump and hardware
problem.Cf1         = 0.01;         % Accumulator C1 capacity 
%problem.Cf2         = 0.01;         % Accumulator C2 capacity

	
% do the optimization	
result = optim(problem);
result
if (result.info ~= 0)
	disp('THE OPTIMIZATION DID NOT CONVERGE, USE RESULTS WITH CAUTION');
end