Hi Robert,
Your system is already marginally stable. You can see the poles if you type
pzamp(H) ~ all are in the left half plane with a root at the origin (type one system).
Overshoot function?? I dont believe there is such a thing. There are equations that relate the damping ratio (phase margin) to the overshoot of the system. Have a look at Dorf and bishop / Ogata / Nise. All of these are great texts.
Your system doesn't satisfy the dominance condition so these relations are not exact, rather approximations only.
For further refinement of the closed loop system try the sistool. You can choose your controller and design specs and it will optimise the controller to meet the criteria specified.
> sisotool(H)
For the nyquist plot
> nyquist(H)
