State Space Modeling

Ibrarur Rahman Choudhury

Lecturer, MPE
Ahsanullah University of Science and Technology (AUST)

March 2, 2025

1 / 29
Objectives

Book: Norman S. Nise and Katsuhiko Ogata

Acknowledgment: Asso. Prof. Dr. Richard Hill (Detroit Mercy)

By the end of this lecture, students will be able to:

Understand the differences between classical and modern
(state-space) control system methods.
Explain how state-space modeling represents dynamic systems with
multiple state variables.
Convert higher-order differential equations to state-space form.
Identify and define state variables in a system.
Apply state-space modeling for dynamic systems, considering both
input-output relationships and energy storage elements.
Recognize the importance of choosing the minimum number of state
variables for efficiency and accuracy in control system design.

2 / 29
Pre-Assessment

Before starting the lecture, answer the following questions to assess your
prior knowledge:
1 What is the classical method of control system design, and how does

it differ from modern methods?

2 What do you understand by a state variable in control systems?

3 Why might state-space models be preferred over transfer function

models in certain applications?

4 Can you describe a dynamic system with time-varying inputs? Provide

an example.
5 What do you think is meant by “state-space representation of a

system”?

3 / 29
DESIGN OF CONTROL SYSTEM (CLASSICAL
METHOD)
Two approaches are available for the analysis and design of feedback control
systems.
The classical, or frequency-domain approach is based on converting a
system’s differential equation to a transfer function.
Thus generating a mathematical model of the system that algebraically
relates a representation of the output to a representation of the input.
Replacing a differential equation with an algebraic equation not only
simplifies the representation of individual subsystems but also simplifies
modeling interconnected subsystems.

ADVANTAGE:
1 Rapidly provide stability and
transient response information.
2 Thus, we can immediately see the
effects of varying system
parameters until an acceptable
design is met.

4 / 29
CLASSICAL METHOD: LIMITATIONS
With the arrival of space exploration, requirements for control systems
increased in scope.
Modeling systems by using LTI differential equations and subsequent
transfer functions became inadequate.
Many systems do not have just a single input and a single output.
Time-varying systems, (for example, missiles with varying fuel levels or lift in
an aircraft flying through a wide range of altitudes) can not be represented.
It can not conveniently handle, systems with nonzero initial conditions.
f (t) Ms 2 + fv s + K X (s) = F (s) Ms 2 + fv s + k X (s) − 5 = F (s) (1)

Figure: System Diagram

5 / 29
ENTER STATE SPACE
The state-space approach (also referred to as the modern, or time-domain
approach) is a unified method for modeling, analyzing, and designing a wide
range of systems.
It can be used to represent nonlinear systems that have backlash,
saturation, and dead zone.
It can conveniently handle systems with nonzero initial conditions.
Time-varying systems can be represented in state space.
Multiple-input, multiple-output systems (such as a vehicle with input
direction and input velocity yielding an output direction and an output
velocity) can be compactly represented in state space with a model similar in
form and complexity to that used for single-input, single-output systems.

6 / 29
ENTER STATE SPACE
The time-domain approach can be used to represent systems with a digital
computer in the loop or to model systems for digital simulation.
With a simulated system, system response can be obtained for changes in
system parameters—an important design tool.
The state space approach is also attractive because of the availability of
numerous state-space software packages for the personal computer.
Both classical and modern approaches can be used to solve LTI systems,
However.
State space approach is not as intuitive as the classical approach.
The designer has to engage in several calculations before the physical
interpretation of the model is apparent, whereas in classical control a few
quick calculations or a graphic presentation of data rapidly yields the
physical interpretation.

7 / 29
STATE SPACE MODELING
A state space model represents a system by a series of first order differential
state equations and algebraic output equations.
State Space models are numerically efficient to solve, can handle complex
systems, allow for a more geometric understanding of dynamic systems, and form
the basis for much of modern control theory, e.g., optimal control theory, Linear
quadratic control, Kalman filters.
Consider the following physical (dynamic) system where

u(t) is the input and ẋ(t) is the output

...
x + 5ẍ + 3ẋ + 2x = u
This is a third order DE. In order to
convert it to a series of first order DE
(generating state space model), we can
change the variables via pure
mathematical manipulation.
By choosing these new variables, we can
convert the initial third order DE to first
order:
x1 = x, x2 = ẋ, x3 = ẍ

8 / 29
STATE SPACE MODELING
...
ẍ + 5 x + 3ẋ + 2x = u
x1 = x, x2 = ẋ, x3 = ẍ
For each of the state variables we will need a single first order DE: (State
Equations:)
ẋ1 = ẋ = x2
ẋ2 = ẍ = x3
...
ẋ3 = x = −5ẍ − 3ẋ − 2x + u
ẋ3 = −5x3 − 3x2 − 2x1 + u
The original 3rd order DE has now been converted to three 1st order DE.
State space modeling also includes output equations.
Given, ẋ is the output:
y = x2 (as ẋ = x2 )

Position can be output, velocity can be output, or both can be output. It depends
on the engineering demand.

The converted state space system has 1 input (u), 1 output (y ), and 3 state
variables (x1 , x2 , x3 ).
9 / 29
STATE SPACE MODELING GENERAL FORM
In general, state space models have the following form:
A derivative for each of the state variables. And these derivatives are function of
other state variables and the inputs.
State Equations:

ẋ1 (t) = f1 (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

ẋ2 (t) = f2 (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

ẋn (t) = fn (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

Functions can be non-linear and time varying, e.g. sin(x1 ), x12 , at ∗ x1

The state variables themselves are not derivatives.

Output Equations(Algebraic equations, no derivative):

y1 (t) = g1 (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

y2 (t) = g2 (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

ym (t) = gm (x1 , x2 , . . . , xn , u1 , u2 , . . . , ur ; t)

10 / 29
STATE SPACE MODELING
For linear systems we can write general form of state space modeling in matrices
form.

11 / 29
STATE SPACE MODELING

12 / 29
STATE SPACE MODELING

Summary of previous lecture:

State space modelling is just repackaging our higher DE into a series of 1st order
DE. Solving these 1st order DE’s is much easier.
In the previous example, we have converted the physical equation to state space
form by simply changing the variables (which was blind mathematical
manipulation).

In this method we might run into these problems:

We may choose more state space variables than are required.
This will increase the number of equations.
The variables may not be linearly independent, which will make the solution
difficult or impossible.
The controller made based on this model won’t work.

To avoid these issues, there is a proper way to choose the state space variable,
which will guarantee the minimum number of variables.

13 / 29
STATE SPACE MODELING

Static system Dynamic system

Initial conditions do not affect the Initial conditions affect the
system. system/model.
Knowledge of the current input is Knowledge of the current input and
sufficient to get the current output. the past input is required to get the
current output.

A state basically captures the history of the system, it captures all the past inputs.

14 / 29
DEFINITION OF STATE VARIABLES

Q: What set of variables will capture all of the history that came before this
instance?
The state of a dynamic system is the smallest set of variables (called state
variables) such that knowledge (values) of these variables at t = t0 , together
with knowledge of the input for t ≥ t0 , completely determines the behavior of the
system for any time t ≥ t0 .
So we basically know the values of the past and the corresponding inputs, that
caused this state.
This information is sufficient for us to predict the response of the system with the
current input.
Min # of state variables = # of independent energy storage elements
The energy stored in the system at the current instant of time affects the
behavior of the system going forward. So if we know the stored energy we can
predict the future.

15 / 29
EXAMPLE

Equations of motion:
m1 ÿ + b(ẏ − ż) + k(y − z) = 0
m2 z̈ + b(ż − ẏ ) + k(z − y ) = u
Not in state space form.
In our previous blind way of choosing state variables, we will end up with these
state variables:
x1 = y , x2 = ẏ , x3 = z, x4 = ż

4 first order DEs But this is not the minimum number of state variables. We can
do with even less.
Solution: Find the number of energy storage elements.
16 / 29
EXAMPLE
Energy storage elements State Variables
Spring (stores elastic P.E) x1 = y − z
1 2
Mass, m1 (stores K.E 2 m 1 v1 ) x2 = ẏ
Mass, m2 (stores K.E 12 m2 v22 ) x3 = ż

Damper doesn’t store energy instead dissipates it.

Stored energy depends on deformation. Here the deformation is (y − z)

If you examine the state space variables, you can see that all initial conditions can
be defined by them.

17 / 29
EXAMPLE

18 / 29
EXAMPLE

19 / 29
EXCEPTIONS

No need to take two spring energy source k and k ′′

Min # of state variables ̸= # of energy storage elements
We can combine the two spring
No need to consider two mass m2 and m3 if they are not sliding against one
another. We can simply combine them.

General rule:
1 If the elements are constrained together, i.e. they are dependent

2 Some equations can not be expressed in terms of the minimum # of state

variables

For example, in the previous example, if y or z was the output we

wouldn’t be able to combine them as x1 = y − z.
20 / 29
BASIC EXAMPLE

21 / 29
STATE SPACE TO TRANSFER FUNCTION

22 / 29
EXAMPLE
Convert the state and output equations shown below to a transfer
function.

23 / 29
EXAMPLE PROBLEM

24 / 29
EXAMPLE PROBLEM

25 / 29
EXAMPLE PROBLEM

26 / 29
EXAMPLE PROBLEM

27 / 29
Post-Assessment

After the lecture, answer the following questions to assess your

understanding:
1 Explain the steps involved in converting a higher-order differential

equation into state-space form.

2 What are the advantages of using state-space models over classical

transfer function methods in modern control systems?

3 How do you determine the minimum number of state variables in a

system? Give an example based on the lecture.

4 What are the typical issues that might arise when choosing state

variables incorrectly, and how can these be avoided?

5 Given a system with two masses connected by springs, explain how

you would model it using state-space equations. What energy storage

elements would you identify?
6 Provide an example where you would need to combine two systems

(like two spring systems) due to their constraints. How does this
affect the state-space representation?

28 / 29
End of Topic

THANK YOU

ANY QUESTIONS?

29 / 29