01 Introduction

Lecture: Introduction
Automatic Control 1
Prof. Alberto Bemporad
University of Trento
Academic year 2010-2011
Faculty of Engineering
Prof. A. Bemporad (University of Trento) Automatic Control 1 Academic Year 2010/2011

Lecture: Introduction Outline
Lecture outline
What is automatic control ?
Application examples
Course information
Prof. A. Bemporad (University of Trento) Automatic Control 1 Academic Year 2010/2011 2 /50
Lecture: Introduction What is control ?
What is control ? A real life example ...

shower
water knobs disturbance
d(t)
desired temperature water inflow water temperature
r(t) u(t) y(t)
skin sensations
Water inflow u(t) must be controlled to reach and maintain

the desired temperature r(t)
Sensors on skin measure water temperature y(t)
Water inflow u(t) manipulated so that y(t)≈r(t) …
… in spite of flow and temperature fluctuations d(t)
What is control ? A real life example ...

steering wheel vehicle on road
disturbance
d(t)
desired lateral steering angle lateral displacement
u(t) y(t)
displacement r(t)
eyes watching current displacement
Steering wheel must be controlled to reach and maintain

the desired lateral displacement r(t) within the lane
(e.g.: staying in the middle of the lane)
Eyes measure current later displacement y(t)
Steering wheel u(t) manipulated so that y(t)≈r(t) …
… in spite of changes of road curvature and of r(t)
What is automatic control ?

embedded control unit process
disturbance
d(t)
reference manipulated input output
r(t) u(t) y(t)
measurements
How to control the inputs u(t) to the process automatically performance

to make the output y(t) track the given reference r(t) ?
How to exploit the measurements of y(t) to track the reference

r(t) in spite of disturbances d(t) acting on the process ? robustness
What is control ? A real life example …

Mom is washing hands,
she drains hot water
(=disturbance entering
the process)
disturbance
d(t)
r(t) u(t) y(t)
measurements
Lecture: Introduction Examples of control systems
Application areas of control engineering

Aeronautics & aerospace
Automotive
Manufacturing
Process control (chemical, pharmaceutical,
steel, pulp & paper, ...)
Power electronics
Telecommunications
Environmental systems
Financial engineering
Supply chains
Power networks
...
Adaptive Cruise Control (ACC)
Adaptive Cruise Control (ACC)
Electronic Control Unit vehicle

road slope
d(t) speed/position
desired speed engine torque/brakes y(t)
r(t) u(t)
measurements
Lane Warning !"#$%&$'$(')*#%+,%-."'/*01%

HALF - HAptical Lane Feedback
!"#$%&
'()"%)*"
+(,$
Optical sensor detects and tracks lane markings
VIP ECU
!"#"$%&'(#)*+
Incipient lane change signalled!"#"$%&'(#)*+
A N E WA RN I N G
to the driver
!-./0&1
by: components
• Automotive
234"%$5
• Power supply: 12Vbat - 500mA
ptical sensor detects andontracks lane markings.
- Haptic signal driver seat (simulation
6"*()-" of• analogue
rumblevideo
strip)
inputs
- Haptic
gnals the feedback
driver incipient on steering
change by means of:• DSP @ 60Mhz
wheel
of lane!&#$%&-
Haptic- feedback
• 2MB DRAM + SRAM
Acousticondirectional signals
steering wheel (HALF)
785$"9 • CAN + RS-232 interface
Haptic signal on driver seat (simulation of rumble strip)
• many configurable Digital I/ Os
Lane can be automatically maintained
Acoustic directional signals
')**/"%4&'(#)*+
Orbassano
')**/"%4&'(#)*+
17th Dec. 2002
electric power steering
Prof. A. Bemporad (University of Trento) Product

Automatic Control 1 & Process Engineering Academic Year 2010/2011 10 /50
Lane Warning
%
Lane Warning
desired lateral Driver

Electronic Control Unit d(t) !"#"$"%
displacement Torque Vehicle (Fiat Brava 1600)
r(t) Torque Steer
Demand actuated
u(t) Demand
lateral
00
!"#$%
!"#$% &'()*+,-.,)-% /.+%y(t)
&'()*+,-.,)-% /.+% 0''*%
0''*% &'1'2)3'&%
&'1'2)3'&% 04%
04% 5676$6
5676$6
displacement
8+9*:%.%$9.,%;-.1.%<=>>6
8+9*:%.%$9.,%;-.1.%<=>>6
&'()"*)"+,-./00/
##
measurements
--
Idle speed control

Objective:
maintain the engine speed at a given rpm
Torque disturbance: power steering Power steering + air conditioning

Active suspensions
Active Suspension System

Ford Mercur XR 40i
active
suspensions
passive
suspensions
(Semi)active suspensions
Ms=suspended mass
Mus=unsprung mass
x4 sprung mass
velocity
suspension
deflection
x2 unsprung mass
velocity
tire
deflection
active suspensions
semiactive suspensions
semiactive suspensions
Traction control
Problem: Improve driver's ability to control a vehicle under adverse
external conditions (wet or icy roads)
indoor tests at ice arena (μ¼ 0.2)
2000 Ford Focus, 2.0l 4-‐cyl engine

5-‐speed manual transmission
Air to Fuel Ratio (AFR) control

Control objectives:
Maintain the stoichiometric Air to Fuel Ratio (AFR) and avoid oxygen saturation
(empty or full catalyst)
Fuel Fuel injector

Tank (Actuator)
Air Exhaust gas Exhaust gas
Delay
Air/Fuel
mixture
UEGO Three Way Catalyst
(Sensor)
Control of suspensions
(Courtesy of Daimler-‐Chrysler)
Active steering
Courtesy of
Segway Human Transporter
www.segway.com

The Segway™ Human Transporter (HT) is the first
self-balancing, electric-powered transportation device.
With dimensions no larger than the average adult body
and the ability to emulate human balance, the Segway
HT uses the same space as a pedestrian, and can go
wherever a person can walk.
Dynamic stabilization is the essence of the Segway

Human Transporter (HT). Dynamic stabilization
enables Segway HT to work seamlessly with the body’s
movements.
Gyroscopes and tilt sensors in Segway HT monitor a user’s center of

gravity at about 100 times a second. When a person leans slightly
forward, Segway HT moves forward. When leaning back, Segway HT
moves back.
Short historical notes

Water level controllers date back to ancient Greece
(Ktesibios of Alexandria, water clock, 270 B.C.), (Philon of Byzantium, 250 B.C.)
(Heron of Alexandria, ≈100 A.D.)
Other examples in the 17th-18th century

(Cornelis Drebbel’s incubator, 1620), (Edmund Lee’s self-regulating wind machine, 1745)
(Thomas Mead’s lift tenter and speed controller, 1787)
James Watt’s flyball governor (1788)
Used to regulate the speed of steam engines.

If speed increases, flyballs spread apart and
the steam flow through the throttle is reduced.
And vice versa.
Frequency domain (1930-1950)

(H. Nyquist, H.W. Bode, N.B. Nichols)
State-space and optimal control (1950-1980)

(R. Kalman, R. Bellman)
Nonlinear, robust, adaptive, optimization-based, (…) control (1980-today)

Is control needed after all ?
Friday November 7, 1997
Baby Benz falls foul of a moose
“Mercedes is recalling 3,000 brand-new

A Class mini cars to correct stability
problems revealed in Scandinavian
tests.”
A Class failed the “moose test”, simulating a sudden swerve
to avoid a moose on the road.
The recall of about 3,000 cars followed motoring press

reports which indicated the A Class is unstable, and likely
to roll over, in extreme lane-changing tests.
Problem solved by introducing the Electronic Stability

Program (ESP), which automatically manipulates brakes
(and engine torque) to keep skid under control.
EC – DG INFSO – SMART 2007/047
WORKSHOP: Brussels the 9th of October 2
Is control needed Worldwide

after all ?Monitoring & Control Markets
Monitoring and control: today's market, its evolution till 2020 and the impact of ICT on these
EC – DG INFSO – SMART 2007/047

Monitoring and control: today's market, its evolution till 2020 and the impact of ICT on these WORKSHOP: Brussels the 9th of October 2008
Worldwide Monitoringby&application Control Markets

by application Total world: !187,9 bn
Worldwide Monitoring Process &industries,
•
Control
industries market
Factory automation, the sum of Manufacturing plus
remains the main market with 58 Vehicles 56,4
(=188
• billions
Factory euro)
automation,
billion euros, comparable to Vehicles (56 billion):
the– sum of Manufacturing plus
Total world: !187,9 bn
With 38 billion euros out of 58, Services dominates with 2/3 of Industry
31,5
Process industries,
industries remainstotal
theFactory
mainAutomation
market with 58
market.
Vehicles
Manufacturing
56,4 57,8
billion euros, comparable to Vehicles (56 billion):
– Among Services the sub segment Integration
Integration, installation and Industry
Monitoring & control ≈ semiconductor
training accounts for 50% of them. industry
– With 38 billion euros out of 58, Services dominates with 2/3 of Industry
Process 26,3
31,5
≈ 2 x mobile phone
total Factory
•
industry
Automation market. Manufacturing
Together, three application markets, Vehicles,
Healthcare 18,3
57,8
– Among Services the sub segment Integration,
Integration installation and Industry Critical
M
Manufacturing
f t i and d Process
P industries
i d t i representt 60% off
Process 26,3 18,0
training accounts for 50% of them. Infrastructures
total Monitoring & Control market.
Still growing, young areas with great potentialsHealthcare Logistics &
18,3
transport
10,9
• Together, three application markets,
betweenVehicles,
(power
M
grids,
Manufacturing
f t i and
building
• Ranked
d Process
automation,
P applications
industries
i d tmarkets
i represent
environment)
10 and 20 billion
t 60%
are in order
euros the next three
: off Critical Environment 8,6
Infrastructures 18,0
total Monitoring & Control– market.
Healthcare
Power grids
– Critical infrastructures Logistics & 7,2
transport
10,9
– Logistic
g & transport
p
• Ranked between 10 and 20 billion euros the next three Building 7,2
Environment 8,6
applications markets
• are in one,
Last order :
Home is, for the moment, a small niche market. Household
appliances 4,3
– Healthcare Hard
Power grids 7,2 Soft
– Critical infrastructures Homes 0,7
Serv
– Logistic
g & transport
p & a presentation prepared by:
A report
Building 7,2
Available for download: http://www.decision.eu/smart2007.htm 25
• Last one, Home is, for the moment, a small niche market. Household
appliances 4,3
Hard
Homes 0,7 Soft
Serv
A report & a presentation prepared by:
Available for download: http://www.decision.eu/smart2007.htm 25
Process control
y y
Control is heavily used in the process industries
Ifwj » wi ! jthe controller does its best to track rj, sacrificing ri tracking if
y
necessary. If w j = 0 , on the other hand, the controller completely ignores
Usually slow processes (heat transfer, chemical reactions, distillation, etc.)
deviations rj–yj.
Choosing the weights is a critical step. You will usually need to tune your
controller, varying the weights to achieve the desired behavior.
Usually several input and output variables
As an example, consider Figure 1-3, which depicts a type of chemical reactor (a
CSTR). Feed enters continuously with reactant concentration CAi. A reaction
takes place inside the vessel at temperature T. Product exits continuously, and
contains residual reactant at concentration CA (<CAi).
The reaction liberates heat. A coolant having temperature Tc flows through
Example: Continuous stirred-tank reactor
coils immersed in the reactor to remove excess heat.
CAi
Tc
T CA
Figure 1-3: CSTR Schematic

Manipulated inputs:
From thereactant concentration
Model Predictive CAi inpoint
Control Toolbox feedfor view, T and CA would be
coolant
plant outputs, and temperature Tcbe inputs. More specifically, CAi would be
CAi and Tc would
an independent disturbance input, and Tc would be a manipulated variable
(actuator).
Controlled outputs:
reactant
There is one concentration
manipulated variable (the Ai in vessel
Ccoolant temperature), so it’s impossible
to hold both T and
vessel CA at setpoints.
temperature T Controlling T would usually be a high
priority. Thus, you might set the output weight for T much larger than that for www.abacusengg.in
CA. In fact, you might set the CA weight to zero, allowing CA to float within an
acceptable region (to be
operating (University
Prof. A. Bemporad of defined
Trento) by constraints). Automatic Control 1 Academic Year 2010/2011 26 /50
Control systems example: Ball and plate
Control systems example: Ball and plate
commands
control ball &
station measurements plate
Robotics
Robotics
Snake Robot (Biorobotics and Biomechanics Lab, Technion, Israel)
Functional electrical stimulation
Aeronautics and aerospace

Control systems heavily used in:
Aircrafts (roll, pitch, yaw)
Helicopters
Satellites
Aeronautics and aerospace. Quadcopter example

Parrot AR Drone
• 6 DoF Inertial Measurement Unit
(3 accelerometers, 3 gyros)
•468MHz ARM9 CPU running Linux

•Ultrasound altimeter/ground detector
•Two cameras:
- Vision-based stabilization (176x144)
- Wifi streaming (640x480) 7 (27)
Prepared Title
M. Lefébure AR Drone Developer Guide
Approved Date Revisio File
quadcopter
n
05/01/2010 1.0 AR Drone Developer Guide Release 1.0.doc
flight control unit
desired angles engine torques angles/positions

r(t) u(t) y(t)
measurements
• 2D-tag: magnetic cylindrical tag

z ψ y
φ
• Manipulated inputs f2
f3 τ2
– total force u f1
– torques τ̃θ , τ̃φ, τ̃ψ τ3
f4 l τ1 x
mg
• Outputs to regulate τ4 θ
x, y, z, θ, φ, ψ, ẋ, ẏ, ż, θ̇, φ̇, ψ̇
mẍ = −u sin θ − β ẋ
mÿ = u cos θ sin φ − β ẏ
mz̈ = u cos θ cos φ − mg − β ż
θ̈ = τ̃θ
φ̈ = τ̃φ
ψ̈ = τ̃ψ
Outputs: position (x,y,z) Inputs: torques and force
Financial engineering. Example: Option hedging

• The financial institution sells a synthetic option to a customer and gets y(0) (€)
• Such money is used to create a portfolio y(t) of underlying assets (e.g. stocks)
whose quantities at time t are u1(t), u2(t), ..., un(t)
• At the expiration date T, the option is worth the payoff r(T) = wealth (€) to be
returned to the customer
Portfolio (*) vs. Payoff
20 100
18 payoff r(T)
80
16
14 option price 60
How to adjust dynamically

12
10 portfolio wealth 40
8 the portfolio so that

6 wealthy(T) = payoff r(T) ? ... 20
4
0
wealth y(T)
2
0 !20
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0 20 40 60 80 100 120 140 160 180 200
time (years) asset price at expiration
Stock price at expiration
Lecture: Introduction Control systems technology
A typical control system
D 1001
A
A 1001 D
D/A converters actuators sensors A/D converters
process
?
control unit
Most used sensors and actuators in control systems
Sensors
temperature
pressure
flow
level
velocity, position
acceleration
force (strain) / deformation
Actuators
electrical motors (DC, brushless, step)
pumps
valves
heaters
Sensors and actuators in control systems
angular position thermocouple pressure sensor
brushless motor
strain gage
liquid flow sensor pump

valve
Lecture: Introduction Open-loop vs. closed-loop control
Open-loop vs. closed-loop control

Closed-loop control (feedback control)
Measurements of the output variables are fed back to the process through the controller
disturbance
measurements
Open-loop control (feedforward control)

The manipulated input variable is generated without measuring the output variable
Lecture: Introduction Control systems design
How to design a (modern) control system

Understand the automation problem:
- Which variables can be manipulated by actuators ?
- What are the output variables of interest ?
- What should we measure ?
- Which are the disturbances ?
Get a reliable simulation model
Get a simplified mathematical model of the main process dynamics
Use design techniques to synthesize the control algorithm
Test in simulation, validate on real process
Control design requires a dynamical model of the open-loop process
Lecture: Introduction Course contents
What you will learn in this course

The tools to study the mathematical properties of dynamical
systems (control theory is often called “systems theory”)
Building simple dynamical models of common physical

processes
Computer-aided tools for analysis, simulation, and control of

dynamical systems (in MATLAB™)
Lecture: Introduction Course contents
What you will learn in this course

Analysis of dynamical systems in continuous time
(differential equations, Laplace transform, stability)
Models of dynamical systems (electrical, mechanical, hydraulic,...)
Analysis of dynamical systems in discrete-time

(difference equations, Zeta transform)
State-feedback control synthesis in the time domain (pole-placement)
State estimation
Output feedback control (dynamic compensator)
Integral action
Automatic Control 2: Frequency-domain analysis and synthesis (loop shaping)

and more advanced control techniques (optimal, nonlinear, predictive, ...)
Lecture: Introduction Textbooks
Suggested references
A. Bemporad - Lecture notes (what you’re looking at right now …)
(maybe enough …)
K.J. Åström, B. Wittenmark, Computer-controlled Systems,

Theory and Design, Prentice-Hall
(good classical textbook on digital control)
G.F. Franklin, J.D. Powell, M. Workman “Digital Control of

Dynamic Systems”, Addison-Wesley Longman.
(good classical textbook on digital control)
T. Kailath, “Linear Systems”, Prentice-Hall

(advanced textbook for state-space concepts)
Lecture: Introduction Textbooks
Suggested references
E. Fornasini, G. Marchesini, “Appunti di Teoria dei Sistemi” ,
Edizioni Libreria Progetto (in Italian)
(advanced textbook for state-space concepts)
P. Bolzern, R. Scattolini, N. Schiavoni, “Fondamenti di

controlli automatici” (in Italian)
(good textbook for frequency-domain)
K.J. Åström, R.M. Murray, “Feedback Systems: An Introduction

for Scientists and Engineers”, (available on-line for download)
(new textbook, plenty of examples)
Lecture: Introduction Logistics
Logistics
Midterm exam: April 21, 2011. No expiration date.

= Final exam of Automatic Control 1 (TLC)
= Exam of AC2 for those elder mechatronic students that already
passed AC1.
Final exam: sometimes in June 2011 (AC1 + AC2).

Note: Submitting a new AC1 test automatically kills the midterm test !
Classroom exercises
Classroom exercises in MATLAB™
Web site
http://www.ing.unitn.it/~bemporad/teaching/ac
Prof. A. Bemporad (University of Trento) 30 novembre 2010
Lecture: Introduction Logistics
Logistics
Lectures Wednesday 10.30 13.00 Room A102

Thursday 10.30 13.00 Room A102
Friday 10.30 13.00 Room A102
Appointments for questions & explanations: send an email
Teaching assistants:
Matteo Rubagotti http://www.ing.unitn.it/~rubagott

Sergio Trimboli http://www.ing.unitn.it/~trimboli
Lecture: Introduction End of lecture
Questions ???

01 Introduction

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

01 Introduction

Загружено:

Авторское право:

Доступные форматы

Lecture: Introduction

Academic year 2010-2011

Prof. A. Bemporad (University of Trento) Automatic Control 1 Academic Year 2010/2011

What is automatic control ?

What is control ? A real life example ...

Water inflow u(t) must be controlled to reach and maintain

Sensors on skin measure water temperature y(t)

Water inflow u(t) manipulated so that y(t)≈r(t) …

… in spite of flow and temperature fluctuations d(t)

What is control ? A real life example ...

eyes watching current displacement

Steering wheel must be controlled to reach and maintain

Eyes measure current later displacement y(t)

Steering wheel u(t) manipulated so that y(t)≈r(t) …

… in spite of changes of road curvature and of r(t)

What is automatic control ?

How to control the inputs u(t) to the process automatically performance

How to exploit the measurements of y(t) to track the reference

What is control ? A real life example …

Application areas of control engineering

Adaptive Cruise Control (ACC)

Adaptive Cruise Control (ACC)

Electronic Control Unit vehicle

Lane Warning !"#$%&$'$(')*#%+,%-."'/*01%

electric power steering

Prof. A. Bemporad (University of Trento) Product

desired lateral Driver

Idle speed control

Torque disturbance: power steering Power steering + air conditioning

Active Suspension System

indoor tests at ice arena (μ¼ 0.2)

2000 Ford Focus, 2.0l 4-­‐cyl engine

Air to Fuel Ratio (AFR) control

Fuel Fuel injector

Air Exhaust gas Exhaust gas

(Courtesy of Daimler-­‐Chrysler)

Segway Human Transporter

Segway Human Transporter

Dynamic stabilization is the essence of the Segway

Gyroscopes and tilt sensors in Segway HT monitor a user’s center of

Segway Human Transporter

Short historical notes

Other examples in the 17th-18th century

James Watt’s flyball governor (1788)

Used to regulate the speed of steam engines.

Frequency domain (1930-1950)

State-space and optimal control (1950-1980)

Nonlinear, robust, adaptive, optimization-based, (…) control (1980-today)

Is control needed after all ?

Friday November 7, 1997

Baby Benz falls foul of a moose

“Mercedes is recalling 3,000 brand-new

The recall of about 3,000 cars followed motoring press

Problem solved by introducing the Electronic Stability

Is control needed Worldwide

EC – DG INFSO – SMART 2007/047

Worldwide Monitoringby&application Control Markets

Available for download: http://www.decision.eu/smart2007.htm 25

Figure 1-3: CSTR Schematic

Control systems example: Ball and plate

Control systems example: Ball and plate

Snake Robot (Biorobotics and Biomechanics Lab, Technion, Israel)

Functional electrical stimulation

Lane Warning !"#$%&$'$(')#%+,%-."'/01%

2000 Ford Focus, 2.0l 4-‐cyl engine

(Courtesy of Daimler-‐Chrysler)