Re: [python-control-discuss] python-control-4b

[Richard M. <mu...@cd...>]

Roberto,

The problem may be that control.Lti objects are no longer compatible with scipy.signal.lti objects, which the scipy.signal.step() function expects.  A couple of fixes:

* Instead of using scipy.signal.step(), use the control.matlab.step() function.  You can do this by just getting rid of the scipy.signal import statement.  Note that we don't have a step2() to function yet, so you may get warnings if you pass a system with a pole at the origin and you don't give a time interval (because it can't figure out the default time scale to use).

* You can get the signal.lti object from a control.Lti object using the returnScipySignalLti() method.  Note that this returns a 2-dimensional set of scipy.signal.lti objects, even for SISO systems (this will eventually be under the control of the toSISO and toMIMO keywords, which I hope to make part of the v0.4c release).

Also, it would be nice to hear how you are implementing the discrete time modifications.  I think the cleanest way to do this may be to create a separate class, derived from Lti, for discrete-time systems (transfer functions and state space).

-richard

On 15 Apr 2011, at 5:04 , Roberto Bucher wrote:

> Hi Richard
> 
> I've finally found some time to look the new modifications for your control 
> toolbox. With few modifications I've reached to put the sampling time in the 
> new matlab.py, statesp.py and xferfcn.py files.
> 
> My examples run correctly again, but I have some problems with the step and 
> step2 functions. Here is the code:
> 
> #-------------------------------------------------------
> from matplotlib.pylab import *
> from control.matlab import *
> from scipy.signal import step, step2
> 
> # Motor parameters
> Kt = 126e-6
> Jm = 1.3024e-3
> Dm = 857.4e-6
> 
> # Transfer function
> g=tf([Kt/Jm],[1,Dm/Jm,0])  
> 
> # Continous state space form
> a=[[0,1],[0,-Dm/Jm]]
> b=[[0],[1]]
> c=[[Kt/Jm,0]];
> d=[0];
> sysc=ss(a,b,c,d)
> 
> t1,y1 = step(g)
> t2,y2 = step(sysc)
> t1,y1 = step2(g)
> t2,y2 = step2(sysc)
> #---------------------------------------------------------
> I have some problems with the step and step2 functions (linalg problems or 
> 
>    695         sys = system
>    696     else:
> --> 697         sys = lti(*system)
>    698     if N is None:
>    699         N = 100
> 
> TypeError: type object argument after * must be a sequence, not instance
> 
> Any Idea?
> 
> Thanks in advance
> 
> Roberto
> 
> -- 
> -----------------------------------------------------------------------------
> Great spirits have always encountered violent opposition 
> from mediocre minds (A. Einstein)
> ----------------------------------------------------------------------------
> University of Applied Sciences of Southern Switzerland
> Dept. Innovative Technologies
> CH-6928 Lugano-Manno
> http://web.dti.supsi.ch/~bucher