Controllability & Observability

of Control Systems
State Space Representation

Control Systems
Course Instructor: Deepthi M S
AY: 2023-24
Controllability
➢ The generic structure of a state-space model of a nth order continuous time
dynamical system with m inputs and p outputs is given by

𝒙ሶ 𝑡 = 𝐴 𝒙 𝑡 + 𝐵 𝒖 𝑡
𝒚 𝑡 =𝐶𝒙 𝑡 +𝐷𝒖 𝑡
• 𝒙 𝑡 → state vector
• 𝒖 𝑡 → control input
• 𝐴 → n x n matrix, 𝐵 → n x m matrix , 𝐶 → p x n matrix , 𝐷 → p x m matrix

➢ Controllable: A system is said to be controllable at time 𝑡0 if it is possible, by means

of any unconstrained input control vector to transfer the system from any initial state
𝑥 𝑡0 to any other state in a finite interval of time
03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 2
Controllability
➢ Completely controllable: A system is said to be completely controllable, if every
state variable of the system can be controlled and able to reach certain objective in
finite time by using unconstrained control vector.

➢ Uncontrollable : if any one of the state variables is independent of control u(t), then
there would be no way of driving this state to desired state in finite, then system is
said to be uncontrollable or not completely controllable.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 3

Controllability: Example of Uncontrollable System
• The below system is an uncontrollable linear system with two state variables.
• 𝒖 𝒕 effects only the state 𝒙𝟏 (𝒕) and the state 𝒙𝟐 (𝒕) is uncontrollable.
• It is not possible to drive 𝒙𝟐 (𝒕) from any initial state 𝒙𝟐 (𝒕𝟎 ) to desired state
𝒙𝟐 (𝒕𝒇 ) in a finite time interval 𝒕𝒇 − 𝒕𝟎 by control vector 𝒖 𝒕 .

The entire system is said to be uncontrollable

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 4

Controllability: Condition
➢ The generic structure of a state-space model of a nth order continuous time
dynamical system with m inputs and p outputs is given by

𝒙ሶ 𝑡 = 𝐴 𝒙 𝑡 + 𝐵 𝒖 𝑡
𝒚 𝑡 =𝐶𝒙 𝑡 +𝐷𝒖 𝑡

• The necessary and sufficient condition for controllability is, the following matrix
has a rank ‘n’

𝑺 = [𝑩 𝑨𝑩 𝑨𝟐 𝑩 … … … . 𝑨𝒏−𝟏 𝑩]
This matrix is called controllability matrix S
Alternative condition for controllability is 𝒅𝒆𝒕. 𝑺 ≠ 𝟎

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 5

Controllability: Example Problem

Rank n = 3 , since det.S is not zero . The system

is completely controllable

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 6

Controllability: Practice Problem
Find whether the given system is controllable or not.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 7

Observability:
• Observability: A system is said to be observable at time 𝒕𝟎 if it is possible to
determine the system state 𝐱 𝒕𝟎 from the observation of the output over finite
time interval.

• This happens only when every state variables of the system affects some of the
outputs.
• If any one of the state variable information cannot be observed by measuring the
output, the state is said to be unobservable and the system is said to be not
observable.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 8

Observability: Example of Unobservable System
• The below system is an unobservable linear system with two state variables.
• The state 𝒙𝟐 (𝒕) is not connected to y 𝒕 in anyway.
• Once we measure 𝐲 𝒕 , we can observe state 𝒙𝟏 (𝒕) , since 𝒙𝟏 𝒕 = 𝒚(𝒕).
• State 𝒙𝟐 (𝒕) cannot be observed from the information on 𝒚(𝒕).

The entire system is said to be unobservable.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 9

Observability:
• The system is said to be completely observable if every transition of the state
eventually affects every element of the output vector.

• The concept of observability is useful in solving problem of reconstructing the

unmeasurable state variables from measured state variables and outputs.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 10

Observability : Condition
➢ The generic structure of a state-space model of a nth order continuous time
dynamical system with m inputs and p outputs is given by

𝒙ሶ 𝑡 = 𝐴 𝒙 𝑡 + 𝐵 𝒖 𝑡
𝒚 𝑡 =𝐶𝒙 𝑡 +𝐷𝒖 𝑡

• The necessary and sufficient condition for observability is, the following matrix
has a rank ‘n’

𝒏−𝟏 𝑻
𝑺𝑶 = [𝑪𝑻 ……. 𝑨𝑻 𝑪𝑻 𝑨𝑻 𝑪 ]
This matrix is called Observability matrix 𝑺𝑶 Alternative condition for
controllability is 𝒅𝒆𝒕. 𝑺𝑶 ≠ 𝟎
03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 11
Observability : Example problem

The rank of matrix 𝑺𝑶 is 1, it is

singular, therefore the system is
unobservable.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 12

Observability : practice problem
• Find whether the given system is observable or not

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 13

Advantages of Controllability & Observability

• Controllability provides how the system responds to different inputs. Hence, it

helps in providing a flexible system response.

• The controllable systems can adapt to changes when the operating conditions are
changed.

• Observability helps in the precise control and monitoring of the system.

• Observability checks the internal behavior of the system which helps in detecting
the faults inside the system

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 14

Disadvantages of Controllability & Observability

• It is very difficult to achieve controllability in a complex system.

• It leads to increasing the complexity of the design.

• To achieve controllability in the system, additional control inputs are required.

• increase in the energy consumption of the system.

• The observability of the system is measured using the sensors of the system.
• Failures and limitations of the performance of the sensor can reduce the accuracy of
the state estimation.

• The observability methods are computationally intensive which increases the cost of the
system due to the addition requirement of the computational resources.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 15

State Space Representation
• In control engineering, a state-space representation is a mathematical model of a
physical system as a set of input, output and state variables related by differential
equations or difference equations.
• The state-space representation → "time-domain approach"
• Provides a convenient and compact way to model and analyze systems with
multiple inputs and outputs

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 16

Continued…
• The system governed by a nth order differential or difference equation or nth order
transfer function is expressed in terms of
• “n” state variables: x1, x2,....... xn

• Selection of state variables is arbitrary.

• The best choice of state variables in electrical network are currents and voltages
in energy storing elements

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 17

 State Space Representation
➢ The generic structure of a state-space model of a nth order continuous time
dynamical system with m inputs and p outputs is given by

𝒙ሶ 𝑡 = 𝐴 𝒙 𝑡 + 𝐵 𝒖 𝑡
State Space Model
𝒚 𝑡 =𝐶𝒙 𝑡 +𝐷𝒖 𝑡

𝑥1 (𝑡)
𝑢1 (𝑡) 𝑦1 (𝑡)
𝑥2 (𝑡)
𝑢2 (𝑡) 𝑦2 (𝑡)
.
. . 𝑺𝒕𝒂𝒕𝒆 𝑽𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒗𝒆𝒄𝒕𝒐𝒓 , 𝒙 𝑡 =
𝑰𝒏𝒑𝒖𝒕 𝒗𝒆𝒄𝒕𝒐𝒓 , 𝒖 𝑡 = 𝑶𝒖𝒕𝒑𝒖𝒕 𝒗𝒆𝒄𝒕𝒐𝒓 , 𝒚 𝑡 = .
. .
.
. . 𝑥𝑛 (𝑡)
𝑢𝑚 (𝑡) 𝑦𝑝 (𝑡)

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 18

Continued…
𝑑𝑥1 (𝑡)/𝑑𝑡
𝑑𝑥2 (𝑡)/𝑑𝑡
𝒅𝒙(𝒕) .
• 𝑫𝒊𝒇𝒇𝒆𝒓𝒆𝒏𝒕𝒊𝒂𝒍 𝑺𝒕𝒂𝒕𝒆 𝑽𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒗𝒆𝒄𝒕𝒐𝒓 , =
𝒅𝒕 .
.
𝑑𝑥𝑛 (𝑡)/𝑑𝑡

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 19

Continued..
✓“t” → denotes time
✓x(t) is 𝑛 × 1 → denotes state vector, it is a function of time.
✓u(t) is m × 1 → denotes the input vector , a function of time
✓y(t) is p x 1 → denotes the output vector, a function of time

➢A is 𝑛 × 𝑛 ; A is the state matrix, a constant

➢B is 𝑛 × 𝑚 ; B is the input matrix, a constant
➢C is p × 𝑛; C is the output matrix, a constant
➢D is p× 𝑚 ; D is the direct transition (or feedthrough) matrix, a constant

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 20

Advantages of State Space Modelling:

➢Takes into account initial conditions.

➢It can be applied to non-linear and as well as time varying systems.

➢Applied to multiple inputs and multiple outputs (MIMO) systems.

➢Any type of input can be considered.

➢Involves matrix algebra , conveniently adopted to digital computers.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 21

Definitions :State, State Variables, State &
Output Equations
• State – Set of variables of smallest possible size that together with any input to
the system is sufficient to determine the future behavior (I.e., output) of the
system.
• Each state variable has “memory” E.g., voltage across the capacitor
• Each state variable has an “initial condition” E.g., its state at time t0
• State variables are typically associated with energy storage

• State vector: set of ‘n’ state variables, necessary to describe the complete
behavior of the system.

NOTE: the number of state variables, n, is equal to the number of independent

energy storage elements in the system.
03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 22
Continued…
State-space models describe a system’s dynamics via two equations
• State equation: The “state equation” describes how the input and state
variables influences the derivative of state variables;
• Output equation: The “output equation” describes how the state and the input
both directly influence the output

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 23

State Space Model for electrical network -
Example
• Consider a series RLC circuit,
• Number of energy storage elements = 2. (state
variables).
• State variables : 𝑖 𝑡 , 𝑣𝑐 (𝑡)
• Output variable: 𝑣𝑜 (𝑡)
• Output equation: 𝑣𝑜 𝑡 = 𝑣𝑐 (𝑡)
• Apply KVL around the loop

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 24

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 25
03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 26
Solution of State Equation:
• In this section let us investigate the solution of state equation

𝒙ሶ 𝑡 = 𝑨 𝒙 𝑡 + 𝑩 𝒖(𝑡)

Where 𝒙(𝑡) is 𝑛 × 1 state vector,𝒖(𝑡) is a input vector, 𝑨 is 𝑛 × 𝑛 real constant

matrix and 𝑩 is 𝑛 × 1 real constant vector.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 27

Homogeneous State Equation: (Unforced)

𝒙ሶ 𝑡 = 𝑨 𝒙 𝑡 + 𝑩 𝒖(𝑡)
• In this case the input is zero and output response is because of state variables
initial condition.
• Example : Consider a RC circuit with zero input voltage, output response will be
obtained due to initial voltage on the capacitor. The capacitor discharges through
the capacitor.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 28

Solution of Homogeneous State Equation :
• Consider scalar first order differential equation:
𝑑𝑥
= 𝑎𝑥 𝑡 , 𝑤𝑖𝑡ℎ 𝑥 0 = 𝑥0
𝑑𝑡
• This equation has the solution,
𝑥 𝑡 = 𝑒 𝑎𝑡 𝑥 0
• By analogy with the scalar case, vector state equation has the solution
𝒙 𝑡 = 𝑒 𝑨𝑡 𝒙 0
𝑨2 𝑡 2 𝑨3 𝑡 3
𝑒 𝑨𝑡 = 𝑰 + 𝑨𝑡 + + + …
2! 3!
Each term in the above equation is a matrix of order n x n and 𝑒 𝑨𝑡 is a matrix
exponential called state transition matrix 𝜙 𝑡 = 𝑒 𝑨𝑡
03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 29
 𝑨𝑡
State Transition Matrix 𝜙 𝑡 = 𝑒
• State transition matrix is a matrix exponential 𝑒 𝑨𝑡 , which drives initial
state 𝑥 0 at 𝑡 = 0 to a state 𝒙 𝑡 at time 𝑡.

• State transition matrix depends only on matrix A.

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 30

Non-Homogeneous State Equation :
𝒙ሶ 𝑡 = 𝑨 𝒙 𝑡 + 𝑩 𝒖(𝑡)
• Rewrite the above equation as,

𝒙ሶ 𝑡 − 𝑨 𝒙 𝑡 = 𝑩 𝒖(𝑡)
• Multiply both the sides of equation by 𝑒 −𝑨𝑡 , we obtain,

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 31

Non-Homogeneous State Equation :
• Integrating with respect to time ‘t’ between the limits 0 & ‘t’, gives

• Multiply both the sides of equation by 𝑒 𝑨𝑡 , we obtain

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 32

Computation of State Transition Matrix using
Laplace Transform
• Consider homogeneous state equation

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 33

Computation of State Transition Matrix using
Laplace Transform
• Consider non-homogeneous state equation

03-01-2025 Deepthi M S, Assistant Professor, Dept of ECE, NIE, Mysuru 34