Subscript indices must either be real positive integers or logicals while k starts from 1 .

% I've fun_dquat called by  another function RK4 as shown below. Error says, Subscript indices must either be real positive integers or
%logicals. 
function dquat=fun_dquat(k,q)
global wwx 
global wwy
global wwz 
wx=wwx;
wy=wwy;
wz=wwz;
ww=[wx;wy;wz]'; %64 by 3 matrix obtained in each iteration 
Wskew=  [ 0,-ww(k,1),-ww(k,2),-ww(k,3);
ww(k,1),  0, ww(k,3),-ww(k,2);
ww(k,2),-ww(k,3),  0, ww(k,1);
ww(k,3), ww(k,2),-ww(k,1),  0]
coef=0.1;
c=coef*(1-q*q');
dquat=0.5*(Wskew*q')'+c*q;
end
% Function RKF utilize fun_dquat to compute dquat

 function yout = RK4(F,t0,h,tfinal,y0)
   % ODE4  Classical Runge-Kutta ODE solver.
   %   yout = ODE4(F,t0,h,tfinal,y0) uses the classical
   %   Runge-Kutta method with fixed step size h on the interval
   %      t0 <= t <= tfinal
   %   to solve
   %      dy/dt = F(t,y)
   %   with y(t0) = y0.
   %   Copyright 2014 - 2015 The MathWorks, Inc.
   
      y = y0;
      yout = y;
      for tt = t0 : h : tfinal-h
         s1 = F(tt,y);
         s2 = F(tt+h/2, y+h*s1/2);
         s3 = F(tt+h/2, y+h*s2/2);
         s4 = F(tt+h, y+h*s3);
         y = y + h*(s1 + 2*s2 + 2*s3 + s4)/6;
         yout = [yout; y]; %#ok<AGROW>
      end
      
      %error message
      RK4 nonunit quaternion integration:
      Subscript indices must either be real positive integers or logicals. 
      Error in fun_dquat (line 9)
      Wskew=  [ 0,-ww(k,1),-ww(k,2),-ww(k,3);    
      Error in RK4 (line 14)
      s1 = F(tt,y);
      Error in PRS20200428 (line 187)
      q_rk4 = RK4(@fun_dquat,0,delt,t_max,q_init); % Integration  to get the next quaternion
      
    %Can anyone help to figureout where the problem of the code is ?  Thanks

[Geoff Hayes]

HN - please copy and paste the full error message to this question. Also, please confirm if the F input parameter to your RK4 function is the fun_dquat. If so, then the first input to this function is possible a non-integer

for tt = t0 : h : tfinal-h
    s1 = F(tt,y);
    % etc.
end

and so the error message makes sense.

[HN]

%error message
      RK4 nonunit quaternion integration:
      Subscript indices must either be real positive integers or logicals. 
      Error in fun_dquat (line 9)
      Wskew=  [ 0,-ww(k,1),-ww(k,2),-ww(k,3);    
      Error in RK4 (line 14)
      s1 = F(tt,y);
      Error in PRS20200428 (line 187)
      q_rk4 = RK4(@fun_dquat,0,delt,t_max,q_init); % Integration  to get the next quaternion
      
   Yes, fun_dquat is input for RK4. 

[Geoff Hayes]

So the k input to your fun_dquat is neither a positive integer nor logical...and so is invalid as an index into the array. Are these the correct values that you want to pass into this function?

[HN]

Yes, ww are the variables I wanted to pass from main file to function fun_dquat. The global declaration seem to confused as well.

Answer 1

[James Tursa]

The RK4( ) function expects to call the derivative function with the signature F(tt,y), i.e. time is the 1st argument.

But your fun_dquat( ) signature has k as the 1st argument, which you are using as an index in your code.

So you have a mismatch, hence the error. Your derivative function needs to have the signature that the RK4 expects when making that F(tt,y) and similar calls.

[HN]

,

ww is passing from the main file. q is predicted from initial point. I used rung-kutta integrator to next q. So I've only ww and q as a variable to to deal with RK4. Thats what I tried to do.

Thanks

[James Tursa]

So why do you have a k in your input argument and why are you using k as an index in your code? I don't see any k in the above formula. What are the global wwx, wwy, wwz in your code? Arrays of sampled rates? And you are trying to pick off the correct sample with k?

The RK4 scheme is going to try and generate a derivative at "fractional" times. If all you have is sampled rate arrays at specific times, you will need to interpolate them within the derivative function to get rates at the desired times for the RK4 scheme to use.