AER 372

1) Background
2) Basic Feedback
System
3) Root Locus
Determination
Control Systems
4) Case Study
5) Design for
Dynamic
Compensation Chapter 5
6) Design
Examples Root Locus Design Method
7) Root-locus
Method
Extensions

Prof. M.R. Emami

Winter 2025

1
 M.R. Emami, 2025
 Outline
1) Background
2) Basic Feedback  Background
System
3) Root Locus
Determination  Basic Feedback System
4) Case Study
5) Design for
Dynamic
 Root Locus Determination
Compensation
6) Design
Examples
 Case Study
7) Root-locus
Method
Extensions
 Design for Dynamic Compensation
 Design Examples
 Root-locus Method Extensions

2
 M.R. Emami, 2025
 Background
What is Root-locus Method?

1) Background
 A simple (graphical) method, consisting of several rules, for finding the
2) Basic Feedback
System
locus of the roots of a system’s characteristic equation as a result of
3) Root Locus
varying some of its parameters.
Determination
4) Case Study  Developed in 1948 by the American engineer and control
5) Design for theorist, Walter R. Evans (1920-1999). He published the
Dynamic
Compensation book “Control System Dynamics” (McGraw-Hill) in 1954.
6) Design
Examples
7) Root-locus  The method is generic, and can be used for studying the roots of any
Method polynomial with respect to any (real) parameter that enters the equation linearly.
Extensions

 In Controls, the method is used specifically to study the effect of the control
gain of closed-loop system on the roots of characteristic equation, thus on the
system’s dynamic behaviour.

3
 M.R. Emami, 2025
 Basic Feedback System
 Closed-loop System Transfer Function:
1) Background 𝑌 𝑠 𝐷 𝑠 𝐺 𝑠
2) Basic Feedback =𝒯 𝑠 =
System 𝑅 𝑠 1+𝐷 𝑠 𝐺 𝑠 𝐻 𝑠
3) Root Locus
Determination
4) Case Study
 Characteristic Equation: 1+𝐷 𝑠 𝐺 𝑠 𝐻 𝑠 =0
5) Design for
Dynamic
𝑎 𝑠 + 𝐾𝑏 𝑠 = 0
Compensation
𝑏 𝑠
6) Design 1+𝐾 =0
Examples 𝑎 𝑠
7) Root-locus
Method 1 + 𝐾𝐿 𝑠 = 0 Root Locus (Evans) Form
Extensions
𝑏 𝑠 1
𝐿 𝑠 = =−
𝑎 𝑠 𝐾
 In most cases, 𝐿 𝑠 is proportional to the open-loop transfer function
𝐷 𝑠 𝐺 𝑠 𝐻 𝑠 of the closed-loop system, and 𝐾 is the gain of the controller.

 Plot the locus of all possible roots of characteristic equation as 𝐾 varies from
zero to infinity, in order to select the best value of controller gain.

 Also, study from the root locus graph the effect of potentially additional poles
and zeros included in 𝐷 𝑠 to modify system dynamics. 4
 M.R. Emami, 2025
 Basic Feedback System
 Closed-loop System Transfer Function:

1) Background 𝑌 𝑠 𝐷 𝑠 𝐺 𝑠
=𝒯 𝑠 =
2) Basic Feedback 𝑅 𝑠 1+𝐷 𝑠 𝐺 𝑠 𝐻 𝑠
System
3) Root Locus
Determination 𝑏 𝑠 1
4) Case Study
 Characteristic Equation: 𝐿 𝑠 = =−
𝑎 𝑠 𝐾
5) Design for
Dynamic 𝑏 𝑠 =𝑠 +𝑏 𝑠 + ⋯+ 𝑏 = 𝑠 − 𝑧 𝑠−𝑧 ×⋯× 𝑠 − 𝑧 = 𝑠−𝑧
Compensation
6) Design
Examples
7) Root-locus 𝑎 𝑠 = 𝑠 +𝑎 𝑠 +⋯+𝑎 = 𝑠 − 𝑝 𝑠−𝑝 ×⋯× 𝑠 − 𝑝 = 𝑠−𝑝
Method
Extensions
𝑎 𝑠 + 𝐾𝑏 𝑠 = 𝑠 − 𝑟 𝑠 − 𝑟 +⋯+ 𝑠 − 𝑟 𝑛≥𝑚

 Roots of 𝑏 𝑠 are open-loop zeros 𝑧 , roots of 𝑎 𝑠 are open-loop poles 𝑝 , and

roots of 𝑎 𝑠 + 𝐾𝑏 𝑠 are closed-loop system poles 𝑟 .
 The order of open-loop denominator polynomial 𝑛 is bigger than (or equal to) that
of open-loop numerator polynomial 𝑚, since 𝐷 𝑠 𝐺 𝑠 𝐻 𝑠 should present a
physical system.
 NOTE: Open-loop zeros and poles are different from those of the closed-loop
transfer function. Here, roots of the closed-loop system characteristic equation
𝑎 𝑠 + 𝐾𝑏 𝑠 , i.e., system poles, are defined as 𝑟 . 5
 M.R. Emami, 2025
 Basic Feedback System
Example: Assume 𝑐 = 1, 𝑅 𝑠 + 𝑈 𝑠 1 𝑌 𝑠
𝐾
1) Background
_ 𝑠 𝑠+𝑐
2) Basic Feedback
Plot the root locus w.r.t. 𝐾.
System
3) Root Locus 𝑌 𝑠 1 𝑌 𝑠 𝐾 𝐾
Determination =𝐺 𝑠 = ; 𝑅 𝑠
=𝒯 𝑠 = =
𝑠 + 𝑠 + 𝐾 𝑎 𝑠 + 𝐾𝑏 𝑠
𝑈 𝑠 𝑠 𝑠+1
4) Case Study
5) Design for 𝑏 𝑠 = 1; 𝑚 = 0; 𝑧 = 𝑒𝑚𝑝𝑡𝑦 ; 1 Open-loop
Dynamic 𝐿 𝑠 =
Compensation
𝑎 𝑠 = 𝑠 + 𝑠; 𝑛 = 2; 𝑝 = 0, −1; 𝑠 𝑠+1 TF
6) Design
Examples Characteristic 𝑎 𝑠 + 𝐾𝑏 𝑠 = 𝑠 + 𝑠 + 𝐾 = 0
7) Root-locus Equation
Method
Extensions
1 1 − 4𝐾
𝑟 ,𝑟 = − ±
2 2
𝐾=0 𝑟 , 𝑟 = −1, 0
𝐾 = 1⁄ 4 𝑟 , 𝑟 = − 1⁄ 2

𝜁 = 0.5 𝑟 , 𝑟 = − 1⁄2 ± 𝑗 3⁄2 𝐾=1

 2 open-loop poles (start points, corresponding

to 𝐾 = 0), 2 branches, 1 breakaway point.
 Choose 𝐾 to have poles at desired locations
 M.R. Emami, 2025
(desired system behaviour, e.g., w.r.t. damping.) 6
 Basic Feedback System
Example: Assume 𝐾 = 1, 𝑅 𝑠 + 𝑈 𝑠 1 𝑌 𝑠
𝐾
1) Background
_ 𝑠 𝑠+𝑐
2) Basic Feedback
Plot the root locus w.r.t. 𝑐.
System
3) Root Locus 𝑌 𝑠 1 𝑌 𝑠 1 1
Determination =𝐺 𝑠 = ; =𝒯 𝑠 = =
𝑈 𝑠 𝑠 𝑠+𝑐 𝑅 𝑠 𝑠 + 1 + 𝑐𝑠 𝑎 𝑠 + 𝑐𝑏 𝑠
4) Case Study
5) Design for 𝑏 𝑠 = 𝑠; 𝑚 = 1; 𝑧 = 0; 𝑠
Dynamic 𝐿 𝑠 =
Compensation
𝑎 𝑠 = 𝑠 + 1; 𝑛 = 2; 𝑝 = +𝑗, −𝑗; 𝑠 +1
6) Design
Examples
𝑎 𝑠 + 𝑐𝑏 𝑠 = 𝑠 + 𝑐𝑠 + 1 = 0
7) Root-locus
Method 𝑐 𝑐 −4
Extensions 𝑟 ,𝑟 = − ±
2 2
𝑐=0 𝑟 , 𝑟 = −𝑗, +𝑗
𝑐=2 𝑟 , 𝑟 = −1

𝜁 = 0.5 𝑟 , 𝑟 = − 1⁄2 ± 𝑗 3⁄2 𝑐=1

 2 poles (start points, corresponding to

𝑐 = 0), 2 branches, 1 break-in point.
 Choose c to have poles at desired locations
 M.R. Emami, 2025
(desired system behaviour, e.g., w.r.t. damping). 7
 Root Locus Determination
Complex Functions (Reminder)
 A complex variable (number) 𝑠 can be shown in the s-plane by either Cartesian
1) Background or polar coordinates:
2) Basic Feedback
Imaginary Axis (Im)
𝑠 = 𝜎 + 𝑗𝜔 = 𝑟𝑒
System
s    j
3) Root Locus
Determination
Magnitude: r  s   
2 2

4) Case Study
5) Design for
Phase:  
  s  tan 1   r

Dynamic 𝑒 +𝑒 Real
Compensation cos 𝜑 = 2 ; e j  cos   j sin   Axis
6) Design 𝑒 −𝑒 (Re)
Examples sin 𝜑 = 2𝑗
7) Root-locus
Method
Extensions  Similarly, a complex function of 𝑠, 𝐿 𝑠 , can be
represented in Cartesian or polar form.
𝐿 𝑠 Function Magnitude
Polar Form: 𝐿 𝑠 = 𝐿 𝑠 𝑒 ∠ ;
∠𝐿 𝑠 Function Phase
𝑏 𝑠 ∏ 𝑠−𝑧 Im s
 For a rational complex function: 𝐿 𝑠 = =
𝑎 𝑠 ∏ 𝑠−𝑝
𝑏 𝑠 ∏ 𝑠−𝑧 ∠ 𝑠−𝑎
𝐿 𝑠 = = a
𝑎 𝑠 ∏ 𝑠−𝑝
Re
∠𝐿 𝑠 = ∠𝑏 𝑠 − ∠𝑎 𝑠 = ∠ 𝑠−𝑧 − ∠ 𝑠−𝑝
8
 M.R. Emami, 2025
 Root Locus Determination
 Definition 1:
The root locus is the set of all possible values of 𝑠 for which the characteristic
1) Background
equation 1 + 𝐾𝐿 𝑠 = 0 holds, as the real parameter 𝐾 varies from 0 to ∞ (or −∞).
2) Basic Feedback
In controls, the characteristic equation is typically for the closed-loop system,
System so the roots on the locus are the system poles. 𝑛≥𝑚
3) Root Locus
1 𝑏 𝑠 ∏ 𝑠−𝑧
Determination  Considering the characteristic equation: 𝐿 𝑠 = − 𝐾 ; 𝐿 𝑠 = =
4) Case Study
𝑎 𝑠 ∏ 𝑠−𝑝
 If 𝐾 is real and positive, then 𝐿 𝑠 must be real
5) Design for
Dynamic
and negative. In polar form, the phase of 𝐿 𝑠 must be 180° (positive locus).
Compensation  If 𝐾 is real and negative, then 𝐿 𝑠 must be real and positive. In polar form,
6) Design the phase of 𝐿 𝑠 must be 0° (negative locus).
Examples
7) Root-locus  Definition 2:
Method
Extensions The root locus is the set of points in the s-
plane where the phase of 𝐿 𝑠 is 180° for
positive locus, or 0° for negative locus.
 If the angle to a test point 𝑠 from an
open-loop zero 𝑧 is 𝜓 and the angle to
the test point from an open-loop pole 𝑝
is 𝜙 , then in order for the test point to be
on the root locus (𝑙 is a positive integer):
positive locus: 𝜓 − 𝜙 = 180° + 360° 𝑙 − 1

negative locus: 𝜓 − 𝜙 = 0° + 360° 𝑙 − 1 9

 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus
Example:
𝑠+1
1) Background 𝐿 𝑠 =
2) Basic Feedback
𝑠 𝑠+5 𝑠+2 +4
System
Test point:
3) Root Locus
Determination 𝑠 = −1 + 𝑗2
4) Case Study
5) Design for Does 𝑠 lie on the root locus of 𝐿 𝑠 for
Dynamic some values of 𝐾 in the characteristic
Compensation
equation 1 + 𝐾𝐿 𝑠 = 0 (𝐾 ≥ 0)?
6) Design
Examples
For 𝑠 to be on the locus we must have:
7) Root-locus
Method
Extensions ∠𝐿 𝑠 = 180° + 360° 𝑙 − 1
∠𝐿 𝑠 = ∠ 𝑠 + 1 − ∠𝑠 − ∠ 𝑠 + 2 − 𝑗2 − ∠ 𝑠 + 2 + 𝑗2 − ∠ 𝑠 + 5

= 𝜓 −𝜙 − 𝜙 − 𝜙 − 𝜙 = 90° − 116.6° − 0° − 76° − 26.6°

= −129.2° ≠ 180° + 360° 𝑙 − 1

 To determine the points of the root locus without having to compute for their
total phase, a set of rules are suggested, which are collectively referred to as
the root-locus method.
10
 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus 𝑛≥𝑚
 Rule 1:
𝑏 𝑠 ∏ 𝑠−𝑧
There are 𝑛 branches of the locus starting at the poles 𝐿 𝑠 = =
of 𝐿 𝑠 (open-loop poles), and 𝑚 (≤ 𝑛) of these branches 𝑎 𝑠 ∏ 𝑠−𝑝
1) Background
2) Basic Feedback end at the zeros of 𝐿 𝑠 (open-loop zeros), and the rest goes to infinity.
System
 The start of the locus is for when 𝐾 = 0, reducing the system
3) Root Locus
Determination
characteristic equation to 𝑎 𝑠 = 0, whose roots are the open-loop poles.
4) Case Study  If 𝐾 approaches ∞, then 𝑠 must be such that either 𝑏 𝑠 = 0 or 𝑠 → ∞. Since
5) Design for there are 𝑚 zeros where 𝑏 𝑠 = 0, 𝑚 branches can end at these points.
Dynamic
Compensation  Rule 2:
6) Design The segments of the locus on the real axis are always
Examples to the left of an odd number of real poles and zeros.
7) Root-locus
Method  For any point on the real axis, the phase angles of
Extensions conjugate poles or zeros cancel out each other.
With respect to poles and zeros on real axis, the
point must be to the left of an odd number of them,
to have a total phase angle of 180° + 360° 𝑙 − 1 .
𝑛−𝑚 =2
 Rule 3: 𝑛−𝑚 =1 𝑛−𝑚 =3 𝑛−𝑚 =4

If 𝑛 > 𝑚 , when 𝐾 → ∞, those

𝑛 − 𝑚 branches that must go
to infinity (because 𝑠 → ∞) are
asymptotic to lines radiating out from real axis at the point 𝑠 = 𝛼 at angles 𝜙 .
∑𝑝 − ∑𝑧 180° + 360° 𝑙 − 1
𝛼= ; 𝜙 = , 𝑙 = 1,2, ⋯ , 𝑛 − 𝑚 11
 M.R. Emami, 2025 𝑛−𝑚 𝑛 − 𝑚
 Root Locus Determination
Positive Root Locus
Example:
𝑅 𝑠 + 𝑈 𝑠 𝑠+1 𝑌 𝑠
Characteristic Equation: 𝐾
1) Background _ 𝑠 𝑠+2 𝑠+3
2) Basic Feedback 𝑠+1
System 1+𝐾 =0
𝑠 𝑠+2 𝑠+3
3) Root Locus
Determination 𝑠+1 𝑏 𝑠 = 𝑠 + 1; 𝑚 = 1; 𝑧 = −1;
4) Case Study 𝐿 𝑠 =
𝑠 𝑠+2 𝑠+3 𝑎 𝑠 =𝑠 𝑠+2 𝑠+3 ; 𝑝 = 0, −2, −3;
5) Design for 𝑛 = 3;
Dynamic
Compensation
6) Design
Examples
Rules 1 , 2
7) Root-locus
Method
Extensions

asymptotes
Rule 3: (not root locus)
∑𝑝 − ∑𝑧 0−2−3+1
𝛼= = = −2
𝑛−𝑚 3−1

180° + 360° 𝑙 − 1
𝜙 = = 90° + 180° 𝑙 − 1 = 90° , 270°
𝑛−𝑚 12
 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus
 Rule 4:
 The angle of departure of a branch of the locus from a single pole is identified
1) Background by: 𝜙 = 𝜓 − 𝜙 − 180°
2) Basic Feedback
System
∑ 𝜙 : Sum of the angles from the remaining poles to the single pole
3) Root Locus
Determination ∑ 𝜓 : Sum of the angles from all the zeros to the single pole
4) Case Study
5) Design for The angles of departure for repeated poles with multiplicity 𝑞 is given by:
Dynamic
Compensation 𝑞𝜙 , = 𝜓 − 𝜙 − 180° − 360° 𝑙 − 1 , 𝑙 = 1,2, ⋯ , 𝑞
6) Design ,
Examples
Note: For 𝑞 repeated poles, there are 𝑞 branches departing from these poles.
7) Root-locus
Method
Extensions  The angle(s) of arrival of a branch at a zero with multiplicity 𝑞 is obtained by:

𝑞𝜓 , = 𝜙 − 𝜓 + 180° + 360° 𝑙 − 1 , 𝑙 = 1,2, ⋯ , 𝑞

,
∑ 𝜙 : Sum of the angles from zero to all the poles
1
∑ 𝜓 : Sum of the angles from zero to other zeros

Note: For 𝑞 repeated zeros, there are 𝑞 branches

arriving at these zeros.
Example: 𝜙 = 0 − 90° + 135° − 180° = −405° 2
13
 M.R. Emami, 2025 = −45°
 Root Locus Determination
Positive Root Locus
 Rule 5:
 There can be multiple roots (of the characteristic polynomial) at points on
1) Background the locus, and branches will depart (or approach) from a point of 𝑞 roots at
2) Basic Feedback angles separated by:
System 180° + 360° 𝑙 − 1
3) Root Locus 𝑙 = 1,2, ⋯ , 𝑞
Determination 𝑞
4) Case Study and will approach (or depart) at angles with the same separation, forming
5) Design for an array of 2𝑞 rays “equally” spaced. If the point is on real axis, the
Dynamic orientation of this array is determined by RULE 2. Otherwise, RULE 4 must
Compensation
be applied.
6) Design 𝑅 𝑠 + 𝑈 𝑠 1 𝑌 𝑠
Examples Example: 𝐾
_ 𝑠 𝑠+𝑐
7) Root-locus
Method Assume 𝑐 = 1,
Extensions
𝑌 𝑠 𝐾 𝐾
=𝒯 𝑠 = =
𝑅 𝑠 𝑠 + 𝑠 + 𝐾 𝑎 𝑠 + 𝐾𝑏 𝑠

𝑎 𝑠 + 𝐾𝑏 𝑠 = 𝑠 + 𝑠 + 𝐾 = 0
1 1 − 4𝐾
𝑟 ,𝑟 = − ±
2 2
1 1
𝐾= multiple roots at 𝑠 = −
4 2
Departure 180° + 360° 𝑙 − 1
= 90° , 270° 14
 M.R. Emami, 2025 Angles 2
 Root Locus Determination
Positive Root Locus
 Rule 6:
1) Background
 The breakaway/break-in points of the locus, i.e., points where two or more
2) Basic Feedback
branches meet and break away, are among the roots of (𝐿 𝑠 is open-loop TF)
System
3) Root Locus
𝑑𝐿 𝑠
=0
Determination 𝑑𝑠
4) Case Study
Example: 𝑅 𝑠 + 𝑈 𝑠 𝑌 𝑠
5) Design for 𝐾
Dynamic
_
Compensation 𝑠+1
Characteristic Equation: 1 + 𝐾 =0
6) Design 𝑠 𝑠+2 𝑠+3
Examples
7) Root-locus 𝑠+1 𝑑𝐿 𝑠 −2𝑠 − 8𝑠 − 10𝑠 − 6
Method 𝐿 𝑠 = = =0
Extensions 𝑠 𝑠+2 𝑠+3 𝑑𝑠 𝑠 𝑠+2 𝑠+3
𝑠 = −2.46 , −0.77 ± 𝑗0.79

For each root, check the real-positivity of

1
𝐾=−
𝐿 𝑠
𝐾 = 0.42 , 1.79 ∓ 𝑗4.28

Departure 180° + 360° 𝑙 − 1

Angles = 90° , 270° 15
 M.R. Emami, 2025 2
 Root Locus Determination
Positive Root Locus

1) Background
 MATLAB Function “rlocus.m”
2) Basic Feedback
System
3) Root Locus
Determination
4) Case Study
5) Design for
Dynamic
Compensation
6) Design
Examples
7) Root-locus
Method
Extensions

s = tf (‘s’);
sys =(s+1)/(s*(s+2)*(s+3));
rlocus (sys)

16
 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus
 Control Gain Selection:
 For each point of the locus:
1) Background
2) Basic Feedback Phase
𝜓 − 𝜙 = 180° + 360° 𝑙 − 1
System Relation
3) Root Locus
Determination Magnitude 1
𝐾=
4) Case Study Relation 𝐿 𝑠
5) Design for
Dynamic
Compensation  Each point of the root locus corresponds to a value of control gain 𝐾.
6) Design  Control gain 𝐾 can be selected for a desired value (location) of roots of
Examples
closed-loop system characteristic equation, which corresponds to a desired
7) Root-locus
Method behaviour of the system.
Extensions 1
Example: 𝐿 𝑠 = 𝑝
𝑠 𝑠 + 4 + 16

𝑏 𝑠 = 1; 𝑚 = 0; 𝑧 = 𝑒𝑚𝑝𝑡𝑦
𝑝
𝑎 𝑠 = 𝑠 + 8𝑠 + 32𝑠; 𝑛 = 3; 𝑝 = 0, −4 ± 𝑗4

Characteristic Equation (Root Locus Form): 𝑝

1
1+𝐾 =0
𝑠 𝑠 + 4 + 16
17
 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus
 Control Gain Selection:
 For each point of the locus:
1) Background
2) Basic Feedback Phase
𝜓 − 𝜙 = 180° + 360° 𝑙 − 1
System Relation
3) Root Locus
Determination Magnitude 1
𝐾=
4) Case Study Relation 𝐿 𝑠
5) Design for
Dynamic
Compensation  Each point of the root locus corresponds to a value of control gain 𝐾.
6) Design  Control gain 𝐾 can be selected for a desired value (location) of roots of
Examples
closed-loop system characteristic equation, which corresponds to a desired
7) Root-locus
Method behaviour of the system.
Extensions 1
Example: 𝐿 𝑠 = 𝑝
𝑠 𝑠 + 4 + 16
Asymptotes:
∑𝑝 − ∑𝑧 0 − 4 + 𝑗4 − 4 − 𝑗4 𝑝
𝛼= = = −2.67
𝑛−𝑚 3−0
180° + 360° 𝑙 − 1
𝜙 = = 60° , 180° , 300°
𝑛−𝑚 𝑝

18
 M.R. Emami, 2025
 Root Locus Determination
Positive Root Locus
 Control Gain Selection:
 For each point of the locus:
1) Background
2) Basic Feedback Phase
𝜓 − 𝜙 = 180° + 360° 𝑙 − 1
System Relation
3) Root Locus
Determination Magnitude 1
𝐾=
4) Case Study Relation 𝐿 𝑠
5) Design for
Dynamic
Compensation  Each point of the root locus corresponds to a value of control gain 𝐾.
6) Design  Control gain 𝐾 can be selected for a desired value (location) of roots of
Examples
closed-loop system characteristic equation, which corresponds to a desired
7) Root-locus
Method behaviour of the system.
Extensions 1 1
Example: 𝐿 𝑠 = =
𝑠 𝑠 + 4 + 16 𝑠 𝑠−𝑠 𝑠−𝑠
Departure Angle: 𝑝

𝜙 | = 𝜓 − 𝜙 − 180°

= 0 − −45° + 45° − 180° = −180° 𝑝

𝜙 | = 0 − +135° + 90° − 180°

= −360° − 45° 𝑝
19
 M.R. Emami, 2025 𝜙 | = 0 − −135° − 90° − 180° = +45°
 Root Locus Determination
Positive Root Locus
 Control Gain Selection:
 For each point of the locus:
1) Background
2) Basic Feedback Phase s = tf (‘s’);
𝜓 − 𝜙 = 180° + 360° 𝑙 − 1
System Relation sysL =1/(s*((s+4)^2+16));
3) Root Locus
Determination Magnitude 1 rlocus (sysL);
𝐾=
4) Case Study Relation 𝐿 𝑠 [K , p] = rlocfind (sysL);
5) Design for
Dynamic
Compensation  Each point of the root locus corresponds to a value of control gain 𝐾.
6) Design  Control gain 𝐾 can be selected for a desired value (location) of roots of
Examples
closed-loop system characteristic equation, which corresponds to a desired
7) Root-locus
Method behaviour of the system.
Extensions 1 1
Example: 𝐿 𝑠 = =
𝑠 𝑠 + 4 + 16 𝑠 𝑠−𝑠 𝑠−𝑠
For 𝜁 = 0.5 : 𝜙 = 90° + 30° = 120°
1
𝐾= = 𝑠 𝑠 −𝑠 𝑠 −𝑠
𝐿 𝑠
sin 𝜁
𝐾 ≅ 4.0 × 2.1 × 7.7 = 65
num = [1];
For finding the den = [1 8 32 0];
real root, sysL =tf (num, den);
use MATLAB. rlocus (sysL);
[K , p] = rlocfind (sysL); 20
 M.R. Emami, 2025 (𝑠 ≅ −4)
 Case Study
Satellite Attitude Control

1) Background
2) Basic Feedback
System
3) Root Locus
Determination
4) Case Study Rigid Body
𝑇 + 𝑀 = 𝐹 𝑑 + 𝑀 = 𝐼𝜃̈
5) Design for Angular Motion
Dynamic Θ 𝑠 11 𝐴
Compensation 𝑀 =0 =𝐺 𝑠 = = Double Integrator TF
𝑇 𝑠 𝐼𝑠 𝑠
6) Design
Examples
Flexible Body 𝑇 = 𝐼 𝜃̈ + 𝑏 𝜃̇ − 𝜃̇ +𝑘 𝜃 −𝜃
7) Root-locus
Method
Angular Motion
Extensions 0 = 𝐼 𝜃̈ + 𝑏 𝜃̇ − 𝜃̇ +𝑘 𝜃 −𝜃
𝑀 =0
𝑏=0 𝑇 = 𝐼 𝑠 + 𝑘 Θ 𝑠 − 𝑘Θ 𝑠

0 = −𝑘Θ 𝑠 + 𝐼 𝑠 + 𝑘 Θ 𝑠

Θ 𝑠 𝐼 𝑠 +𝑘
=
𝑇 𝑠 𝑘 𝑘 Collocated
𝐼𝐼 𝑠 𝑠 + +
𝐼 𝐼
Θ 𝑠 𝑘
=
𝑇 𝑠 𝑘 𝑘 Non-collocated 21
 M.R. Emami, 2025 𝐼𝐼 𝑠 𝑠 + +
𝐼 𝐼
 Case Study
Satellite Attitude Control (Rigid Body)

Rigid Body Θ 11
̈ =
1) Background
Angular Motion 𝑇 = 𝐼𝜃 𝑇 𝐼𝑠 𝑌 𝑠 𝐴
2) Basic Feedback =𝐺 𝑠 =
System Rigid Body 𝑋 1 1 𝑈 𝑠 𝑠
𝐹 = 𝑚𝑥̈ =
3) Root Locus Linear Motion 𝐹 𝑚𝑠 Double Integrator TF
Determination
4) Case Study
5) Design for 1
Dynamic 𝑌 𝑠 𝑘 𝑅 𝑠 + 𝑈 𝑠 1 𝑌 𝑠
Closed-loop TF = 𝑠 𝑘
Compensation
𝑅 𝑠 1+𝑘
1 _ 𝑠
6) Design 𝑠
Examples
Characteristic 1
7) Root-locus 1+𝑘 = 0 Root Locus
Method Equation 𝑠 Form
Extensions

 Root Locus:

 RULE 1: Two branches both starting

at 𝑠 = 0.
 RULE 2: No segment on the real axis.
 RULE 3: Two asymptotes intersecting
at 𝑠 = 0, and at the angles ±90° .
 RULE 4 (& 5): Segments depart from
22
 M.R. Emami, 2025 𝑠 = 0 at the angles ±90° .
 Case Study
Satellite Attitude Control (Rigid Body)
1 𝑅 𝑠 + 𝑈 𝑠 1 𝑌 𝑠
𝑌 𝑠 𝑘 +𝑘 𝑠 𝑘 +𝑘 𝑠
1) Background Closed-loop TF = 𝑠 _ 𝑠
2) Basic Feedback 𝑅 𝑠 1
1+ 𝑘 +𝑘 𝑠
System 𝑠
3) Root Locus
𝑘
Determination
𝐾=𝑘 and =1 Characteristic 𝑠+1
𝑘 Equation 1+𝐾 =0
4) Case Study 𝑠
 Root Locus:
5) Design for s = tf (‘s’);
Dynamic  RULE 1: Two branches both starting at 𝑠 = 0, one
Compensation sysL =(s+1)/(s^2);
terminating at 𝑠 = −1 (the zero) and the other goes infinity.
6) Design rlocus (sysL);
Examples
 RULE 2: The real axis on the left of 𝑠 = −1 is on the locus.
7) Root-locus
Method  RULE 3: Since 𝑛 − 𝑚 = 1, one asymptote along
Extensions
negative real axis.
 RULE 4: Departure angles from double poles
at 𝑠 = 0 are ±90° .
 RULE 5: The two branches meet on real axis
at angles ±90° .
 RULE 6: The two branches meet on real axis
at 𝑠 = −2.

 NOTE: The addition of a zero has pulled the locus into the LHP. 23
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Rigid Body)
𝑝𝑘 𝑅 𝑠 + 𝑘 𝑠 𝑈 𝑠 1 𝑌 𝑠
𝐾 = 𝑘 + 𝑝𝑘 and 𝑧= 𝑘 +
1) Background
𝐾 _ 𝑠 ⁄𝑝 + 1 𝑠
𝑠+𝑧 Lead
2) Basic Feedback 𝐷 𝑠 =𝐾 0<𝑧<𝑝
System 𝑠+𝑝 Compensator
3) Root Locus 𝑠+𝑧
Determination Characteristic Equation: 1+𝐷 𝑠 𝐺 𝑠 =1+𝐾 =0
𝑠 𝑠+𝑝
4) Case Study
 Root Locus: ( 𝑧 = 1 and 𝑝 = 12 ) s = tf (‘s’);
5) Design for sysL =(s+1)/(s^2)*(s+12);
Dynamic  RULE 1: Three branches, two starting at 𝑠 = 0, one starting
Compensation rlocus (sysL);
at 𝑠 = −12. One branch ends at 𝑠 = −1 and two at infinity.
6) Design
Examples
 RULE 2: The real axis segment −12 ≤ 𝑠 ≤ −1 is on the locus.
7) Root-locus
Method  RULE 3: 𝑛 − 𝑚 = 3 − 1 = 2 asymptotes centred at
Extensions
𝛼= = − 11⁄2 and at angles ±90° .
 RULE 4: Departure angles at 𝑠 = 0 are ±90° ,
and departure angle at 𝑠 = −12 is 0° .
 RULE 5: The two branches departing at 𝑠 = 0
meet on real axis at angles ±90° .
 RULE 6: The two branches departing at 𝑠 = 0
meet on real axis at 𝑠 = −2.3, and the other two
branches depart at 𝑠 = −5.2.
24
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Rigid Body)
𝑝𝑘 𝑅 𝑠 + 𝑘 𝑠 𝑈 𝑠 1 𝑌 𝑠
𝐾 = 𝑘 + 𝑝𝑘 and 𝑧= 𝑘 +
1) Background
𝐾 _ 𝑠 ⁄𝑝 + 1 𝑠
𝑠+𝑧 Lead
2) Basic Feedback 𝐷 𝑠 =𝐾 0<𝑧<𝑝
System 𝑠+𝑝 Compensator
3) Root Locus 𝑠+𝑧
Determination Characteristic Equation: 1+𝐷 𝑠 𝐺 𝑠 =1+𝐾 =0
𝑠 𝑠+𝑝
4) Case Study
 Root Locus: ( 𝑧 = 1 and 𝑝 = 4 ) s = tf (‘s’);
5) Design for sysL =(s+1)/(s^2)*(s+4);
Dynamic  RULE 1: Three branches, two starting at 𝑠 = 0, one starting
Compensation rlocus (sysL);
at 𝑠 = −4. One branch ends at 𝑠 = −1 and two at infinity.
6) Design
Examples
 RULE 2: The real axis segment −4 ≤ 𝑠 ≤ −1 is on the locus.
7) Root-locus
Method  RULE 3: 𝑛 − 𝑚 = 3 − 1 = 2 asymptotes centred at
Extensions
𝛼= = − 3⁄2 and at angles ±90° .
 RULE 4: Departure angles at 𝑠 = 0 are ±90° ,
and departure angle at 𝑠 = −4 is 0° .

 RULE 6: The two branches departing at

𝑠 = 0 do not meet at any point on the real
axis.

 NOTE: The addition of a pole moving in from

far left tends to push locus branches to the
right (causing oscillatory responses). 25
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Rigid Body)
𝑝𝑘 𝑅 𝑠 + 𝑘 𝑠 𝑈 𝑠 1 𝑌 𝑠
𝐾 = 𝑘 + 𝑝𝑘 and 𝑧= 𝑘 +
1) Background
𝐾 _ 𝑠 ⁄𝑝 + 1 𝑠
𝑠+𝑧 𝑘 Lead
2) Basic Feedback 𝐷 𝑠 =𝐾 0<𝑧<
System 𝑠+𝑝 𝑘 Compensator
3) Root Locus 𝑠+𝑧
Determination Characteristic Equation: 1+𝐷 𝑠 𝐺 𝑠 =1+𝐾 =0
𝑠 𝑠+𝑝
4) Case Study
 Root Locus: ( 𝑧 = 1 and 𝑝 = 9 ) s = tf (‘s’);
5) Design for sysL =(s+1)/(s^2)*(s+9);
Dynamic  RULE 1: Three branches, two starting at 𝑠 = 0, one starting
Compensation rlocus (sysL);
at 𝑠 = −9. One branch ends at 𝑠 = −1 and two at infinity.
6) Design
Examples
 RULE 2: The real axis segment −9 ≤ 𝑠 ≤ −1 is on the locus.
7) Root-locus
Method  RULE 3: 𝑛 − 𝑚 = 3 − 1 = 2 asymptotes centred at
Extensions
𝛼= = −4 and at angles ±90° .
 RULE 4: Departure angles at 𝑠 = 0 are ±90° ,
and departure angle at 𝑠 = −9 is 0° .
 RULE 6: Three branches departing at 𝑠 =
0 and 𝑠 = −9 meet at 𝑠 = −3.

 RULE 5: The three branches departing at

𝑠 = 0 and and 𝑠 = −9 meet on real axis
at angles ±60° and 180° .
26
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Rigid Body)
+
_ 1
𝐷 𝑠 =𝐾 1+𝐾 =0
1) Background 𝑠
2) Basic Feedback
System 𝑘
3) Root Locus 𝐷 𝑠 =𝑘 +𝑘 𝑠 =𝑘 𝑠+
Determination
𝑘 𝑘
=1
𝐾=𝑘 𝑘 1 𝑘 1
4) Case Study
1+𝐾 𝑠+ =0 1+𝐾 𝑠+1 =0
5) Design for 𝑘 𝑠 𝑠
Dynamic
Compensation
𝑠+𝑧
𝐷 𝑠 =𝐾
6) Design 𝑠+𝑝 𝑠+1 1
𝑝𝑘 𝑧=1
Examples 𝐾 = 𝑘 + 𝑝𝑘 ; 𝑧 = 1+𝐾 =0
7) Root-locus 𝐾 𝑝 = 12 𝑠 + 12 𝑠
Method
Extensions
𝑠+𝑧
𝐷 𝑠 =𝐾
𝑠+𝑝 𝑧=1 𝑠+1 1
𝑝𝑘
𝐾 = 𝑘 + 𝑝𝑘 ; 𝑧 = 1+𝐾 =0
𝐾 𝑝=9 𝑠+9 𝑠

𝑠+𝑧
𝐷 𝑠 =𝐾
𝑠+𝑝 𝑧=1 𝑠+1 1
𝑝𝑘
𝐾 = 𝑘 + 𝑝𝑘 ; 𝑧 = 1+𝐾 =0
𝐾 𝑝=4 𝑠+4 𝑠
27
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Collocated)
 Sensor and actuator are on the same rigid body, but
1) Background instrument (or antenna) is flexible w.r.t. the satellite body.
2) Basic Feedback
System Θ 𝑠 𝑠 + 0.1 + 6 𝑠+1
𝐺 𝑠 = = ; 𝐷 𝑠 =𝐾
3) Root Locus 𝑇 𝑠 𝑠 𝑠 + 0.1 + 6.6 𝑠 + 12
Determination
4) Case Study 𝑠+1 𝑠 + 0.1 + 6
Characteristic Equation: 1 + 𝐷 𝑠 𝐺 𝑠 = 1 + 𝐾 =0
5) Design for 𝑠 + 12 𝑠 𝑠 + 0.1 + 6.6
Dynamic
Compensation  RULE 1: Five branches, three approaching finite zeros
6) Design and two go infinity along asymptotes.
Examples
 RULE 2: Real axis segment −12 ≤ 𝑠 ≤ −1 on the locus.
7) Root-locus
Method  RULE 3: Two asymptotes centred at
Extensions . . . .
𝛼= = − 11⁄2 and at ±90° .
 RULE 4: 𝜙 = +90° +90° +𝑡𝑎𝑛 6.6
− +90° +90° +90° + 𝑡𝑎𝑛 6.6⁄12 − 180° = 142.6°
 RULE 5: The two branches departing at 𝑠 = 0 meet on
real axis at angles ±90° , and the other two real
branches depart at angles ±90° .
 RULE 6: The two branches departing at 𝑠 = 0
meet on real axis at 𝑠 ≅ −2.1, and the other two
branches depart at 𝑠 ≅ −4.8.
 NOTE: Single flexible mode brings lightly damped roots. 28
 M.R. Emami, 2025
 Case Study
Satellite Attitude Control (Non-collocated)

1) Background  Sensor and actuator are not on the same rigid body.
2) Basic Feedback
Θ 𝑠 1 𝑠+1
System 𝐺 𝑠 = = ; 𝐷 𝑠 =𝐾
3) Root Locus 𝑇 𝑠 𝑠 𝑠 + 0.1 + 6.6 𝑠 + 12
Determination
4) Case Study 𝑠+1 1
Characteristic Equation: 1 + 𝐷 𝑠 𝐺 𝑠 = 1 + 𝐾 =0
5) Design for 𝑠 + 12 𝑠 𝑠 + 0.1 + 6.6
Dynamic
Compensation  RULE 1: Five branches, one approaching a finite zero
6) Design and three go infinity along asymptotes.
Examples
 RULE 2: Real axis segment −12 ≤ 𝑠 ≤ −1 on the locus.
7) Root-locus
Method  RULE 3: Four asymptotes centred at
Extensions
.
𝛼= = − 11.2⁄4 and at ±45° , ±135° .

 RULE 4: 𝜙 = +𝑡𝑎𝑛 6.6

− +90° +90° +90° + 𝑡𝑎𝑛 6.6⁄12 − 180° = −37.4°
 RULE 5: The two branches departing at 𝑠 = 0
meet on real axis at angles ±90° .
 RULE 6: The two branches departing at 𝑠 = 0
meet on real axis at 𝑠 ≅ −2.1, and the other two
branches depart at 𝑠 ≅ −7.8.
29
 M.R. Emami, 2025  NOTE: Lead compensator can barely stabilize non-collocated flexibility.
 Case Study
Complex Multiple Roots
1
Characteristic Equation: 1 + 𝐷 𝑠 𝐺 𝑠 = 1 + 𝐾 =0
𝑠 𝑠+2 𝑠+1 +4
1) Background
2) Basic Feedback  RULE 1: Four branches, all approaching infinite zeros
System along asymptotes.
3) Root Locus
Determination
 RULE 2: Real axis segment −2 ≤ 𝑠 ≤ 0 on the
4) Case Study
locus.
5) Design for
Dynamic
Compensation
 RULE 3: Four asymptotes centred at
±
6) Design 𝛼= = −1 and at ±45° , ±135° .
Examples
7) Root-locus  RULE 4:
Method 2 2
Extensions 𝜙 = 𝜙 = −𝑡𝑎𝑛 − 𝑡𝑎𝑛 −90° −180°
−1 1
= −90°

 RULE 5: The two branches on real axis depart at

angles ±90° , and the other two pairs of branches
(conjugate) depart at angles 0° and 180° .

 RULE 6: The two branches on real axis depart at

𝑠 = −1, and the other two pairs of branches
(conjugate) depart at 𝑠 = −1 ± 𝑗1.22.
30
 M.R. Emami, 2025
 Design for Dynamic Compensation
 Design the controller to alter (improve)
1) Background the dynamic behaviour of the plant, i.e.,
2) Basic Feedback transient response, steady-state error, etc.
System
𝑠+𝑧
3) Root Locus Lead Compensation: 𝐷 𝑠 =𝐾 𝑧<𝑝
Determination 𝑠+𝑝
4) Case Study  For a sinusoidal input, its output leads the input, i.e., the
5) Design for output signal has a positive phase shift.
Dynamic
Compensation  Approximates the PD control, speeding up the response
6) Design (by lowering rise time) and decreasing overshoot.
Examples
7) Root-locus
𝑠+𝑧
Method
Lag Compensation: 𝐷 𝑠 =𝐾 𝑧>𝑝
𝑠+𝑝
Extensions
 For a sinusoidal input, its output lags the input, i.e., the
output signal has a negative phase shift.
 Approximates the PI control, decreasing steady-state error.
𝑠 + 2𝜁𝜔 𝑠 + 𝜔
Notch Compensation: 𝐷 𝑠 =𝐾
𝑠+𝜔
 Attenuates the input around a prescribed (unwanted)
frequency. Overall, achieves stability for plants with
lightly damped flexible modes.
 Typically consists of two complex zeros, for capturing 31
 M.R. Emami, 2025 troublesome poles, and two real poles far out in the LHP.
 Design for Dynamic Compensation
𝜃𝑚 𝑡
 Example: Servomechanisms bl T

𝜏 << 𝜏 𝑙 𝐽 la ra
h JIl arm
1) Background <<
𝑟 𝑏 𝑣 𝑡
ia Tm q m b m
2) Basic Feedback + ql Tl
+
_ va vb
System _
Jm
𝐾 motor
Jg
3) Root Locus Θ 𝑠 𝑟 𝐾 𝑇 𝑡
Determination = =
𝑉 𝑠 𝐾𝐾 𝑠 𝜏𝑠 + 1
𝑠 𝐽 𝑠+ 𝑏 + 𝑟 𝐾 𝐽 𝑟
4) Case Study
Θ 𝑠 −𝜂 𝐶 ; 𝐾=
𝑏 𝑟 +𝐾 𝐾
; 𝜏=
𝑏 𝑟 +𝐾 𝐾
5) Design for
Dynamic
= =
𝑇 𝑠 𝐾𝐾 𝑠 𝜏𝑠 + 1 −𝜂𝑟
Compensation 𝑠 𝐽 𝑠+ 𝑏 + 𝑟 𝐶=
𝑏 𝑟 +𝐾 𝐾
6) Design T s
Examples h
7) Root-locus
Tm s ++
1 Q
Method vVas Kt 1 qm s
Extensions  J m s   bm 
K e Kt 
ra
 ra  s

Θ 𝑠 𝐾𝐷 𝑠
=
Θ 𝑠 𝑠 𝜏𝑠 + 1 + 𝐾ℎ𝐷 𝑠
Θ 𝑠 −𝐾𝜂
=
𝑇 𝑠 𝑠 𝜏𝑠 + 1 + 𝐾ℎ𝐷 𝑠

+
+ +
 
_

32
 M.R. Emami, 2025
 Design for Dynamic Compensation
Lead Compensation
 Practical implementation of PD control, since pure differentiation of a signal
1) Background (e.g., the output error) always creates more noise in the system.
2) Basic Feedback 𝑌 𝑠 1 P Control
2nd-order Position System PD Control
System =𝐺 𝑠 = 𝑠𝑖𝑛 𝜁
3) Root Locus
(Servomechanism) 𝑈 𝑠 𝑠 𝑠+1
Determination
P Control: 𝐷 𝑠 =𝐾 𝜔 =2
4) Case Study
5) Design for  weak damping behaviour (𝜔 =2 𝜁 = 0.25 )
Dynamic
Compensation PD Control: 𝐷 𝑠 =𝐾 𝑠+2
6) Design
Examples  improved damping behaviour (𝜔 = 2 𝜁 = 0.75)
7) Root-locus 𝑠+2
Method Lead Compensation: 𝐷 𝑠 = 𝐾
Extensions 𝑠+𝑝
 For small gains 𝐾, lead compensator behaves
like PD control.
 For large values of 𝑝, lead compensator
behaves like PD control.
 The additional pole (slightly) lowers the
damping, but with low gains and high values of
the pole the effect is negligible.
 Typically, the zero is placed near the desired closed-loop 𝜔 (1/4 to 1 times 𝜔 ),
and the (real) pole 5-25 times the value of the zero location. (The farther the
pole, the more noise creation, getting closer to signal differentiation, as well as33
 M.R. Emami, 2025
more steady-state error.)
 Design for Dynamic Compensation
Lead Compensation
𝑌 𝑠 1
Example: =𝐺 𝑠 =
𝑈 𝑠 𝑠 𝑠+1
1) Background
 Design a lead compensator for the position system to
2) Basic Feedback
System
provide an overshoot of no more than 20% and rise time
3) Root Locus
of no more than 0.3 seconds.
Determination  From Chap. 3, a damping ratio of 𝜁 ≥ 0.5 and a natural
4) Case Study frequency of 𝜔 ≥ . ⁄ . = 6 𝑟𝑎𝑑 ⁄𝑠 should satisfy the
5) Design for requirements. (Choose 𝜔 ≥ 7 for some margin.)
Dynamic
Compensation 𝑠+2
6) Design
 First Trial: 𝐷 𝑠 =𝐾
Examples
𝑠 + 10
s = tf (‘s’);
For 𝐾 = 70, sysG = 1/(s*(s+1));
7) Root-locus
𝜁 = 0.56 and 𝜔 = 7.7 𝑟𝑎𝑑⁄𝑠 sysD1 = (s+2)/(s+10);
Method
rlocus (sysG*sysD1);
Extensions
 Rise Time < 0.3 sec sysD = 70*(s+2)/(s+10);
sysCL= feedback(sysG*sysD,1);
 Overshoot > 20% step (sysCL);

𝑠+2
 Second Trial: 𝐷 𝑠 = 𝐾
𝑠 + 13
For 𝐾 = 91,
s = tf (‘s’);
𝜁 = 0.67 and 𝜔 = 8.6 𝑟𝑎𝑑⁄𝑠 sysG =1/(s*(s+1));
sysD1 = (s+2)/(s+13);
rlocus (sysG*sysD1);
 Rise Time < 0.3 sec sysD =91*(s+2)/(s+13);
 Overshoot < 20% sysCL= feedback(sysG*sysD,1);
step (sysCL); 34
 M.R. Emami, 2025
 Design for Dynamic Compensation
Lead Compensation Design Process
1) Determine where the closed-loop roots need to be in the s-plane in order to meet the
1) Background
desired specifications on the speed of response and damping (or overshoot).
2) Basic Feedback
a) pick the limits for 𝜔 and 𝜁, or
System b) pick the limits for 𝜎 and 𝜔 .
3) Root Locus 2) Create the root locus w.r.t. K with no compensation.
Determination
4) Case Study 3) If more damping is required, select a value of z of the Lead Compensator to be around
5) Design for 1/4 to 1 times the value of the desired 𝜔 or 𝜔 and pick p to be (5-25)×z.
Dynamic
Compensation 4) Examine the resulting root locus; and adjust as necessary to meet the required
6) Design specifications as determined in step 1.
Examples a) Decrease p if less damping is needed,
7) Root-locus b) Increase p if more damping is needed, and/or decrease z ,
Method
Extensions c) It is desirable to keep the value of p / z as low as possible (p / z  25) in order
to minimize the error and amplification of sensor noise by the compensation.
5) When the values of z and p are selected so that the resulting locus passes through an
acceptable region of the s-plane, determine the value of K to select the closed-loop
root locations.
6) Verify that all time domain specifications are met by examining the time response to a
unit step input, and adjust the desired s-plane specifications if needed and go back to
step 2.
7) Determine if the resulting value of K meets the steady-state error requirements, if any.
If a value of K cannot be found that meets the requirement, then add Integral Control
or a Lag Compensator. 35
 M.R. Emami, 2025
 Design for Dynamic Compensation
Lag Compensation
 Typically, added to the Lead Compensator to decrease the steady-state error
1) Background (at very low frequencies when 𝑠 = 0), without affecting the transient response
2) Basic Feedback caused by the Lead Compensator.
System
3) Root Locus Example: Increase velocity constant 𝐾 of a type-1 system (unity feedback) to
Determination decrease the steady-state error to a ramp input, without changing its
4) Case Study transient response. 𝐾 = lim 𝑠𝐷 𝑠 𝐺 𝑠
5) Design for →
Dynamic  The additional TF to 𝐷 𝑠 should have a significant gain at 𝑠 = 0 , but should be
Compensation
nearly unity (no effect) at the higher frequency 𝜔 , where dynamic (transient)
6) Design
Examples response is determined.
𝑠+𝑧 𝑧 and 𝑝 are very small compared to 𝜔 .
7) Root-locus
Method
𝐷 𝑠 = 𝑧 > 𝑝 
𝑠+𝑝 𝐷 0 = 𝑧⁄𝑝 = 3 𝑡𝑜 10 .
Extensions
Second-order 1
𝐺 𝑠 =
Position System 𝑠 𝑠+1
Lead 𝑠+2
Compensator 𝐷 𝑠 = 91 ×
𝑠 + 13
𝐾 = lim 𝑠𝐷 𝑠 𝐺 𝑠 = 14
→

New 𝐾 = 5 × 14 = 70
𝑧 ⁄𝑝 = 5
𝑧 = 0.05 𝑠 + 0.05
Choose: 𝐷 𝑠 = 36
 M.R. Emami, 2025 𝑝 = 0.01 𝑠 + 0.01
 Design for Dynamic Compensation
Lag Compensation Design Process

1) Background
2) Basic Feedback 1) Determine the amount of gain amplification to be contributed by the Lag
System Compensation at low frequencies, in order to achieve the desired 𝐾 or 𝐾 or 𝐾 as
3) Root Locus error constants.
Determination
4) Case Study 2) Select the value of z in the Lag Compensator so it is approximately a factor of 100 to
5) Design for 200 smaller than the system dominant natural frequency.
Dynamic
Compensation
3) Select the value of p in the Lag Compensator so that z / p is equal to the desired gain
6) Design
amplification determined in step 1.
Examples
4) Examine the resulting root locus to verify that the frequency and damping of the
7) Root-locus
Method
dominant closed-loop roots are still satisfactory. If not, adjust the lead compensation
Extensions as needed.
5) Verify that all time domain specifications are met by examining the time response to
a unit step input. If the slow root introduced by the Lag Compensation is too slow,
increase the values of z and p somewhat while keeping z / p constant, and go back to
step 4. However, do so with the understanding that the closer the values of the Lag
Compensator's z and p come to the dominant roots of the closed-loop system, the
more they will affect those dominant root characteristics.

37
 M.R. Emami, 2025
 Design for Dynamic Compensation
Notch Compensation
 Added to the controller to dampen the oscillation at a resonant frequency,
1) Background e.g., due to the flexibility of a non-collocated case (without affecting the
2) Basic Feedback transient response and steady-state error).
System
3) Root Locus Example: The TF for the non-collocated version of the 2nd-order position system
Determination with a resonant frequency at 𝜔 = 50 𝑟𝑎𝑑⁄𝑠 can be represented as:
4) Case Study 1 2500 1 2500
5) Design for 𝐺 𝑠 = ≅
Dynamic
𝑠 𝑠+1 𝑠 + 𝑠 + 2500 𝑠 𝑠+1 𝑠 + 0.5 + 50
Compensation  Assume that the lead-lag control previously 𝑠 + 2 𝑠 + 0.05
6) Design designed is implemented, but an additional 𝐷 𝑠 = 91
Examples 𝑠 + 13 𝑠 + 0.01
notch filter is needed to remove the resonance.
7) Root-locus
Method Notch Compensation:
Extensions 𝑠 + 2𝜁𝜔 𝑠 + 𝜔 𝑠 + 0.8𝑠 + 3600 𝑠 + 0.4 + 60
𝐷 𝑠 = = ≅
𝑠+𝜔 𝑠 + 60 𝑠 + 60
 Filter gain at 𝑠 = 0 is
unity (no effect on 𝑒 ).

 Filter poles are far left

in LHP (no effect on
transient response).

 Some compromise in
overshoot specs.
38
 M.R. Emami, 2025
 Design Examples
Quadrotor Drone Control

1) Background
2) Basic Feedback
System
3) Root Locus
Determination
4) Case Study
𝜃 𝑠 1
5) Design for =𝐺 𝑠 =
Dynamic 𝑇 𝑠 𝑠 𝑠+2
Compensation
6) Design 𝜔 ≥ 1 𝑟𝑎𝑑 ⁄𝑠 𝑠 + 0.5 𝜔 ≅ 1.03 𝑟𝑎𝑑⁄𝑠
Examples Specs: 𝐷 𝑠 = 30 𝐷 𝑠 =𝐾
7) Root-locus 𝜁 ≥ 0.44 𝑠 + 15 𝜁 ≅ 0.446
Method
Extensions

𝑠+1 𝑠 + 0.5
𝐷 𝑠 =𝐾 𝐷 𝑠 =𝐾
𝑠 + 10 𝑠 + 15
39
 M.R. Emami, 2025
 Design Examples
Quadrotor Drone Control

5
1) Background
2) Basic Feedback
System
3) Root Locus
Determination
4) Case Study  Non-collocated Case: Flexibility between actuators and the body.
5) Design for 𝜃 𝑠 1 225
Dynamic =𝐺 𝑠 =
Compensation 𝑇 𝑠 𝑠 𝑠+5 𝑠 + 0.1 + 15
6) Design 𝑡 ≤1𝑠 𝜔 ≥ 1.8 𝑟𝑎𝑑⁄𝑠
Examples
7) Root-locus
𝑀 % ≤ 40 𝜁 ≥ 0.3
Method Specs: 𝑡 ≤ 10 𝑠 𝜎 ≥ 0.46 𝐺 𝑠
Extensions
𝐾 ≥ 12 𝑟𝑎𝑑 Use Lag and Notch filter.
No high frequency oscillation in the response.

𝑠 + 0.5
Lead: 𝐷 𝑠 =𝐾
𝑠 + 10
 Choose: 𝑧 = 0.5 ≅ 0.3𝜔
 Choose: 𝑝 = 10 = 20𝑧

 From the root locus, 𝑠 + 0.5

𝐷 𝑠 = 80 𝐷 𝑠 𝐺 𝑠
Choose: 𝐾 = 80 𝑠 + 10
 Check the time response.
40
 M.R. Emami, 2025
 Design Examples
Quadrotor Drone Control

5
1) Background
2) Basic Feedback
System
3) Root Locus
Determination 𝜃 𝑠 1 225
=𝐺 𝑠 =
4) Case Study 𝑇 𝑠 𝑠 𝑠+5 𝑠 + 0.1 + 15
5) Design for 𝑡 ≤1𝑠 𝜔 ≥ 1.8 𝑟𝑎𝑑⁄𝑠
Dynamic
Compensation 𝑀 % ≤ 40 𝜁 ≥ 0.3
6) Design Specs: 𝑡 ≤ 10 𝑠 𝜎 ≥ 0.46
Examples
𝐾 ≥ 12 𝑟𝑎𝑑 Use Lag and Notch filter.
7) Root-locus
Method No high frequency oscillation in the response 𝐷 𝑠 𝐷 𝑠 𝐺 𝑠
Extensions
𝑠 + 0.05 + 16
Notch: 𝐷 𝑠 =
𝑠 + 16
 Choose the zeros above (and close to) plant (unwanted) poles:
 Choose the poles to have
a DC gain of nearly 1.

 Check the time response.

 Modify the Lead

Compensator and
control gain if necessary.
𝑠 + 0.5 𝑠 + 0.05 + 16 𝑠 + 0.5 𝑠 + 0.05 + 16
𝑠 + 0.5 𝐷 𝑠 = 75 𝐷 𝑠 = 80
𝑠 + 13 𝑠 + 16 𝑠 + 10 𝑠 + 16
𝐷 𝑠 = 75 41
 M.R. Emami, 2025 𝑠 + 13
 Design Examples
Quadrotor Drone Control

5
1) Background
2) Basic Feedback
System
3) Root Locus
Determination 𝜃 𝑠 1 225
=𝐺 𝑠 =
4) Case Study 𝑇 𝑠 𝑠 𝑠+5 𝑠 + 0.1 + 15
5) Design for 𝑡 ≤1𝑠 𝜔 ≥ 1.8 𝑟𝑎𝑑⁄𝑠
Dynamic
Compensation 𝑀 % ≤ 40 𝜁 ≥ 0.3
6) Design Specs: 𝑡 ≤ 10 𝑠 𝜎 ≥ 0.46
Examples
𝐾 ≥ 12 𝑟𝑎𝑑 Use Lag and Notch filter.
7) Root-locus
Method No high frequency oscillation in the response
Extensions
𝐷 𝑠 𝐷 𝑠 𝐷 𝑠 𝐺 𝑠
𝑠+𝑧
Lag: 𝐷 𝑠 = ; 𝐾 = lim 𝑠 𝐷 𝑠 𝐷 𝑠 𝐺 𝑠 = 0.58
𝑠+𝑝 →
𝑧 12
≥ = 20.7
𝑝 0.58
 Choose: 𝑧 = 0.02 ≅ 0.01𝜔
1
 Choose: 𝑝 = 0.001 = 𝑧
20
𝑠 + 0.02
𝐷 𝑠 =
𝑠 + 0.001
 Check the time response. 𝐷 𝑠 = 65 𝑠 + 0.5 𝑠 + 0.02 𝑠 + 0.05 + 16
𝑠 + 13 𝑠 + 0.001 𝑠 + 16
 Modify control 𝐾 = 65 42
 M.R. Emami, 2025 gain if necessary.
 Design Examples
Small Airplane Control
𝑃𝑖𝑡𝑐ℎ 𝐴𝑡𝑡𝑖𝑡𝑢𝑑𝑒 (𝑑𝑒𝑔) 𝜃 𝑠 160 𝑠 + 2.5 𝑠 + 0.7
=𝐺 𝑠 = =
1) Background 𝐸𝑙𝑒𝑣𝑎𝑡𝑜𝑟 𝐴𝑛𝑔𝑙𝑒 (𝑑𝑒𝑔) 𝛿 𝑠 𝑠 + 5𝑠 + 40 𝑠 + 0.03𝑠 + 0.06
Piper Dakota Plane
2) Basic Feedback  Design a controller for a unit step response with rise
System
time of 1 sec or less and overshoot of 10% or less.
3) Root Locus
Determination 1.8
𝑡 ≅ 𝜔 > 1.8 𝑟𝑎𝑑 ⁄𝑠
4) Case Study  For a 2 -order system:
nd 𝜔
5) Design for 𝑀 % ≤ 10 𝜁 > 0.6
Dynamic
Compensation  Start with a proportional control: The two faster roots
6) Design always have a damping ratio less than 0.4, causing
Examples excess oscillation & overshoot. s = tf (‘s’);
7) Root-locus 𝑠+3 sysG = (160*(s+2.5)(s+0.7))/((s^2+5*s+
Method  Proceed with a lead compensator: 𝐷 𝑠 = 𝐾 40)*(s^2+0.03*s+0.06));
Extensions 𝑠 + 20 sysD =(s+3)/(s+20);
𝐾 = 1.5 𝜁 >≅ 0.52 𝑡 ≅ 0.9 rlocus(sysG*sysD);
𝜔 ≅ 15 𝑟𝑎𝑑⁄𝑠 𝑀 %≅8 sysCL= feedback(1.5*sysG*sysD,1);
step (sysT);

43
 M.R. Emami, 2025
 Design Examples
Small Airplane Control

 To act off a constant

1) Background
disturbance moment, a
2) Basic Feedback
System (small) trim tab, actuated
3) Root Locus by a separate (small) servomotor,
Determination is added to the tail. Adjust the controller
4) Case Study to use the tab angle 𝛿 to balance the
5) Design for constant disturbance at steady state without
Dynamic
Compensation having any 𝛿 .
6) Design
Examples
 To let 𝛿 create the required moment to balance off the disturbance at steady state,
7) Root-locus
we can feed the integral of command 𝛿 to the trim tab servomotor for it to
Method eventually provide the balancing moment, thus eliminating the need for a steady-
Extensions
state 𝛿 .
𝐾 𝐾
𝐷 𝑠 = 𝐾𝐷 𝑠 1 + Characteristic Equation: 1 + 𝐾𝐷 𝐺 + 𝐾𝐷 𝐺 = 0
𝑠 𝑠
𝐾 ⁄𝑠 𝐾𝐷 𝐺 1 𝐾𝐷 𝐺
1+ =0 Open-loop TF: 𝐿 𝑠 =
1 + 𝐾𝐷 𝐺 𝑠 1 + 𝐾𝐷 𝐺

44
 M.R. Emami, 2025
 Design Examples
Small Airplane Control

 Damping of the fast roots decreases

1) Background
as 𝐾 increases (typical case when
2) Basic Feedback
System
integral control is added). So, keep 𝐾
3) Root Locus
as low as possible. s = tf (‘s’);
sysG = (160*(s+2.5)(s+0.7))/((s^2+5*s+
Determination
Trial & Error: 𝐾 = 0.15 40)*(s^2+0.03*s+0.06));
4) Case Study 𝑡 < 1% ≅ 33 𝑠𝑒𝑐 sysD =(s+3)/(s+20);
5) Design for For slow root at 𝑠 = −0.14 sysT = feedback(1.5*sysG*sysD,1);
Dynamic
𝜏 = 1⁄0.14 ≅ 7 𝑠𝑒𝑐 sysIn =1/s;
Compensation rlocus(sysT*sysIn);
6) Design sysCL= feedback(0.15*sysT*sysIn,1);
Examples step (sysT);
7) Root-locus
Method
Extensions

45
 M.R. Emami, 2025
 Root Locus Method Extension
Successive loop Closure
 Some feedback control systems may contain multiple loops, e.g., an inner
1) Background loop around an actuator or other parts of the plant and an outer loop around
2) Basic Feedback the entire plant.
System
3) Root Locus  The outcome is typically a more effective control system than a single closed
Determination loop.
4) Case Study
Example: Servomechanism
5) Design for
Dynamic 𝐾 𝐾
Compensation Characteristic 1 + + =0
6) Design
𝑠 𝑠+1 𝑠+1
Equation:
Examples
𝑠 +𝑠+𝐾 +𝐾 𝑠 =0
7) Root-locus
Method
Extensions
Iteration Process: Set the outer gain to a nominal/initial value, find a proper inner
gain using root locus method, then find a new outer gain by fixing the inner gain.
Continue the iteration until satisfactory performance is achieved.
𝑠
𝐾 =4 1+𝐾 =0 𝐾 =1
𝑠 +𝑠+4
1
𝐾 =4+𝐾 1+𝐾 =0
𝑠 + 2𝑠 + 4

NOTE: Open poles of the new locus are

the roots on the previous locus at the
selected 𝐾 .
46
 M.R. Emami, 2025
 Root Locus Method Extension
Successive loop Closure
Example: Drone Horizontal Position Control
𝑥 𝑠 𝑔
1) Background 𝐹 = −𝑚𝑔 𝑠𝑖𝑛 𝜃 ≅ −𝑚𝑔 𝜃 = 𝑚𝑥̈ =𝐺 𝑠 =−
𝜃 𝑠 𝑠
2) Basic Feedback
System 𝜃 𝑠 1 (𝑔 ≅ 32.2 𝑓𝑝𝑠 )
=𝐺 𝑠 =
3) Root Locus 𝑇 𝑠 𝑠 𝑠+2 𝐺 𝑠
Determination
4) Case Study
𝜔 ≥ 0.4 𝑟𝑎𝑑 ⁄𝑠
5) Design for
Specs:
𝜁 ≥ 0.5
Dynamic
Compensation 𝑠 + 0.5
Inner-loop
6) Design 𝐷 𝑠 =𝐾
Examples
Design: 𝑠 + 15
7) Root-locus 𝑠 + 15
Method 𝐺 𝑠 =
Extensions 𝑠 + 17𝑠 + 30𝑠 + 𝐾 𝑠 + 0.5
 Choose: 𝐾 = 30 to make 𝑒 to step input zero.
Outer-loop 𝐷 𝑠 + 0.1
𝑠 =𝐾
Design: 𝑠 + 10
𝑠 + 0.1 𝑔 𝑠 + 15
𝐿 𝑠 =
𝑠 + 10 𝑠 𝑠 + 17𝑠 + 30𝑠 + 30𝑠 + 15
𝜔 ≅ 0.4 𝜔 ≅ 0.9 𝑟𝑎𝑑 𝑠
 Choose: 𝐾 = 0.081 ;
𝜁 ≅ 0.7 𝜁 ≅ 0.6
 Iterate on 𝐷 and 𝐷 if need be. 47
 M.R. Emami, 2025
 Root Locus Method Extension
Negative (0° ) Root Locus
 Positive Feedback System (with non-negative gain TFs):
1) Background
Characteristic 1 − 𝐴𝐿 𝑠 = 1 + −𝐴 𝐿 𝑠 = 1 + 𝐾𝐿 𝑠 = 0 ; 𝐾 = −𝐴 < 0
2) Basic Feedback
Equation
System  Nonminimum Phase System (a zero in RHP): 𝐾 = −𝐴 < 0
3) Root Locus
Determination 1 + 𝐴 𝑧 − 𝑠 𝐿 𝑠 = 1 + −𝐴 𝑠 − 𝑧 𝐿 𝑠 = 1 + 𝐾𝐿 𝑠 = 0
4) Case Study
 In such cases, 𝐾 is real and negative, and 𝐿 𝑠 must be real and positive.
5) Design for
Dynamic In polar form, the phase of 𝐿 𝑠 must be 0° (negative locus).
Compensation
6) Design Negative Locus: ∠𝐿 𝑠 = 𝜓 − 𝜙 = 0° + 360° 𝑙 − 1
Examples
7) Root-locus  𝜓 is the angle to a test point 𝑠 from a zero 𝑧 , and 𝜙 is the angle to the
Method
Extensions
test point from a pole 𝑝 , and 𝑙 is an integer:
 Rule 1 (same as for Positive Root Locus):
 The 𝑛 branches of the locus leave the poles, 𝑚 branches approach the
zeros, and 𝑛 − 𝑚 branches approach the asymptotes.
 Rule 2:
 The locus on the real axis is to the left of an even number of real poles plus
zeros (assuming no poles plus zeros as an even number.)
 Rule 3: 0° + 360° 𝑙 − 1
∑𝑝 − ∑𝑧
 The asymptotes are determined as: 𝛼 = ; 𝜙 = 𝑛−𝑚
𝑛−𝑚 48
 M.R. Emami, 2025 𝑙 = 1,2, ⋯ , 𝑛 − 𝑚
 Root Locus Method Extension
 Rule 4: Negative (0° ) Root Locus
 Departure angles from poles and arrival angles to zeros are found by
1) Background searching in the near neighbourhood of the poles or zeros where the phase
2) Basic Feedback of 𝐿 𝑠 =0 (𝑞 is the order of the pole or zero):
System
3) Root Locus
𝑞𝜙 , = 𝜓 − 𝜙 − 0° − 360° 𝑙 − 1
Determination , ; 𝑙 = 1,2, ⋯ , 𝑞
4) Case Study
𝑞𝜓 , = 𝜙 − 𝜓 + 0° + 360° 𝑙 − 1
5) Design for
Dynamic ,
Compensation NOTE: For a single pole/zero the angles of departure/arrival for positive and negative
6) Design locus branches are 180° apart, and for a double pole/zero the two positive branches
Examples
are 180° apart and the two negative branches are 90° to the positive branches.
7) Root-locus
Method  Rule 5 (same as for Positive Root Locus):
Extensions
 There can be 𝑞 multiple roots at points on the locus, where 𝑞 branches will
depart (or approach) at angles separated by:
180° + 360° 𝑙 − 1
; 𝑙 = 1,2, ⋯ , 𝑞
𝑞
and will approach (or depart) at angles with the same separation, forming an
array of 2𝑞 rays equally spaced. If the point is on real axis, the orientation of
this array is determined by RULE 2. Otherwise, RULE 4 must be applied.
 Rule 6 (same as for Positive Root Locus):
 The points of multiple roots (breakaway / 𝑑𝐿 𝑠
 M.R. Emami, 2025 break-in) are among the roots of: =0 49
𝑑𝑠
 Root Locus Method Extension
Negative (0° ) Root Locus

Example: Airplane Altitude Control

1) Background
2) Basic Feedback 𝐴𝑙𝑡𝑖𝑡𝑢𝑑𝑒 (𝑚) ℎ 𝑠 6−𝑠
System =𝐺 𝑠 = =
3) Root Locus
𝐸𝑙𝑒𝑣𝑎𝑡𝑜𝑟 𝐴𝑛𝑔𝑙𝑒 (𝑑𝑒𝑔) 𝛿 𝑠 𝑠 𝑠 + 4𝑠 + 13
Determination
4) Case Study Characteristic Equation:
5) Design for
𝑠−6
Dynamic 1+𝐷 𝑠 𝐺 𝑠 = 1+𝐾 =0 ; −∞ < 𝐾 < 0
Compensation 𝑠 𝑠 + 4𝑠 + 13
6) Design
Examples  RULE 1: 3 branches and 2 asymptotes.
7) Root-locus
Method  RULE 2: Real-axis segments to the right of 𝑠 = 6, and to the left of 𝑠 = 0 .
Extensions
 RULE 3: Asymptotes intersect at 𝛼 = = −5 at angles
°
𝜙 = = 0° , 180° .
 RULE 4: Branch departing the pole 𝑠 = −2 + 𝑗3
at angle
3 3
𝜙 = 𝑡𝑎𝑛 − 𝑡𝑎𝑛 − 90° + 360° 𝑙 − 1
−8 −2
°
= −54.3
 RULE 6: Two branches meet at 𝑠 = +9.84.
50
 M.R. Emami, 2025  RULE 5: The break-in angle is ±90° .