Systems Analysis and Control

Matthew M. Peet
Arizona State University

Lecture 1: Introduction to Control Systems and Historical Perspective

Rorschach Test for Control? Which One is Different?

Atlas Robot from Boston Dynamics

Lt. Cmdr. Data from Star Trek TNG

Star Wars Marionette

M. Peet Lecture 1: Control Systems 2 / 44

What do we mean by “Control”?
Control is the study of how to make things do what you want.
The most robust Controller is Human Consciousness.
• Fly an airplane.
• Drive a car without crashing.
The Subconscious is also not bad
• Keeping your eyes on the instructor.
• Keeping your heart beating.
• Standing up.

In this class we focus on Automatic Control.

• Machines are clay. To function, they must be controlled
I They only speak Math.
• Your subconscious is very good at Math.
Unfortunately, you conscious mind is NOT (no offense).
M. Peet Lecture 1: Control Systems 3 / 44
Who Am I?
Website: http://control.asu.edu

Research Interests: Computation, Optimization and Control

Focus Areas: Control Specialization:
• Control of Nuclear Fusion • Optimization
• Immunology • Control of Delayed Systems
• Thermostats, Renewable Energy, • Control of PDE Systems
and Power Distribution • Control of Nonlinear Systems
• Soft Robotics

My Background:
• B.Sc. University of Texas at Austin
• Ph.D. Stanford University
• Postdoc at INRIA Paris
• NSF CAREER Awardee
Office: ERC 253; Lab: GWC 531

M. Peet Lecture 1: Introduction 4 / 44

What is a Control System?
Well... What is a System?
• AKA: “process”, “machine”, “plant”, “thing”, et c.
• Can be a “real” thing or a mathematical representation (model).

Definition 1.
A System is anything with Inputs and Outputs

There should ALWAYS be Inputs and Outputs!

• If No Inputs: You can’t change anything.
• IF No Outputs: Then it doesn’t matter anyway.
Let start with some examples.
M. Peet Lecture 1: Introduction 5 / 44
Example: A Stereo Receiver

Inputs and Outputs depend on what we are trying to do.

• A system may have multiple Inputs and Outputs.
• But only some of these are relevant to you.

M. Peet Lecture 1: Introduction 6 / 44

Example of a System: Education

Definition 2.
The System to be controlled is called the Plant.

M. Peet Lecture 1: Introduction 7 / 44

Example of a System: Education
Societal Perspective

WASTE is ALSO an Output!

M. Peet Lecture 1: Introduction 8 / 44
Example of a System
Aircraft

M. Peet Lecture 1: Introduction 9 / 44

Example: The economy

This is a tough one...

• Consider a small-scale economy.

M. Peet Lecture 1: Introduction 10 / 44

Example: Dinner at House of Tricks

There can be Multiple Subsystems!

• You tip the waiter based on quality of service.
• The waiter can improve his service to increase his tip.
Feedback: When the Outputs affect the Inputs.

M. Peet Lecture 1: Introduction 11 / 44

Inputs and Actuators

Inputs can be created by Actuators.

Definition 3.
An Actuator is any mechanism/signal/communication which can affect the
Outputs.

Examples:
• Ailerons, Rudder
• Force Transducers: Servos/Motors
I Robots

I Engines

• Money (Prices)

M. Peet Lecture 1: Introduction 12 / 44

Sensors, Outputs and Feedback
Outputs are measured by Sensors.
Definition 4.
A Sensor is any mechanism/communication/signal which can be used to
measure an Output.

Definition 5.
A Closed Loop Controller creates a loop between Sensors and Actuators. This
loop is referred to as Feedback.

Control
System

Sensor

Plant

More Examples of Feedback:

• Pool Maintenance
M. Peet Lecture 1: Introduction 13 / 44
Back to the economy

Prices are the feedback

• Markets price goods by how much they are needed.
I Scarcity and Demand to Price

• People produce goods based on how much money they can make.
I Price and Cost of production to Production

M. Peet Lecture 1: Introduction 14 / 44

Control Systems

Definition 6.
A Control System is a system which modifies the inputs to the plant to
produce a desired output.

M. Peet Lecture 1: Introduction 15 / 44

Example of a System
Paper Mill

Things to look for:

• Dynamics
• Sensors
• Actuators
• Regulated Outputs
• Disturbances

M. Peet Lecture 1: Introduction 16 / 44

Example of a System
Nuclear Reactor

Things to look for:

• Dynamics • Regulated Outputs
• Sensors • Disturbances
• Actuators
M. Peet Lecture 1: Introduction 17 / 44
Example of a System
Pantograph (High-Speed Electric Train)

Things to look for:

• Dynamics • Regulated Outputs
• Sensors • Disturbances
• Actuators

M. Peet Lecture 1: Introduction 18 / 44

Of Course there is also Open-Loop Control

Definition 7.
An Open Loop Controller has an actuator, but makes no measurements.
• No way to tell how well you are doing...

M. Peet Lecture 1: Introduction 19 / 44

Examples of Open-Loop Control

• Choosing a Class
• A Pop-up Toaster
I Actuators, Inputs,

Outputs?
• Irrigation Systems

Problems?

M. Peet Lecture 1: Introduction 20 / 44

Lets Compare Open- and Closed-Loop Driving

3D Google Maps Driving Simulator

2D Google Maps Driving Simulator

2D Driving Simulator without Google Maps

M. Peet Lecture 1: Introduction 21 / 44

Model-Based vs. Model-Independent Control
A Last Bit of Nomenclature

Open-Loop Control requires detailed knowledge of the System and Environment.

• How much knowledge is needed?
• How is this knowledge represented?

Definition 8.
A Mathematical Model of a System is any tool which allows us to predict
the output of the system for any given input.
Examples: Differential Equations, Transfer Functions, Functions
• Comes from: Physical Principles; System Identification; Machine Learning

Definition 9.
A Model-Based Controller is a controller which uses a Mathematical Model of
the system to map inputs to outputs (sensor signals to actuator signals).

Definition 10.
A Model-Independent Controller is a controller which is designed to work for
ANY system.
M. Peet Lecture 1: Introduction 22 / 44
Automatic Control throughout History: The Measurement
of Time
Egyptian Water Clocks 1600BC

Significant in Commerce, Industry, Science, Medicine and Military

Time left is given by the amount of water left in the pot.

Problem: Measurement is limited to time left and by amount of water in pot.
M. Peet Lecture 1: Introduction 23 / 44
History of Water Clocks

Time passed is amount of water in pot.

Problem: Water flow varies by amount of water in the top pot.
Solution: Maintain a constant water level in top pot.
M. Peet Lecture 1: Introduction 24 / 44
History of Water Clocks

Problem: Manually refilling the top pot is labor intensive and inaccurate.
Solution: Design a control System (Inputs, Outputs?).

M. Peet Lecture 1: Introduction 25 / 44

History of Water Clocks
Ctesibius c. 220-285 BC

Father of pneumatics
• Lived in Abject Poverty
• Created most accurate clock until Huygens (1657 AD)
• Overshadowed by better-known student Heron (Hero) of
Alexandria

M. Peet Lecture 1: Introduction 26 / 44

History of Water Clocks
Heron (Hero) of Alexandria c. 10 AD

As any good student, Hero used Ctesibius’ water clock

to perform party tricks.

The self-replenishing wine bowl. (Inputs, Outputs?)

M. Peet Lecture 1: Introduction 27 / 44

History of Water Clocks
The Pipe Organ

Ctesibius himself applied the principle of pneumatic control to create a pipe

organ.

M. Peet Lecture 1: Introduction 28 / 44

History of Water Clocks
What do we use Ctsebius’ water clock for today?

M. Peet Lecture 1: Introduction 29 / 44

The Industrial Revolution
More Serious Applications

In addition to wine bowls, Heron also developed the steam engine.

Unfortunately, the Aeolipile was NOT CONTROLLED.

• No Feedback
• Missed chance for an industrial age?
M. Peet Lecture 1: Introduction 30 / 44
The Modern Aeolipile

M. Peet Lecture 1: Introduction 31 / 44

Modern (Relatively) Steam Engines
The Flyball Governor

Problem: To be useful, steam engines must rotate a piston at a fixed speed.

M. Peet Lecture 1: Introduction 32 / 44

The Flyball Governor

Flyballs regulate rotation rate.

• Faster rotation = More centrifugal
force.
• Centrifugal force lifts the flyballs,
which closes a valve, reducing flow
of steam.
• Reduced flow of steam decreases
engine/piston speed.
Gain: A key parameter is the ratio of
lift to throttle.

Identify the inputs and outputs.

M. Peet Lecture 1: Introduction 33 / 44

The Flyball Governor

M. Peet Lecture 1: Introduction 34 / 44

The Flyball Governor
Block Diagram Representation

Decompose the Problem:

• Inputs and Outputs
• Plant and Controller
• Disturbances?

Valve
Flyball

Boiler
Steam Engine
RPM Steam Pressure

M. Peet Lecture 1: Introduction 35 / 44

The Flyball Governor in Operation
Stuart-Turner No9 Steam Engine

Factoid: What does it mean to be going “Balls Out”?

M. Peet Lecture 1: Introduction 36 / 44

What are the challenges for control in the 21st century?

Megatrends:
• Increased Complexity (Embedded Computation and Control)
• Increased Connectivity (Internet of Things)
• Robots, Drones and Self-Driving Cars
• Increased Demands (Higher Standards)
• Mobile Computing (Mobile Apps)

M. Peet Lecture 1: Introduction 37 / 44

Challenges for Control in the 21st century
Privatization of Space Travel

Challenges
• Safety
• Complexity
• Uncertainty

Links:
Blue Origin Successful Landing
Blue Origin Successful Landing: Flight 3
SpaceX Landing, Second Attempt
Proton M launch Failure (FCS was for wrong rocket)
Kepler Space Telescope
M. Peet Lecture 1: Introduction 38 / 44
Challenges for Control in the 21st century
UAVs and Drones (Delay, Sampled-Data)
Safe Interaction with
• Crowded Airspace
• Real-Time Obstacle
Avoidance

Precision Control with

• Delayed Feedback
ẋ(t) = Ax(t) + Bu(t − τ )

• Lossy Connections
ẋ(t) = Ax(t) + Bu(tk )

Links:
X47 Drone Carrier Landing
Raff’s TED talk
M. Peet Lecture 1: Introduction 39 / 44
Challenges for Control in the 21st century
Self-Driving Vehicles
Challenges:
• Safety (Provable)
• Uncertainty (in model,
environment)
• Other Drivers (Multi-Agent)
• Obstacles

Self-Driving Vehicles
• Google (Waymo)
• Über
• Tesla, Mobileye
• Toyota, Nutonomy

Links:
Toyota’s Research Expansion in Automation
Uber’s self-driving Taxis are in Pittsburg
Self-Driving Cars Flood into Arizona
M. Peet Lecture 1: Introduction 40 / 44
Challenges for Control in the 21st century
Interconnectivity (Decentralized Control)

m N
0 15 30 60 Miles
Credits : layer1 : E sri, HERE, DeLorme, MapmyIndia, © O penS treetMap contributors, and the GIS user community; State Layers : ; layer0 : E sri, HERE ,
DeLorme, MapmyIndia, © O penStreetMap contributors, and the GIS user community

Surface Coal Mine Natural Gas Power Plant Wind Power Plant HGL Market Hub (z)
Underground Coal Mine Nuclear Power Plant Petroleum Refinery
Natural Gas Market Hub (z)
Biomass Power Plant Other Power Plant Biodiesel Plant
Electricity Border Crossing
Coal Power Plant Petroleum Power Plant Ethanol Plant
Natural Gas Pipeline Border Crossing
Geothermal Power Plant Pumped Storage Power Plant Natural Gas Processing Plant (z)
Hydroelectric Power Plant Solar Power Plant Ethylene Cracker

M. Peet Lecture 1: Introduction 41 / 44

Challenges for Control in the 21st century
Robotics (Hybrid and Nonlinear Dynamics, PDE systems)

HARD Robots
• Uncertain Terrain
• Interactions with the environment
If x(t)>0:
ẋ(t) = Ax(t)

If x1 (t)=0 AND x2 (t)<0: Set

x2 (t) = −x2 (t)
Link:
Boston Dynamics, Atlas Mark 3

SOFT Robots
• Infinite Degrees of Freedom
• Material Dynamics
Link:
Robotic Worm

M. Peet Lecture 1: Introduction 42 / 44

Challenges for Control in the 21st century
Arduino and Raspberry Pi
Trends:
• Rapid prototyping
• Internet of Things
• Control is Everywhere

Challenges
• Noisy Sensors
• Data-Driven Modeling
• Dynamics with logical switching
ẋ = Ax + Bu(t)

If Occupied=True :
u(t) = K1 x(t)

Else : u(t) = K2 x(t)

M. Peet Lecture 1: Introduction 43 / 44

Overview of Course Objectives
Will that be on the final???

Part 1: System Analysis

• Given a system model:
I Given an input, find the output
I Predict the effect of Standard Inputs (Impulse, Step, Ramp, etc.)
I Determine Stability
• Introduction to the Frequency Domain
• Given a desired response:
I Determine the characteristic root locations

Part 2: Controller Design

• Given a system model:
I Plot the effect of proportional gain (Root-Locus)
I Plot the response to sinusoidal input (Bode Plot)
• Given a desired response:
I Propose feedback controllers (PID and lead-lag) to achieve the desired

response.

M. Peet Lecture 1: Introduction 44 / 44