From Wikipedia, the free encyclopedia
The Observability Gramian is a Gramian used in control theory to determine whether or not a linear system is observable .
For a linear system described by
x
˙
(
t
)
=
A
(
t
)
x
(
t
)
+
B
(
t
)
u
(
t
)
{\displaystyle {\dot {x}}(t)=A(t)x(t)+B(t)u(t)}
y
(
t
)
=
C
(
t
)
x
(
t
)
+
D
(
t
)
u
(
t
)
{\displaystyle y(t)=C(t)x(t)+D(t)u(t)\,}
The observability Gramian for a linear time variant system is given by
W
o
(
t
0
,
t
1
)
=
∫
t
0
t
1
Φ
T
(
s
,
t
0
)
C
T
(
s
)
C
(
s
)
Φ
(
s
,
t
0
)
d
s
{\displaystyle W_{o}(t_{0},t_{1})=\int _{t_{0}}^{t_{1}}\Phi ^{T}(s,t_{0})C^{T}(s)C(s)\Phi (s,t_{0})ds}
,
where
Φ
{\displaystyle \Phi }
is the state transition matrix .
The system is observable on the interval
t
∈
[
t
0
,
t
1
]
{\displaystyle t\in [t_{0},t_{1}]}
if and only if
W
o
(
t
0
,
t
1
)
{\displaystyle W_{o}(t_{0},t_{1})}
is nonsingular . In the case of a linear time invariant system, this can be simplified to finding the rank of the "observability matrix". If
x
(
t
)
{\displaystyle x(t)}
is a
n
{\displaystyle n}
-dimensional real-valued vector, then the system is observable if and only if
rank
[
C
T
,
A
T
C
T
,
.
.
.
,
(
A
T
)
n
−
1
C
T
]
=
n
{\displaystyle {\text{rank}}[C^{T},A^{T}C^{T},...,(A^{T})^{n-1}C^{T}]=n}
See also
References
External links