There are a number of problems with your problem formulation.
I think the primary one may be that your system, as defined, is unstable.
You can calculate roots(den) and see that it has poles in the right half plane.
Did you intend for the system to be unstable? I would have to dig into it more but my intuition is that identifying frequency responses of unstable systems using a chirp may be problematic.
Also, if you are looking for a chirp with a linearly increasing frequency I think you should define it as:
in = a*cos(2*pi*(beta*t+fIni));
in = a*cos(2*pi*(beta*t+fIni).*t);
which results in the time value being squared.