aerskine / azimuthproject Goto Github PK

I think Simulink is the ideal tool to do these kind of simulations, however, 
this is the quick and dirty Matlab code to do this, if anyone wants to try:

beta=1;lambda=0.1;

dv=@(t,v) [-v(2);v(1)]+beta*v-v’*v*v+lambda*randn(2,1);

[t,v]=ode45(dv,[0 10],[0 2]);

plot(v(:,1),v(:,2));axis([-2 2 -1.5 2])

The first line sets up the parameters, the second sets up the dynamical system 
(where v is the vector having x and y as components), the third solves the ode 
from 0 to 10 seconds, with an initial condition of [0 2], and the fourth line 
plots the results (use plot(t,v(:,1)) to plot the first component versus time).

The plots are smoother than the one shown above, perhaps because I have assumed 
that the variance of the white noise is 1, and maybe a larger one was assumed 
above. Multiplying lambda by 10 produces very similar plots.