Gaussian White Noise Input into ode solver?

Hey all,

Here is the jist of it: I want to input Band-limited Guassian White Noise into a nonlinear system of ordinary differential equations, but want the white noise signal preserved so that I can plot it against the final solution. Any ideas? The full story is below:

So I have this function:

function vprime = nagumo(t,v)

vprime=[v(1)-(1/3)*(v(1))^3-v(2); 
    0.08*(v(1)+0.7-0.8*v(2))];

end

I call this function by setting:

tspan=[0 100];
x0=[0 0];
[t,x]=ode45(@nagumo, tspan, x);

This is a pretty simple ode calculation. However, where I get confused is when I'm trying to do the same thing as above but input Band-limited Guassian White Noise into the system such that the new equation would look like this:

function vprime = nagumo(t,v)
  vprime=[v(1)-(1/3)*(v(1))^3-v(2)+WHITENOISE; 
      0.08*(v(1)+0.7-0.8*v(2))];
end

I am able to input white noise by simply adding wgn(1,1,0) and generating a response but have no way of saving that inputted white noise signal so that I can plot the output (what the ode solves) against the inputting white noise. Any help as to how I can do that?

[EL890]

Hi Mario,

I am currently encountering a similar issue with adding white noise to a nonlinear model in Matlab. Were you able to resolve this?

[Kwin]

You could generate the noise in advance and interpolate, with something like zero- or first-order-hold. I think ode solvers would prefer foh over zoh since it is continuous. This will result into band-limited noise, so by choosing the interpolation interval appropriately you can assure that certain frequencies get excited.