Control System II (EE16101)
Signal Processing in Digital Control
1. Introduction
2. Signals
3. Signal Processing
Introduction
. Signal processing emerged soon after World War I in the form of
electrical filtering.
. With the invention of the digital computer and the rapid
advances in VLSI technology during the 1960s, a new way of
processing signals emerged: digital signal processing.
. To start with, a classification of the various types of signals
encountered in today’s technological world is provided.
. Then the sampling process is described as a means of
converting analog into digital signals.
Signals
. Typically one assumes that a signal is an electrical signal, for
example, a radio, radar, or TV signal.
However, in DSP a signal is any quantity that depends on one or
more independent variables.
A radio signal represents the strength of an electromagnetic
wave that depends on one independent variable, namely, time.
Signals Cont’d
. In our generalized definition of a signal, there may be more
than one independent variable and the independent variables
may be any quantity other than time.
For example, a digitized image may be thought of as light
intensity that depends on two independent variables, the
distances along the x and y axes; as such a digitized image is, in
effect, a 2-dimensional signal.
A video signal is made up of a series of images which change
with time; thus a video signal is light intensity that depends on
the distances along the x and y axes and also on the time; in
effect, a video signal is a 3-dimensional signal.
. Some signals arise naturally, others are man-made.
Signals Cont’d
Natural signals are found, for example, in:
. Acoustics, e.g., speech signals, sounds made by dolphins and
whales
. Astronomy, e.g., cosmic signals originating in galaxies and
pulsars, astronomical images
. Biology, e.g., signals produced by the brain and heart
. Seismology, e.g., signals produced by earthquakes and
volcanoes
. Physical sciences, e.g., signals produced by lightnings, the
room temperature, the atmospheric pressure
Signals Cont’d
Man-made signals are found in:
. Audio systems, e.g., music signals
. Communications, e.g., radio, telephone, TV signals
. Telemetry, e.g., signals originating from weather stations and
satellites
. Control systems, e.g., feedback control signals
. Medicine, e.g., electrocardiographs, X-rays, magnetic
resonance imaging
. Space technology, e.g., the velocity of a space craft
. Politics, e.g., the popularity ratings of a political party
. Economics, e.g., the price of a stock at the TSX, the TSX index,
the gross national product
Signals Cont’d
Two general classes of signals can be identified:
. Continuous-time signals
. Discrete-time signals
Continuous-Time Signals
. A continuous-time signal is a signal that is defined at each
and every instant of time.
. Typical examples are:
– An electromagnetic wave originating from a distant galaxy
– The sound wave produced by a dolphin
– The ambient temperature
– The light intensity along the x and y axes in a photograph
. A continuous-time signal can be represented by a function
x(t) where − ∞ < t < ∞
Continuous-Time Signals Cont’d
x(t)
t
Discrete-Time Signals
. A discrete-time signal is a signal that is defined at discrete
instants of time.
. Typical examples are:
– The closing price of a particular commodity on the stock
exchange
– The daily precipitation
– The daily temperature of a patient as recorded by a nurse
Discrete-Time Signals Cont’d
. A discrete-time signal can be represented as a function
x(nT) where − ∞ < n < ∞
and T is aconstant.
. The quantity x(nT) can represent a voltage or current level or any
other quantity.
. In DSP, x(nT) always represents a series of numbers.
. Constant T usually represents time but it could be any other
physical quantity depending on the application.
Discrete-Time Signals Cont’d
x(nT)
nT
T
Nonquantized and Quantized Signals
. Signals can also be classified as:
– Nonquantized
– Quantized
. A nonquantized signal is a signal that can assume any value
within a given range, e.g., the ambient temperature.
. A quantized signal is a signal that can assume only a finite
number of discrete values, e.g., the ambient temperature as
measured by a digital thermometer.
Nonquantized and Quantized Signals Cont’d
x(t) x(nT)
t nT
(a) Continuous-time, nonquantized (b) Discrete-time, nonquantized
x(t) x(nT)
t nT
(c) Continuous-time, quantized (d) Discrete-time, quantized
Sampling Process
To be able to process a nonquantized continuous-time signal
by a digital system, we must first sample it to generate a
discrete-time signal.
We must then quantize it to get a quantized discrete-time
signal.
That way, we can generate a numerical representation of the
signal that entails a finite amount of information.
Sampling Process Cont’d
A sampling system comprises three essential components:
– sampler
– quantizer
– encoder
Sampling Process Cont’d
Sampler
x(nT) xq(nT)
x(t) Quantizer Encoder xq'(nT)
Clock
nT
Sampling system
Sampling Process Cont’d
A sampler in its bare essentials is a switch controlled by a
clock signal which closes momentarily every T seconds
thereby transmitting the level of the input signal x (t) at
instant nT, i.e., x(nT), to its output.
Parameter T is called the samplingperiod.
Sampler
x(nT) xq(nT)
x(t) Quantizer Encoder xq'(nT)
Clock
nT
Sampling system
Sampling Process Cont’d
A quantizer is a device that will sense the level of its input and
produce as output the nearest available level, say, xq(nT), from a
set of allowed levels, i.e., a quantizer will produce a quantized
continuous-time signal.
Sampler
x(nT) xq(nT)
x(t) Quantizer Encoder xq'(nT)
Clock
nT
Sampling system
Sampling Process Cont’d
An encoder is essentially a digital device that will sense the
voltage or current level of its input and produce a
corresponding binary number at its output, i.e., it will convert a
quantized continuous-time signal into a corresponding discrete-
time signal in binary form.
Sampler
x(nT) xq(nT)
x(t) Quantizer Encoder xq'(nT)
Clock
nT
Sampling system
Sampling Process Cont’d
The sampling system described is essentially an
analog-to-digital converter and its implementation can assume
numerous forms.
These devices go by the acronym of A/D converter or ADC
and are available in VLSI chip form as off-the-shelf devices.
Sampler
x(nT) xq(nT)
x(t) Quantizer Encoder xq'(nT)
Clock
nT
Sampling Process Cont’d
A quantized discrete-time signal produced by an A/D
converter is, of course, an approximation of the original
nonquantized continuous-time signal.
The accuracy of the representation can be improved by
increasing
– the sampling rate, and/or
– the number of allowable quantization levels in the quantizer
The sampling rate is simply 1/T = fs in Hz or 2π/T = ωsin radians
per second (rad/s).
Sampling Process Cont’d
Once a discrete-time signal is generated which is an accurate
representation of the original continuous-time signal, any
required processing can be performed by a digital system.
If the processed discrete-time signal is intended for a person,
e.g., a music signal, then it must be converted back into a
continuous-time signal.
Just like the sampling process, the conversion from a discrete-
to a continuous-signal requires a suitable digital-to-analog
interface.
Sampling Process Cont’d
Typically, the digital-to-analog interface requires a series of two
cascaded modules, a digital-to-analog (or D/A) converter and a
smoothing device:
D/A y′(nT) Smoothing
y(nT) y(t)
converter device
Sampling Process Cont’d
A D/A converter will receive an encoded digital signal in
binary form like that in Fig. (a) as input and produce a
corresponding quantized continuous-time signal such as that
in Fig. (b).
The stair-like nature of the quantized signal is, of course,
undesirable and a D/A converter is normally followed by some
type of smoothing device, typically a lowpass filter, that will
eliminate the uneveness in the signal.
y(nT) y'(t)
t
nT
(a) (b)
Sampling Process Cont’d
Complete DSP system
Smoothing
Sampler Quantizer Encoder Digital system D/A converter
device
x(nT) xq(nT) x'q(nT) y(nT) y′(nT)
x(t) y(t)
Clock
nT
Sampling Process Cont’d
The quality of the conversion from a continuous- to a discrete-
time signal and back to a continuous-time signal can be
improved
– by understanding the processes involved and/or
– by designing the components of the sampling system carefully.
