STABILITY

In the design of a control system, the requirements are transient response, stability and
steady-state errors. Stability is the most important system specification and the maintenance
of system stability is one of the fundamental design objectives in control.

𝑐𝑐(𝑡𝑡) = 𝑐𝑐𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 (𝑡𝑡) + 𝑐𝑐𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 (𝑡𝑡)

A linear, time invariant system is stable if the natural response approaches zero as time
approaches infinity; the system is unstable if the natural response grows without bound as
time approaches infinity and the system is marginally stable if the natural response neither
decays now grows but remains constant or oscillates as time approaches infinity.
Another definition of stability is BIBO definition and it states that if a system is subjected to
a bounded input or disturbance and the response is bounded in magnitude for all time, the
system is said to be stable. A system is unstable if any bounded input yields an unbounded
output.
Concept of stability using a cone on a plane horizontal surface is shown in Figure 1.

Stable Neutral Unstable

Figure 1: Stability concept

Poles on left s-plane (stable) – Pure exponential decay or damped sinusoidal natural response.
Poles on right s-plane (unstable) – Purely exponentially increasing or exponentially
increasing sinusoidal natural response.
- Poles of multiplicity greater than 1 on the imaginary axis
𝐴𝐴𝑡𝑡 𝑛𝑛 cos⁡
(𝜔𝜔𝜔𝜔 + 𝜙𝜙)
Imaginary axis poles of multiplicity 1 and poles in left half plane (marginal stability).
Feedback is often used to stabilize a system that is unstable. In some cases, a sufficiently high
gain can actually destabilize a closed-loop system. Also, there are many systems for which
increasing the gain will not destabilize the system.
For an open-loop stable system, the introduction of feedback may destabilize it. For a general
(possibly nonlinear or time-varying) system, the problem of stability is usually one of the
most difficult to analyze. Fortunately, for linear, time-invariant (LTI) systems, the question of
stability is much easier to determine.
To determine the stability of a feedback control system, one could determine the roots of the
characteristic equation. The disadvantage of this is that system parameters must be assigned
numerical values, which makes it difficult to find the range values of a parameter that ensures

1
stability. Furthermore, in a feedback control system, even if we know the open-loop system
poles, we still need to determine the closed-loop poles. Thus to determine if a system is
stable, use one of the following methods: s-plane approach; frequency plane (jω) and time-
domain approach.

Figure 2: Stability in terms of system poles in s-plane

History of Feedback
Feedback was conceived by Harold Stephen Black (1891-1983). Black revolutionized
telecommunications by inventing systems that eliminated feedback distortion in telephone
calls and it took him six years to come up with the principles and equations on his way to
work aboard a ferry.

R(s) + C(s)
K(s)
- E(s)

H(s)

This resulted in feedback equation popularly known as Black’s formula given by:
𝐶𝐶(𝑠𝑠) 𝐾𝐾(𝑠𝑠)
= .
𝑅𝑅(𝑠𝑠) 1+𝐾𝐾(𝑠𝑠)𝐻𝐻(𝑠𝑠)

Negative feedback had wider applications than transcontinental and transatlantic

telecommunications, including industrial, military, and consumer electronics, weaponry,
analog computers, and biomedical devices such as pacemakers.

Polynomial and its General Properties

Let g(s) be a transfer function representation of an LTI system. Then g(s) can be written as
𝛽𝛽 𝑚𝑚 𝑠𝑠 𝑚𝑚 +𝛽𝛽 𝑚𝑚 −1 𝑠𝑠 𝑚𝑚 −1 +⋯+𝛽𝛽 1 𝑠𝑠+𝛽𝛽 0
𝑔𝑔(𝑠𝑠) =
𝛼𝛼 𝑛𝑛 𝑠𝑠 𝑛𝑛 +𝛼𝛼 𝑛𝑛 −1 𝑠𝑠 𝑛𝑛 −1 +⋯+𝛼𝛼 1 𝑠𝑠+𝛼𝛼 0

where the denominator polynomial is called the characteristic polynomial of g(s). The
equation 𝛼𝛼𝑛𝑛 𝑠𝑠 𝑛𝑛 + 𝛼𝛼𝑛𝑛−1 𝑠𝑠 𝑛𝑛−1 + ⋯ + 𝛼𝛼1 𝑠𝑠 + 𝛼𝛼0 = 0 is called the characteristic equation for g(s).
The roots of the characteristic equation are called the poles of g(s). The order of the system
g(s) is defined to be the degree of the characteristic polynomial.

2
A left half-plane root of a characteristic polynomial (pole) is called a stable root (pole). A
right half-plane root is called an unstable root. While an imaginary axis root is called a
marginally stable root.
Assume all polynomial coefficients are real. Consider the second order polynomial

𝑃𝑃(𝑠𝑠) = 𝑠𝑠 2 + 𝛼𝛼1 𝑠𝑠 + 𝛼𝛼0 = (𝑠𝑠 − 𝑝𝑝1 )(𝑠𝑠 − 𝑝𝑝2 ) = 𝑠𝑠 2 − (𝑝𝑝1 + 𝑝𝑝2 )𝑠𝑠 + 𝑝𝑝1 𝑝𝑝2

⇒ 𝛼𝛼1 = −(𝑝𝑝1 + 𝑝𝑝2 ), 𝛼𝛼0 = 𝑝𝑝1 𝑝𝑝2

Hence, 𝑝𝑝1 and 𝑝𝑝2 are either real, or else complex conjugate. If 𝑝𝑝1 and 𝑝𝑝2 are stable, we have
𝛼𝛼1 > 0 and 𝛼𝛼2 > 0. Consider next the third order polynomial

𝑃𝑃(𝑠𝑠) = 𝑠𝑠 3 + 𝛼𝛼2 𝑠𝑠 2 + 𝛼𝛼1 𝑠𝑠 + 𝛼𝛼0 = (𝑠𝑠 − 𝑝𝑝1 )(𝑠𝑠 − 𝑝𝑝2 )(𝑠𝑠 − 𝑝𝑝3 )

= 𝑠𝑠 3 − (𝑝𝑝1 + 𝑝𝑝2 + 𝑝𝑝3 )𝑠𝑠 2 + (𝑝𝑝1 𝑝𝑝2 + 𝑝𝑝2 𝑝𝑝3 + 𝑝𝑝1 𝑝𝑝3 )𝑠𝑠 − 𝑝𝑝1 𝑝𝑝2 𝑝𝑝3

⇒ 𝛼𝛼2 = −(𝑝𝑝1 + 𝑝𝑝2 + 𝑝𝑝3 ), 𝛼𝛼1 = (𝑝𝑝1 𝑝𝑝2 + 𝑝𝑝2 𝑝𝑝3 + 𝑝𝑝1 𝑝𝑝3 ), 𝛼𝛼0 = −𝑝𝑝1 𝑝𝑝2 𝑝𝑝3
Again, if all the roots are stable, all the coefficients will have the same sign.
Consider a general nth order polynomial

𝑃𝑃(𝑠𝑠) = 𝑠𝑠 𝑛𝑛 + 𝛼𝛼𝑛𝑛 −1 𝑠𝑠 𝑛𝑛−1 + ⋯ + 𝛼𝛼1 𝑠𝑠 + 𝛼𝛼0 = (𝑠𝑠 − 𝑝𝑝1 )(𝑠𝑠 − 𝑝𝑝2 ) … (𝑠𝑠 − 𝑝𝑝𝑛𝑛 ).
By analogy with the previous examples, we have:

𝛼𝛼𝑛𝑛−1 = − ∑𝑛𝑛𝑖𝑖=1 𝑝𝑝𝑖𝑖

𝛼𝛼𝑛𝑛−2 = ∑ 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 2 𝑎𝑎𝑎𝑎 𝑎𝑎 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

𝛼𝛼𝑛𝑛−3 = − ∑ 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 3 𝑎𝑎𝑎𝑎 𝑎𝑎 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡

⋮
𝛼𝛼0 = (−1)𝑛𝑛 ∏𝑛𝑛𝑖𝑖=1 𝑝𝑝𝑖𝑖

We conclude:
1. If all the roots are stable, all the polynomial coefficients will be positive.
2. If any coefficient is negative, at least one root is unstable.
3. If any coefficient is zero, not all roots are stable.

Routh-Hurwitz Stability Criterion

The stability of an LTI system requires that all the poles (roots of the characteristic equation)
must lie in the left half-plane. It is a necessary and sufficient condition for stability. It is a tool
to determine system stability by checking for RHP poles. In other words, it is an analytical
procedure for determining if all roots of a polynomial lie in the left half-plane, without
evaluating these roots.
Routh’s stability criterion tells us whether or not there are unstable roots in a polynomial
equation without actually solving for them. It applies to polynomial with only a finite number
of terms. It does not indicate where poles are located. This method may also be used to
establish the limiting values for a variable parameter beyond which a system would become

3
unstable. This method is powerful for design where the range of an unknown parameter can
be determined easily, which is difficult for a calculator.
The procedure for Routh’s stability criterion is as follows:
1. Write the polynomial in s in the following form:
𝑎𝑎4 𝑠𝑠 4 + 𝑎𝑎3 𝑠𝑠 3 + 𝑎𝑎2 𝑠𝑠 2 + 𝑎𝑎1 𝑠𝑠 + 𝑎𝑎0 = 0
where the coefficients are real quantities. We assume that 𝑎𝑎0 ≠ 0; that is any zero
roots has been removed. Or find the equivalent closed-loop system and the
characteristic equation is of the form above.

2. If any of the coefficients are zero or negative in the presence of at least one positive
coefficient, there is a root or roots that are imaginary or that have positive real parts.
This determines absolute stability or it is a necessary condition for stability.

3. If all coefficients are positive, then generate the Routh table (array) in which the first
two rows are formed from the characteristic polynomial coefficients.

𝑠𝑠 4 𝑎𝑎4 𝑎𝑎2 𝑎𝑎0

3
𝑠𝑠 𝑎𝑎3 𝑎𝑎1 0
𝑠𝑠 2 𝑏𝑏1 𝑏𝑏2 𝑏𝑏3
𝑠𝑠 1
𝑐𝑐1 𝑐𝑐2 𝑐𝑐3
0
𝑠𝑠 𝑑𝑑1 𝑑𝑑2 𝑑𝑑3
The remaining rows are formed as follows:
𝑎𝑎4 𝑎𝑎2 𝑎𝑎4 𝑎𝑎0 𝑎𝑎 0
�𝑎𝑎 𝑎𝑎 � �𝑎𝑎 � � 4 �
0 𝑎𝑎 0
𝑏𝑏1 = − 3 1
; 𝑏𝑏2 = − 3 ; 𝑏𝑏3 = − 3 = 0;
𝑎𝑎3 𝑎𝑎3 𝑎𝑎3
𝑎𝑎3 𝑎𝑎1 𝑎𝑎 0 𝑎𝑎 0
� 𝑏𝑏 𝑏𝑏 � � 3 � � 3 �
1 2 𝑏𝑏 1 0 𝑏𝑏1 0
𝑐𝑐1 = − ; 𝑐𝑐2 = − = 0; 𝑐𝑐3 = − = 0;
𝑎𝑎3 𝑎𝑎3 𝑎𝑎3
𝑏𝑏 𝑏𝑏2 𝑏𝑏 0 𝑏𝑏 0
� 1 � � 1 � � 1 �
𝑐𝑐1 0 𝑐𝑐1 0 𝑐𝑐1 0
𝑑𝑑1 = − ; 𝑑𝑑2 = − = 0; 𝑑𝑑3 = − = 0.
𝑎𝑎3 𝑎𝑎3 𝑎𝑎3
Once the Routh table is completed, we have the Routh stability criterion.
4. Interpreting the Routh table – The Routh-Hurwitz criterion states that the number of
roots of the characteristic equation with positive real parts (unstable roots) is equal to
the number of changes in sign of the first column of the array.

Example
1. Make a Routh table and tell how many roots of the following polynomial are in the right
half-plane and in the left half-plane.
𝑃𝑃(𝑠𝑠) = 3𝑠𝑠 7 + 9𝑠𝑠 6 + 6𝑠𝑠 5 + 4𝑠𝑠 4 + 7𝑠𝑠 3 + 8𝑠𝑠 2 + 2𝑠𝑠 + 6

Solution

First label the rows with the powers of s and form the first two rows of the Routh table
using the coefficients of the polynomial given by starting with the coefficient of highest
power of s and skipping every other power of s. Next, use the powers of s skipped in the
previous row. Finally, calculate the other coefficients of the table to complete it using
determinants.
4
 𝑠𝑠 7 3 6 7 2
𝑠𝑠 6 9 4 8 6
𝑠𝑠 5 4.67 4.33 0 0
𝑠𝑠 4 −4.34 8 6 0
𝑠𝑠 3 12.94 6.46 0 0
𝑠𝑠 2 10.17 6 0 0
𝑠𝑠1 −1.17 0 0 0
𝑠𝑠 0 6 0 0 0

From the Routh table above, we can see that there are four sign changes from s5 (4.67) to s4(-
4.34), s4 (-4.34) to s3(12.94), s2 (10.17) to s1(-1.17), and s1 (-1.17) to s0(6). Hence, there are
four poles on the right half-plane and three poles on the left half-plane. The system is
unstable.

Special Cases of Routh-Hurwitz Criterion

There are two special cases namely:
1. Zero Only in the First Column – This is an indication of instability or at best marginal
stability. The array could not be continued by the normal method of construction,
since this would involve division by zero. If the root distribution is required, an
alternate technique must be adopted.
• Epsilon Method: The procedure is modified by replacing that first element
(zero) with a small number ε and the value of 𝜀𝜀 is allowed to approach zero
from either the positive or the negative side (i.e.|𝜀𝜀| ≪ 1), after which the
signs of the entries in the first column can be determined.

Test the stability of a system having the characteristic equation

𝑠𝑠 5 + 2𝑠𝑠 4 + 2𝑠𝑠 3 + 4𝑠𝑠 2 + 𝑠𝑠 + 1 = 0
Solution
Form the Routh table as
𝑠𝑠 5 1 2 1
4
𝑠𝑠 2 4 1
𝑠𝑠 3 0 → 𝜀𝜀 0.5 0
𝑠𝑠 2 −1/𝜀𝜀 1 0
𝑠𝑠1 0.5 0 0
𝑠𝑠 0 1 0 0
Since there are two sign changes in the first column, the system is unstable
with two positive (right half-plane) roots. Solving for the roots of the
characteristic equation gives −0.09 ± 𝑗𝑗0.533, +0.069 ± 1.274, and -1.957.

• Reverse Coefficients Method: This is another method used for a zero

appearing in the first column. It involves taking the reciprocal of the original
roots and this new polynomial will not have a zero in the first column. It is
computationally easier than epsilon method.
The new polynomial is simply the original polynomial with its coefficients
written in reverse order.
𝑠𝑠 𝑛𝑛 + 𝑎𝑎𝑛𝑛−1 𝑠𝑠 𝑛𝑛−1 + ⋯ + 𝑎𝑎1 𝑠𝑠 + 𝑎𝑎0 = 0
Replace s with 1/d to obtain

5
 1 𝑛𝑛 1 𝑛𝑛−1 1
�𝑑𝑑 � + 𝑎𝑎𝑛𝑛−1 �𝑑𝑑 � + ⋯ + 𝑎𝑎1 �𝑑𝑑 � + 𝑎𝑎0 = 0
Factorizing we get
1 𝑛𝑛 1 −1 1 1−𝑛𝑛 1 −𝑛𝑛
�𝑑𝑑 � [1 + 𝑎𝑎𝑛𝑛−1 �𝑑𝑑 � + ⋯ + 𝑎𝑎1 �𝑑𝑑 � + 𝑎𝑎0 �𝑑𝑑 � =0
1 𝑛𝑛
�𝑑𝑑 � [1 + 𝑎𝑎𝑛𝑛−1 𝑑𝑑 + ⋯ + 𝑎𝑎1 𝑑𝑑𝑛𝑛−1 + 𝑎𝑎0 𝑑𝑑𝑛𝑛 ] = 0

Determine the stability of the characteristic equation given as

𝑠𝑠 5 + 2𝑠𝑠 4 + 3𝑠𝑠 3 + 6𝑠𝑠 2 + 5𝑠𝑠 + 3 = 0

Solution
Write the reverse order of the characteristic equation to get
𝐷𝐷(𝑠𝑠) = 3𝑠𝑠 5 + 5𝑠𝑠 4 + 6𝑠𝑠 3 + 3𝑠𝑠 2 + 2𝑠𝑠 + 1 = 0
Form the Routh table as
𝑠𝑠 5 3 6 2
𝑠𝑠 4 5 3 1
3
𝑠𝑠 4.2 1.4 0
𝑠𝑠 2 1.33 1 0
𝑠𝑠 1 −1.75 0 0
𝑠𝑠 0 1 0 0

The system has two sign changes and hence, there are two poles on the right
half-plane and is therefore unstable.

2. Entire Row is Zero – This condition occurs when the polynomial contains
singularities that are symmetrically located about the origin (both real and the
imaginary axes) of the s-plane. The roots are of the form 𝑠𝑠 = ±𝜎𝜎; 𝑠𝑠 = ±𝑗𝑗𝑗𝑗 𝑜𝑜𝑜𝑜 𝑠𝑠 =
−𝜎𝜎 ± 𝑗𝑗𝑗𝑗 𝑎𝑎𝑎𝑎𝑎𝑎 𝑠𝑠 = +𝜎𝜎 ± 𝑗𝑗𝑗𝑗. This problem is circumvented by utilizing auxiliary
polynomial which immediately precedes the zero entry in the Routh array. The order
of the auxiliary polynomial is always even and indicates the number of symmetrical
root pairs. An even or auxiliary polynomial is one in which the powers of s are even
integers or zero only: 𝑠𝑠 0 , 𝑠𝑠 2 , 𝑠𝑠 4 …

Determine that stability of the characteristic equation

𝑠𝑠 5 + 7𝑠𝑠 4 + 6𝑠𝑠 3 + 42𝑠𝑠 2 + 8𝑠𝑠 + 52 = 0

Solution
Form the Routh table as
𝑠𝑠 5 1 6 8
𝑠𝑠 4 7→1 42 → 6 56 → 8
𝑠𝑠 3 0 → 4 → 1 0 → 12 → 3 0 → 0 → 0
𝑠𝑠 2 3 8 0
𝑠𝑠 1 1/3 0 0
𝑠𝑠 0 8 0 0
The second and third rows can be divided through by seven (7) and four (4)
respectively.
Form the auxiliary equation from the 𝑠𝑠 4 row as
𝑃𝑃(𝑠𝑠) = 𝑠𝑠 4 + 6𝑠𝑠 2 + 8
𝑑𝑑𝑑𝑑 (𝑠𝑠)
𝑑𝑑𝑑𝑑
= 4𝑠𝑠 3 + 12𝑠𝑠 (replace 𝑠𝑠 3 row with this).

6
 All entries in the first column are positive i.e. no sign change. Hence, the system is
stable and there are no right half-plane poles.

Example

Use the Routh-Hurwitz criterion to find how many poles of the following closed-loop
system T(s) are in the rhp, lhp and on the jω axis.
𝑠𝑠 3 +7𝑠𝑠 2 −21𝑠𝑠+10
𝑇𝑇(𝑠𝑠) = 𝑠𝑠 6 +𝑠𝑠 5 −6𝑠𝑠 4 +0𝑠𝑠 3 −𝑠𝑠 2 −𝑠𝑠+6

Solution
The Routh table is
𝑠𝑠 6 1 −6 −1 6
𝑠𝑠 5 1 0 −1 0
𝑠𝑠 4 −6 0 6 0
𝑠𝑠 3 0 → −24 0 → 0 0 → 0 0 → 0
𝑠𝑠 2 0 → 𝜀𝜀 6 0 0
𝑠𝑠1 144/𝜀𝜀 0 0 0
𝑠𝑠 0 6 0 0 0

Complete row of zero is found in s3, and then the auxiliary equation is formed using s4
row as
𝑃𝑃(𝑠𝑠) = −6𝑠𝑠 4 + 0𝑠𝑠 2 + 6
𝑑𝑑𝑑𝑑 (𝑠𝑠)
𝑑𝑑𝑑𝑑
= −24𝑠𝑠 3 − 0𝑠𝑠 (replace 𝑠𝑠 3 row with this).

Next, replace the zero in the first row with epsilon and continue with the computation.

From the Routh table, there is one sign change after the auxiliary equation (i.e. from -
24 to epsilon). Hence, the even polynomial has one left half-plane pole and one right
half-plane pole due to symmetry. The other two poles are on the jω axis. From the top
to the even polynomial there is one sign change from 1 to -6, hence one left half-plane
pole and one right half-plane pole.

The total for the system is two left half-plane poles, two right half-plane poles and
two imaginary (jω) axis poles.

Stability for First and Second Order Systems

For a first order system with characteristic polynomial

𝑃𝑃(𝑠𝑠) = 𝑎𝑎1 𝑠𝑠 + 𝑎𝑎0

It is clear that stability requires both coefficients to have the same sign.
A second order system with characteristic polynomial

𝑃𝑃(𝑠𝑠) = 𝑎𝑎2 𝑠𝑠 2 + 𝑎𝑎1 𝑠𝑠 + 𝑎𝑎0

Has a Routh table

7
 𝑠𝑠 2 𝑎𝑎2 𝑎𝑎0
𝑠𝑠1 𝑎𝑎1 0
𝑠𝑠 0 𝑎𝑎0 0
A necessary and sufficient condition for stability is that all coefficients have the same
sign.

Thus, for 1st and 2nd order systems, stability properties can be determined by
inspection. For higher order systems, we can only conclude that a necessary condition
for stability is that all the coefficients have the same sign.

Applications in Feedback Design

Consider the feedback system involving a plant G(s) and compensator K(s).
+ C(s)
R(s) K(s) G(s)
-

The closed-loop transfer function T(s) is

𝐺𝐺(𝑠𝑠)𝐾𝐾(𝑠𝑠)
𝑇𝑇(𝑠𝑠) = 1+𝐺𝐺(𝑠𝑠)𝐾𝐾(𝑠𝑠)

The closed-loop poles are given by the roots of the characteristic equation
1 + 𝐺𝐺(𝑠𝑠)𝐾𝐾(𝑠𝑠) = 0

Suppose we write
𝑁𝑁𝑔𝑔 (𝑠𝑠) 𝑁𝑁 (𝑠𝑠)
𝐺𝐺(𝑠𝑠) = ; 𝐾𝐾(𝑠𝑠) = 𝐷𝐷 𝑘𝑘 (𝑠𝑠)
𝐷𝐷𝑔𝑔 (𝑠𝑠) 𝑘𝑘
where 𝑁𝑁𝑔𝑔 (𝑠𝑠), 𝐷𝐷𝑔𝑔 (𝑠𝑠), 𝑁𝑁𝑘𝑘 (𝑠𝑠)𝑎𝑎𝑎𝑎𝑎𝑎 𝐷𝐷𝑘𝑘 (𝑠𝑠) are polynomials.

Then, the closed-loop transfer function is given by

𝑁𝑁 𝑔𝑔 (𝑠𝑠)𝑁𝑁 𝑘𝑘 (𝑠𝑠)
𝐷𝐷 𝑔𝑔 (𝑠𝑠)𝐷𝐷 𝑘𝑘 (𝑠𝑠) 𝑁𝑁𝑔𝑔 (𝑠𝑠)𝑁𝑁𝑘𝑘 (𝑠𝑠)
𝑇𝑇(𝑠𝑠) = 𝑁𝑁 𝑔𝑔 (𝑠𝑠)𝑁𝑁 𝑘𝑘 (𝑠𝑠) =
1+ 𝑁𝑁𝑔𝑔 (𝑠𝑠)𝑁𝑁𝑘𝑘 (𝑠𝑠)+𝐷𝐷𝑔𝑔 (𝑠𝑠)𝐷𝐷𝑘𝑘 (𝑠𝑠)
𝐷𝐷 𝑔𝑔 (𝑠𝑠)𝐷𝐷 𝑘𝑘 (𝑠𝑠)

The poles of the closed-loop system are also given by the roots of the characteristic
polynomial
𝑁𝑁𝑔𝑔 (𝑠𝑠)𝑁𝑁𝑘𝑘 (𝑠𝑠) + 𝐷𝐷𝑔𝑔 (𝑠𝑠)𝐷𝐷𝑘𝑘 (𝑠𝑠) = 0

The poles of the open-loop system G(s) are the roots of its characteristic polynomial
𝐷𝐷𝑔𝑔 (𝑠𝑠) = 0
which are generally different from closed-loop poles.

A fundamental design objective in control is the stabilization of unstable systems. The

Routh-Hurwitz stability criterion can be used as an aid in feedback design.

8
 Examples
1. For a unity feedback system with the forward transfer function
𝐾𝐾(𝑠𝑠+20)
𝐺𝐺(𝑠𝑠) = 𝑠𝑠(𝑠𝑠+2)(𝑠𝑠+3)
find the range of K to make the system stable.

Solution

Find the closed-loop transfer function as

𝐺𝐺(𝑠𝑠) 𝐾𝐾(𝑠𝑠+20)
𝑇𝑇(𝑠𝑠) = 1+𝐺𝐺(𝑠𝑠) = 𝑠𝑠 3 +5𝑠𝑠 2 +(6+𝐾𝐾)𝑠𝑠+20𝐾𝐾

The characteristic equation is given by

𝑠𝑠 3 + 5𝑠𝑠 2 + (6 + 𝐾𝐾)𝑠𝑠 + 20𝐾𝐾 = 0
The Routh table is
𝑠𝑠 3 1 6 + 𝐾𝐾
𝑠𝑠 2 5 20𝐾𝐾
1 6 − 3𝐾𝐾 0
𝑠𝑠
0 20𝐾𝐾 0
𝑠𝑠
From the table, K=2 from 6-3K=0, hence the value of K will have to be less than
2. Also, K should be greater than zero. Thus, for the system to be stable we need
0 < 𝐾𝐾 < 2.

2. Let G(s) be given by

𝐾𝐾
𝐺𝐺(𝑠𝑠) = 𝑠𝑠 3 +5𝑠𝑠 2 +2𝑠𝑠−8 .

Find the range of values of K (a constant controller) for which the closed-loop is
stable.

Solution
Find the closed-loop transfer function as
𝐾𝐾
𝑇𝑇(𝑠𝑠) = 𝑠𝑠 3 +5𝑠𝑠 2 +2𝑠𝑠−8+𝐾𝐾

The closed-loop poles are the roots of the characteristic equation

𝑠𝑠 3 + 5𝑠𝑠 2 + 2𝑠𝑠 + (𝐾𝐾 − 8) = 0
The Routh table is
𝑠𝑠 3 1 2
2
𝑠𝑠 5 (𝐾𝐾 − 8)
1
𝑠𝑠 0.2(18 − 𝐾𝐾) 0
𝑠𝑠 0 𝐾𝐾 − 8 0
From the table, K=18 from 0.2(18-K), hence the value of K will have to be less
than 18. Also, K should be greater than 8. Thus, for the system to be stable we
need 8 < 𝐾𝐾 < 18.

Stability in State Space

We know that the poles of a system are the eigenvalues of the system matrix A.

9
det(𝑠𝑠𝑠𝑠 − 𝐴𝐴) = 0 , determines the poles of the system and is known as the characteristic
equation.

Let λ be the eigenvalues of the system and x the eigenvectors, then

𝐴𝐴𝐴𝐴 = 𝜆𝜆𝜆𝜆 ⇒ (𝜆𝜆𝜆𝜆 − 𝐴𝐴)𝑥𝑥 = 0

𝑎𝑎𝑎𝑎𝑎𝑎 (𝜆𝜆𝜆𝜆 −𝐴𝐴)
𝑥𝑥 = (𝜆𝜆𝜆𝜆 − 𝐴𝐴)−1 0 = 𝑑𝑑𝑑𝑑𝑑𝑑 (𝜆𝜆𝜆𝜆 −𝐴𝐴)
0

The stability of a system represented in state-space can be determined by finding the

eigenvalues of the system matrix A and determining their locations on the s-plane.

Examples

1. Given the system

0 3 1 10
𝑥𝑥̇ = � 2 8 1 � 𝑥𝑥 + � 0 � 𝑢𝑢; 𝑦𝑦 = [1 0 0]𝑥𝑥
−10 −5 −2 0
find out how many poles are in the left half-plane, in the right half-plane and on jω-
axis.

Solution
𝑠𝑠 −3 −1
𝑠𝑠𝑠𝑠 − 𝐴𝐴 = �−2 𝑠𝑠 − 8 1 �
10 5 𝑠𝑠 + 2
det(𝑠𝑠𝑠𝑠 − 𝐴𝐴) = 𝑠𝑠 3 − 6𝑠𝑠 2 − 7𝑠𝑠 − 52 = 0
The Routh table is
𝑠𝑠 3 1 −7
𝑠𝑠 2 −6 → −3 −52 → −26
47
𝑠𝑠1 − → −1 0
𝑠𝑠 0 3 0
−26
Since there is only one sign change in the first column, the system is unstable and has
one right half-plane pole and two left half-plane poles.

2. For the state-space system

2 1 1 0
𝑥𝑥̇ = � 1 7 1 � 𝑥𝑥 + �0� 𝑟𝑟; 𝑦𝑦 = [0 1 0]𝑥𝑥
−3 4 −5 1
find out how many poles are in the left half-plane, in the right half-plane and on the
jω-axis.

Solution
𝑠𝑠 − 2 −1 −1
𝑠𝑠𝑠𝑠 − 𝐴𝐴 = � −1 𝑠𝑠 − 7 −1 �
3 −4 𝑠𝑠 + 5
3
det(𝑠𝑠𝑠𝑠 − 𝐴𝐴) = 𝑠𝑠 − 64 − 33𝑠𝑠 + 55 = 0

10
The Routh table is
𝑠𝑠 3 1 −33
2
𝑠𝑠 −4 55
𝑠𝑠1 −19.25 0
0 55 0
𝑠𝑠
Since there are two sin changes in the first column, the system is unstable and has two
right half-plane poles and one left half-plane pole.

Patience E. Orukpe

11