Lec 1.

Control Systems
• Reading: Chapter 1

• What is control?
– Use of algorithms and feedback to affect the operation
of physical objects of interest (systems) to achieve
some desired performance

Input output
Physical Objects:
Plant, Process, System

Example: Room Temperature Control

Turning on/off heater/cooler to maintain the room temperature at

a pre-specified level

Similar examples: Cruise control of cars (maintaining constant speed)

2
 Example: Robotic Arm

Control the voltage applied on the motors so that the robot hand
moves along a specified trajectory (and grasp some object)

Example: Air Traffic Control

National airspace

Air traffic controllers give directions for pilots so that aircraft can arrive at their
destinations in time and safely.

Similar example: Internet congestion control

3
 Some Observations
• Control is everywhere
– Other examples: electrical, mechanical, ecological, and financial
systems

• Physical process under control can be very complicated

– Air traffic control example
– Bioengineering example

• Often need to work in adverse situations

– Environmental noises
– Part failure
– Human errors, etc

Two Perspectives of Systems

• Black box
– Learn by training
– Example: adaptive neural network, fuzzy logic, expert systems
– Advantage: no need of physical understanding
– Disadvantage: hard to analyze, not suitable for high performance sys.

• Model-based (this course)

– Build a math model to relate the system input and output
– Advantage: easy to analyze, high performance
– Disadvantage: physical models not always available

4
 Example: Satellite

gas jet

(courtesy of Lawrence Berkeley National Laboratory)

Suppose that the antennas of the satellite need to point to the earth.
Thus the satellite can only rotate around a particular axis

Satellite is a system
Input: A force F can be generated by reaction jet, which is controllable
Output: orientation of the satellite is given by the angle θ

Satellite System Model

Newton’s Second Law:

gas jet

Angular acceleration
Total torque Moment of interia

Take Laplace transform (assuming

Torque generated by the thrust F: zero initial condition)

System model (frequency domain):

System model (time domain):
i.e., transfer function:

5
 Example: Servomotor System
armature
resistance
Basic motor properties:
Generated torque proportional
to current:

Motor voltage proportional to

shaft angular velocity:
friction B

Servo System is an electro-mechanical system

Input: voltage source e(t), which can be controlled
Output: shaft angular position θ(t)

Basic circuit properties (KVL): Newton law:

DC motor model
System model (time domain):

System model (frequency domain):

6
 Task of Control
• For a given system, adjust the input signal in a
proper manner to achieve, e.g.,
– Reduced steady-state error
– Faster transient response
– Disturbance rejection
– Low sensitivity to model error

Open Loop vs. Closed Loop Control

Open Loop Control

Input output
Physical Objects
Controller
Plant, Process, System

Controller determines the controlled input directly (in a pre-programmed way)

Closed Loop (Feedback) Control

Input output
Physical Objects
Controller
Plant, Process, System

Using output to help determine the controlled input (emphasis of this course)

7
 Control = Sensing + Computation + Actuation

Actuate Sensing
(gas pedal) (speedometer)

Computation
(human controller)

We will focus on computation part.

Sensing and Actuation parts will be covered by other courses

Modern Digital Control Systems

noises noises

Actuators System Sensors

process
y(t)
ū(t)
Digital controller
Data hold
y(kT )
u(kT ) D/A Computer A/D k = 0, 1, . . .
k = 0, 1, . . .

• The signal that the digital controller bases its decision on is sampled at
discrete time
• The command the digital controller sends to actuators is also discrete time

8
 Antenna Pointing System
Voltage proportional to desired angle

Difference Power
Amplifier Amplifier Motor θ
Vi
+ Vi − Vo K1
K s(s+a)
−

Vo
Data Hold Digital Shaft Encoder
Voltage proportional to angle θ
Binary code

Course Content
• Digital System Modeling
– How to construct math model of a digital system?

• Digital Control System Analysis

– How does system output change under certain inputs?

• Digital Controller Design

– How to design digital controller so that the system
output behaves desirably?