Bexley asked . 2020-11-28

How to solve evalfis error?

I try to run the program "Modeling Inverse Kinematics in a Robotic Arm"

l1 = 10; % length of first arm
l2 = 7; % length of second arm

theta1 = 0:0.1:pi/2; % all possible theta1 values
theta2 = 0:0.1:pi; % all possible theta2 values

[THETA1,THETA2] = meshgrid(theta1,theta2); % generate a grid of theta1 and theta2 values

X = l1 * cos(THETA1) + l2 * cos(THETA1 + THETA2); % compute x coordinates
Y = l1 * sin(THETA1) + l2 * sin(THETA1 + THETA2); % compute y coordinates

data1 = [X(:) Y(:) THETA1(:)]; % create x-y-theta1 dataset
data2 = [X(:) Y(:) THETA2(:)]; % create x-y-theta2 dataset

plot(X(:),Y(:),'r.'); 
  axis equal;
  xlabel('X','fontsize',10)
  ylabel('Y','fontsize',10)
  title('X-Y coordinates generated for all theta1 and theta2 combinations using forward kinematics formula','fontsize',10)
  
opt = anfisOptions;
opt.InitialFIS = 7;
opt.EpochNumber = 150;
opt.DisplayANFISInformation = 0;
opt.DisplayErrorValues = 0;
opt.DisplayStepSize = 0;
opt.DisplayFinalResults = 0;

disp('--> Training first ANFIS network.')

anfis1 = anfis(data1,opt);

disp('--> Training second ANFIS network.')

opt.InitialFIS = 6;
anfis2 = anfis(data2,opt);

x = 0:0.1:2; % x coordinates for validation
y = 8:0.1:10; % y coordinates for validation

[X,Y] = meshgrid(x,y);

c2 = (X.^2 + Y.^2 - l1^2 - l2^2)/(2*l1*l2);
s2 = sqrt(1 - c2.^2);
THETA2D = atan2(s2,c2); % theta2 is deduced

k1 = l1 + l2.*c2;
k2 = l2*s2;
THETA1D = atan2(Y,X) - atan2(k2,k1); % theta1 is deduced

Until I run the coding below, I got error using evalfis

XY = [X(:) Y(:)];
THETA1P = evalfis(anfis1,XY); % theta1 predicted by anfis1
THETA2P = evalfis(anfis2,XY); % theta2 predicted by anfis2

The error that I get:

Error using evalfis (line 51)
The second argument must be a FIS structure.

matlab , control system , fuzzy logic ,