Module 14:

Controller Design
Rick Hill, PhD
Dept of Mechanical Engineering
Module 14: Controller Design

1. Control requirements
2. Design by pole placement

Controller Design
3. Introduction to PID control
Control Requirements
• Approach to control design: Translate
engineering specifications into control
requirements, then design a controller to
meet those specifications

Controller Design
• Example transient specifications
• τ, Mp, tp, ts, tr
• Example steady-state specifications
• Typically it is required that steady-state error be
less than some amount, ess < 0.02
• Can use F.V.T. for different types of reference
inputs e  lim sE ( s)
ss s 0
Pole Placement
• Poles of a system affect the time response
• Choose controller gains (for ex. KP, KI, and
KD) so that poles of the closed-loop system

Controller Design
meet given requirements
• If system is simple (first order, or second
order) can employ algebra
• Process is generally iterative
Example
• Find the closed-loop transfer function for a
cruise control system with PI controller
 𝑉(𝑠) 𝐾𝑃 𝑠 + 𝐾𝐼

Example (continued)
=
𝑉𝑑𝑒𝑠 (𝑠) 𝑚𝑠 2 + 𝑏 + 𝐾𝑃 𝑠 + 𝐾𝐼

• With two parameters (KP and KI) can place poles of

this second order system anywhere we like (2 d.o.f.)
• What is the effect of changing the control gains?
 𝑉(𝑠) 1 𝐾𝑃 𝑠 + 𝐾𝐼

Example (continued)
= ∙ 2
𝑉𝑑𝑒𝑠 (𝑠) 𝑚 𝑠 + 𝑠 𝑏 + 𝐾𝑃 /𝑚 + 𝐾𝐼 /𝑚

• Let m=5 and b=1. Choose KP and KI to achieve a

settling time of 2 seconds and a peak time of 1 second.
 𝑉(𝑠) 1 𝐾𝑃 𝑠 + 𝐾𝐼

Example (continued)
= ∙ 2
𝑉𝑑𝑒𝑠 (𝑠) 5 𝑠 + 𝑠 1 + 𝐾𝑃 /5 + 𝐾𝐼 /5

• Will the closed-loop system with this controller

actually have a settling time of 2 seconds and a
peak time of 1 second?
Example (continued)
• If system is not canonical (has zero, higher order),
can use calculations based on canonical system as
a starting point

• If possible, can place dominant one or two poles at

the desired location, then place other poles to be
much faster (farther to left)

• Often requirements are given as inequalities

Example (continued)
• Plot region in complex plane where poles of a
canonical 2nd order system must be located in order
to achieve a settling time less than 2 seconds and a
peak time less than 1 second
 19𝑠 + 69.3 1

Example 𝑠 5𝑠 + 1

• What is the resulting steady-state error of this system

for a unit ramp reference?
PID Control
• PID controllers are ubiquitous
• Very intuitive, easy to implement
• Provide sophistication in that the control is

Controller Design
based not only on the current error, but also
the history of the error (integral of error) and
the anticipated future error (derivative of
error)
PID Control
• We will examine effect of PID control on a
canonical 2nd order system to gain insight

Controller Design
b
P( s)  2
s  as  b

Y (s) C ( s) P( s) bC ( s)
  2
R( s ) 1  C ( s) P( s) s  as  b(1  C (s ))
P Control
• Control effort is proportional to the amount of
error
Y (s) bK P
C ( s)  K P   2
R( s ) s  as  b(1  K P )

Controller Design
n2  b(1  K P )  n  b(1  K P )
a a
2n  a    
2n 2 b(1  K P )

• Cannot set ωn and ζ independently (only one

d.o.f.)
P Control
• Effect on steady-state performance
Steady-state value for a unit step reference
bK p 1
yss  lims0 sY (s)  lim s 0 s 2
s  as  b(1  K P ) s

Controller Design
bK P KP
 1
y(t)
=
ss error b(1  K P ) 1  K P
1
t

larger Kp makes yss 1 … error goes to zero

PD Control
• With derivative control, do not have to wait
for error to get large before control action
becomes large, control anticipates the error
• Never use D control by itself, amplifies noise

Controller Design
Y (s) b( K P  K D s )
C ( s)  K P  K D s   2
R( s ) s  (a  bK D ) s  b(1  K P )
• Not canonical form, but trends are
meaningful
• Can place two poles anywhere (2 d.o.f.)
PI Control
• Control effort gets larger as error is
accumulated  KI 
b K   P 
C (s)  K P 
KI

Y ( s)
  s 
s R( s)  KI 

Controller Design
s  as  b 1  K P 
2

 s 
b  KPs  KI 
 3
s  as 2  b 1  K P  s  bK I

• Cannot use canonical relations because third

order with a zero, but tends to make system
slower and more oscillatory
PI Control
• Effect on steady-state performance
Steady-state value for a unit step reference

yss  lims0 sY (s)

Controller Design
b  KPs  KI  1 bK I
 lim s 0 s 3  1
s  as  b 1  K P  s  bK I s bK I
2

therefore, steady-state error is zero for a step

reference, even for small KI (just takes longer to
reach steady state)
PID Summary
• Some intuition about the effect of the terms:
• Increasing KP: Same amount of error generates a
proportionally larger amount of control … makes
system faster, but overshoot more (less stable)

Controller Design
• Increasing KD: Allows controller to anticipate an
increase in error, adds damping to the system
(reduces overshoot) … can amplify noise
• Increasing KI: Control effort builds as error is
integrated over time, helps reduce steady-state
error, but can be slow to respond
Note: these guidelines do not hold for all situations
… look specifically at how poles are affected
PID Control
• For systems that are not canonical first or second
order, need to use trial and error … can look for
reduced-order approx
• Following are helpful, though not always true

Controller Design
KP ↑ KD ↑ KI ↑
ss error
↓ --- ↓
rise time
↓ --- ?
settling time
--- ↓ ? (↑)
overshoot
↑ ↓ ? (↑)
Module 13 Summary
• Control requirements
• Pole placement/algebraic approach to control
• Interpret requirements in terms of pole
location, transfer function parameters

Controller Design
• Find closed-loop transfer function in terms of
unknown gains
• Solve for gains to match desired
poles/parameters
• PID control
• Effect of KP, KD, KI