Repository navigation
"Interconnect" input specification. #1015
Description
Hi,
In this code I want "interconnect" to build a system with just one input "sys.u1". It tries to create two inputs and as a result catches an exception. Am I getting it wrong?
import numpy as np
import control as ct
f = np.array([[-35, 1], [-225, -5]])
j = np.array([[0], [5]])
g = np.array([[-965], [-8775]])
cobs = np.eye(2)
dobs = np.zeros((2, 2))
a = np.array([[0, 1, 0], [0, -1, 1], [0, 0, -5]])
b = np.array([[0], [0], [5]])
c = np.array([[1, 0, 0]])
d = 0
plant = ct.ss(a, b, c, d,
inputs=["u1"], outputs=["y1"], name="plant")
observer = ct.ss(f, np.hstack((g, j)), cobs, dobs,
inputs=["u1", "u2"], outputs=["y1", "y2"], name="observer")
sys = ct.interconnect([plant, observer],
connections=[["observer.u1", "plant.y1"]],
inplist=[["plant.u1", "observer.u2"]],
outlist=[["plant.y1"], ["observer.y1"], ["observer.y2"]],
inputs=["u1"],
outputs=["y1", "y2", "y3"],
debug=True)
Activity
murrayrm commentedon Jun 29, 2024
Looks like a bug in the way input (and output) lists are handled. The code is not catching the fact that a single input is going to multiple subsystems.
Fix coming in a bit.
fix interconnect() issue python-control#1015
bonidydy commentedon Jun 30, 2024
It seems there is also a mistake in the documentation of "interconnect", "inplist" argument:
inplist (list of input connections, optional) –
List of connections for how the inputs for the overall system are mapped
to the subsystem inputs. The input specification is similar to the form
defined in the connection specification, except that connections do not
specify an input-spec, since these are the system inputs. The entries for a
connection are thus of the form:
[input-spec1, input-spec2, …]
I believe the bold text should be output-spec.
updated documentation per comment from @bonidydy in python-control#1015
Merge pull request #1017 from murrayrm/interconnect_bug-29Jun2024