Control Systems/Controllability and

Observability
1

System Interaction

Controllability

A state x0 is controllable at time t0 if for some nite

time t1 there exists an input u(t) that transfers the state
In the world of control engineering, there are a slew of x(t) from x0 to the origin at time t1 .
systems available that need to be controlled. The task of a A system is called controllable at time t0 if every state
control engineer is to design controller and compensator x0 in the state-space is controllable.
units to interact with these pre-existing systems. However, some systems simply cannot be controlled (or, more
often, cannot be controlled in specic ways). The con- 2.1 Controllability Matrix
cept of controllability refers to the ability of a controller
For LTI (linear time-invariant) systems, a system is reachto arbitrarily alter the functionality of the system plant.
able if and only if its controllability matrix, , has a full
The state-variable of a system, x, represents the internal
row rank of p, where p is the dimension of the matrix A,
workings of the system that can be separate from the regand p q is the dimension of matrix B.
ular input-output relationship of the system. This also
needs to be measured, or observed. The term observ- [Controllability Matrix]
ability describes whether the internal state variables of
the system can be externally measured.
[
]
= B AB A2 B Ap1 B Rppq
A system is controllable or Controllable to the origin
when any state x1 can be driven to the zero state x = 0 in
a nite number of steps.

Complete state controllability (or simply controllability if A system is controllable when the rank of the system mano other context is given) describes the ability of an exter- trix A is p, and the rank of the controllability matrix is
nal input to move the internal state of a system from any equal to:
initial state to any other nal state in a nite time interval
We will start o with the denitions of the term controlRank() = Rank(A1 ) = p
lability, and the related term reachability
If the second equation is not satised, the system is not .
Controllability A system with internal state vector x is MATLAB allows one to easily create the controllability
called controllable if and only if the system states matrix with the ctrb command. To create the controllacan be changed by changing the system input.
bilty matrix simply type
Reachability A particular state x1 is called reachable if
there exists an input that transfers the state of the = ctrb(A, B)
system from the initial state x0 to x1 in some nite where A and B are mentioned above. Then in order to
time interval [t0 , t).
determine if the system is controllable or not one can use
the rank command to determine if it has full rank.
We can also write out the denition of reachability more
If
precisely:
A state x1 is called reachable at time t1 if for some nite
initial time t0 there exists an input u(t) that transfers the Rank(A) < p
state x(t) from the origin at t0 to x1 .
Then controllability does not imply reachability.
A system is reachable at time t1 if every state x1 in the
Reachability always implies controllability.
state-space is reachable at time t1 .
Controllability only implies reachability when the
state transition matrix is nonsingular.

Similarly, we can more precisely dene the concept of

controllability:
1

3 OBSERVABILITY

2.2

Determining Reachability

t1

(t1 , )B( )B ( ) (t1 , )d

There are four methods that can be used to determine if Wr (t0 , t1 ) =
t
0
a system is reachable or not:
The system is reachable if the rank of the reachability
1. If the p rows of (t, )B(t) are linearly independent gramian is the same as the rank of the system matrix:
over the eld of complex numbers. That is, if the
rank of the product of those two matrices is equal to
rank(Wr ) = p
p for all values of t and
2. If the rank of the controllability matrix is the same <chemistry>/control{range}
as the rank of the system matrix A.
3. If the rank of rank[I A, B] = p for all eigenval- 2.5 Controllability Gramian
ues of the matrix A.
We can dene the controllability gramian of a system
4. If the rank of the reachability gramian (described (A, B) as:
below) is equal to the rank of the system matrix A.
[Controllability Gramian]
Each one of these conditions is both necessary and sucient. If any one test fails, all the tests will fail, and the
t1
system is not reachable. If any test is positive, then all the
Wc (t0 , t1 ) =
(t0 , )B( )B ( ) (t0 , )d
tests will be positive, and the system is reachable.
t0

2.3

Gramians

Gramians are complicated mathematical functions that

can be used to determine specic things about a system.
For instance, we can use gramians to determine whether
a system is controllable or reachable. Gramians, because
they are more complicated than other methods, are typically only used when other methods of analyzing a system
fail (or are too dicult).

The system is controllable if the rank of the controllability

gramian is the same as the rank of the system matrix:

rank(Wc ) = p

If the system is time-invariant, there are two important

points to be made. First, the reachability gramian and
the controllability gramian reduce to be the same equation. Therefore, for LTI systems, if we have found one
gramian, then we automatically know both gramians.
All the gramians presented on this page are all matrices Second, the controllability gramian can also be found as
with dimension p p (the same size as the system matrix the solution to the following Lyapunov equation:
A).
All the gramians presented here will be described usAWc + Wc A = BB
ing the general case of Linear time-variant systems. To
change these into LTI (time-invariant equations), the fol- Many software packages, notably MATLAB, have funclowing substitutions can be used:
tions to solve the Lyapunov equation. By using this last
relation, we can also solve for the controllability gramian
using these existing functions.
A(t )
(t, ) e

(t, ) eA (t )

3 Observability

Where we are using the notation X' to denote the transpose of a matrix X (as opposed to the traditional notation The state-variables of a system might not be able to be
measured for any of the following reasons:
XT ).

2.4

Reachability Gramian

We can dene the reachability gramian as the following

integral:
[Reachability Gramian]

1. The location of the particular state variable might

not be physically accessible (a capacitor or a spring,
for instance).
2. There are no appropriate instruments to measure the
state variable, or the state-variable might be measured in units for which there does not exist any measurement device.

3.2

Observability Matrix

3. The state-variable is a derived dummy variable 3.2 Observability Matrix

that has no physical meaning.
The observability of the system is dependant only on the
If things cannot be directly observed, for any of the rea- system states and the system output, so we can simplify
sons above, it can be necessary to calculate or estimate our state equations to remove the input terms:
the values of the internal state variables, using only the in- Matrix Dimensions:
put/output relation of the system, and the output history A: p p
of the system from the starting time. In other words, we B: p q
must ask whether or not it is possible to determine what C: r p
the inside of the system (the internal system states) is like, D: r q
by only observing the outside performance of the system
(input and output)? We can provide the following formal
denition of mathematical observability:
Observability A system with an initial state, x(t0 ) is
observable if and only if the value of the initial
state can be determined from the system output y(t)
that has been observed through the time interval
t0 < t < tf . If the initial state cannot be so determined, the system is unobservable.

x (t) = Ax(t)
y(t) = Cx(t)

Therefore, we can show that the observability of the system is dependant only on the coecient matrices A and
C. We can show precisely how to determine whether a
system is observable, using only these two matrices. If
Complete Observability A system is said to be com- we have the observability matrix Q:
pletely observable if all the possible initial states [Observability Matrix]
of the system can be observed. Systems that fail this
criteria are said to be unobservable.

C
Detectability A system is Detectable if all states that
CA

cannot be observed decay to zero asymptotically.

2

Q = CA
..
.
Constructability A system is constructable if the
CAp1
present state of the system can be determined from
the present and past outputs and inputs to the system. If a system is observable, then it is also con- we can show that the system is observable if and only if
structable. The relationship does not work the other the Q matrix has a rank of p. Notice that the Q matrix
has the dimensions pr p.
way around.
MATLAB allows one to easily create the observability
A system state x is unobservable at a given time t if the matrix with the obsv command. To create the observzero-input response of the system is zero for all time t. If abilty matrix Q simply type
a system is observable, then the only state that produces a
zero output for all time is the zero state. We can use this
Q=obsv(A,C)
concept to dene the term state-observability.
where A and C are mentioned above. Then in order to
State-Observability A system is completely statedetermine if the system is observable or not one can use
observable at time t0 or the pair (A, C) is observable
the rank command to determine if it has full rank.
at t0 if the only state that is unobservable at t0 is the
zero state x = 0.

3.3 Observability Gramian

3.1

Constructability

We can dene an observability gramian as:

A state x is unconstructable at a time t1 if for every nite [Observability Gramian]

time t < t1 the zero input response of the system is zero
for all time t.
t1
A system is completely state constructable at time t1 if Wo (t0 , t1 ) =
(, t0 )C ( )C( )(, t0 )d
t
0
the only state x that is unconstructable at t0 is x = 0.
If a system is observable at an initial time t0 , then it is A system is completely state observable at time t0 < t <
constructable at some time t > t0 , if it is constructable at t1 if and only if the rank of the observability gramian is
equal to the size p of the system matrix A.
t1 .

If the system (A, B, C, D) is time-invariant, we can construct the observability gramian as the solution to the Lyapunov equation:

A Wo + Wo A = C C

3.4

Constructability Gramian

We can dene a constructability gramian as:

[Constructability Gramian]

t1

Wcn (t0 , t1 ) =

(, t1 )C ( )C( )(, t1 )d

t0

A system is completely state observable at an initial time

t0 if and only if there exists a nite t1 such that:

rank(W0 ) = rank(Wcn ) = p
Notice that the constructability and observability gramians are very similar, and typically they can both be calculated at the same time, only substituting in dierent
values into the state-transition matrix.

Duality Principle

The concepts of controllability and observability are very

similar. In fact, there is a concrete relationship between
the two. We can say that a system (A, B) is controllable if
and only if the system (A', C, B', D) is observable. This
fact can be proven by plugging A' in for A, and B' in for
C into the observability Gramian. The resulting equation will exactly mirror the formula for the controllability
gramian, implying that the two results are the same.

DUALITY PRINCIPLE

Text and image sources, contributors, and licenses

5.1

Text

Control Systems/Controllability and Observability Source: https://en.wikibooks.org/wiki/Control_Systems/Controllability_and_

Observability?oldid=2999948 Contributors: Whiteknight, Adrignola, Tim.greatrex, Nonstandard, Esj88, Davetcoleman and Anonymous:
15

5.2

Images

5.3

Content license

Creative Commons Attribution-Share Alike 3.0