Unit Step Function

Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Unit Step Function

Related terms:

Delta Function, Dirac Delta Function, Fourier Transforms, Heaviside Unit Step,
Initial-Value Problem, Laplace Transform, Heaviside, Laplace Transforms, Unitstep

View all Topics

Laplace transform
Huw Fox, Bill Bolton, in Mathematics for Engineers and Technologists, 2002

Laplace transforms for step and impulse function


Consider the unit step function u(t) shown in Figure 6.3. The Laplace transform is
given by equation [1] as:

Figure 6.3. Unit step at time t = 0

[2]

Thus a unit size step input signal to an engineering system occurring at time t = 0
will have a Laplace transform of 1/s.

Key points

The Laplace transforms of signals commonly used as inputs to systems are:

Unit impulse: 1

Unit step: 1/s

Unit ramp: 1/s2


Now consider obtaining the unit impulse function (represented as (t)). Such an
impulse can be considered to be a unit area rectangular pulse which has its width
k decreased to give the unit impulse in the limit when k → 0. For the unit area
rectangular pulse shown in Figure 6.4, the Laplace transform is:

Figure 6.4. Unit area rectangular pulse

We can replace the exponential by a series, thus obtaining:

Thus in the limit as k → 0, the Laplace transform tends to the value 1 and so:

[3]

Thus a unit size impulse input signal occurring at time t = 0 to an engineering system
will have a Laplace transform of 1.

> Read full chapter

Introduction to the Laplace Transform


Martha L. Abell, James P. Braselton, in Introductory Differential Equations (Fourth
Edition), 2014

Piecewise Defined Functions: The Unit Step Function


An important function in modeling many physical situations is the unit step function
, shown in Figure 8.1 and defined as follows.

Figure 8.1. Graph of (t − a).


Definition 35 (Unit Step Function). The unit step function (t − a), where a is a given
number, is defined by

Example 8.4.1

Graph (a) (t − 5) and (b) (t).

Solution
(a) In this case, the jump occurs at t = 5. We graph (t − 5) in Figure 8.2(a). (b) Here
(t) = (t − 0), so (t) = 1 for t ≥ 0. We graph this function in Figure 8.2(b).

Figure 8.2. (a) Graph of (t − 5). (b) Graph of (t).

The unit step function is useful in defining functions that are piecewise continuous.
Consider the function f(t) = (t − a) − (t − b). If t < a, then f(t) = 0 − 0 = 0. If a ≤ t < b, then
f(t) = 1 – 0 = 1. Finally, if t ≥ b, then f(t) = 1 − 1 = 0. Hence, Thus, we can define the
function using the unit step function as g(t) = h(t)((t − a) − (t − b)), which is illustrated
in Figure 8.3. Similarly, a function such as can be written in terms of the unit step
function as
Figure 8.3. (a) Graph of a function h(t). (b) h(t)(t − a). (c) h(t)(t − b). (d) g(t) = h(t)(t − a-
) − h(t)(t − b) = h(t)(t − a) − (t − b). The graph of the function obtained by subtracting
h(t)(t − b) from h(t)(t − a) is the graph of the function h(t) between a and b.

Thus, in terms of the unit step function, a function such as is expressed as f(t) = g(t-
)[(t − a) − (t − b)] + h(t)[(t − b) − (t − c)]. See Figure 8.4.

Figure 8.4. (a) Graph of g(t) in black and h(t) in gray. (b) Graph of g(t)[(t − a) − U (t − b)]
in black and h(t)[(t − a) − (t — b)] in gray. (c) Graph of g(t)[(t − a) − (t − b)] + h(t)[(t − a)
— (t − b)].

Write in terms of the unit step function.

The reason for writing piecewise-continuous functions in terms of the unit step
function is because we encounter functions of this type when solving initial-value
problems. Using the methods in previous chapters, we solve the problem over each
subinterval on which the function was continuous (i.e., “each piece of the function”).
However, the method of Laplace transforms can be used to avoid the repeated
calculations encountered then. We start to explain the process with the following
theorem.

Theorem 47. Suppose that F(s) =  {f(t)} exists for s > b ≥ 0. If a is a positive constant,
then

Proof. Using the definition of the Laplace transform, we obtain

Changing the variables with u = t − a (where du = dt and t = u + a) and changing the
limits of integration, we have

Example 8.4.2

Find (a) [(t − a)}, a > 0; (b) [(t − 3)5 (t − s)}; and (c) {sin(t – π/6)(t − π/6)}

Solution
(a)Because (b)In this case a = 3 and f(t) = t5. Thus,(c)Here a = π/6 and f(t) = sin t.
Therefore,Find {cos (t − π/6) (t − π/6)}.

In most cases, we must calculate {g(t) (t − a)} rather than {f(t − a)(t − a)}. To solve
this problem, we let g(t) = f(t − a), so f(t) = g(t + a). Therefore,

Example 8.4.3

Calculate (a) {t2 (t − 1)} and (b) {sin t (t − π)}.

Solution
(a)Because g(t) = t2 and a = 1,(b)In this case, g(t) = sin t and a = π. Notice that sin(t +
π) = sin t cos π + cos t sin π = − sin t. Thus,Find {cos t (t − π)}.

From the previous theorem, it follows:

Theorem 48. Suppose that F(s) = L{f(t)} exists for s > b ≥ 0. If a is a positive constant and
f (t) is continuous on [0,∞), then

Example 8.4.4

Find (a) and (b)

Solution
(a)If we write the expression e− 4ss− 3 in the form e-asF(s), we see that a = 4 and F(s) = s− 3.
Hence, and(b)In this case, a = π/2 and F(s) = 1/(s2 + 16). Then, and
With the unit step function, we can solve initial-value problems that involve piece-
wise- continuous functions.

Example 8.4.5

Solve subject to y(0) = y (0) 0.

Solution
To solve this initial-value problem we must compute L{f (t)}, where

Because this is a piecewise-continuous function, for t ≥ 0 we write it in terms of


the unit step function asThen,Hence, applying to the differential equation gives
usThen,Consider In the form of − 1{e− asF(s)}, a = π and F(s) = 1/[s(s2 + 9)]. Now,
f (t) = L− 1{F(s)} can be found with either a partial fraction expansion or with the
formulaThen with cos(3 t − 3π) = cos 3 t cos 3π + sin 3 t sin 3π = − cos 3 t, we
haveCombining these results yields the solutionNotice that without using the unit
step function, u, we can rewrite this function as a piecewise defined function aswhich
is graphed in Figure 8.5.

Figure 8.5. Graphs of three solutions of y&quot; + 9y = (t) _ (t _ π) that satisfy y(0) = 0.


Which one is the graph of the solution that satisfies y'(0) = 0?

> Read full chapter

Nonstandard Generalised Functions


R.F. Hoskins, in Delta Functions (Second Edition), 2011

Examples
(i) The unit step function appears in the Schwartz model as the functional
defined directly on byIn the axiomatic formulation it is simply defined as
the (distributional) derivative of the continuous function t+, which is such that
t+(t) = t for all t ≥ 0 and which vanishes for all negative values of t.
(ii) The delta distribution is usually defined as that functional which maps each
function into the number (0):It can be identified in the above axiomatic
formulation as the (distributional) second derivative of a continuous function (iii)
by means of the equation(11.27)
Similarly the pseudofunction t+− 3/2 can be defined as a distribution in the Silva
formulation as the second order (distributional) derivative of the continuous
function .

> Read full chapter

Laplace Transform Methods


Martha L. Abell, James P. Braselton, in Differential Equations with Mathematica
(Fourth Edition), 2016

8.4.1 Piecewise-Defined Functions: The Unit Step Function


An important function in modeling many physical situations is the unit step func-
tion, .

Definition 35 Unit Step FunctionThe unit step function, , where a is a number is


defined by (8.30)We can use the function UnitStep to define the unit step function:

so . Note: The functions UnitStep and HeavisideTheta have nearly identical function-
ality and can usually be used interchangeably.

Example 8.4.1

Graph (a) , (b) , and (c) .

Solution(a) Here, , so for t ≥ 2.(b)In this case, , so the “jump” occurs at t = 5. (c) .
These functions are graphed using Plot and UnitStep in Figure 8-5.Figure 8-5. Plots
of combinations of various step functionsThe unit step function is useful in defining
functions that are piecewise continuous. For example, we can define the function

as

Similarly, a function like

can be written as

The reason for writing piecewise continuous functions in terms of step functions is
that we encounter functions of this type in solving initial-value problems. Using our
methods in Chapters 4 and 5, we had to solve the problem over each piece of the
function. However, the method of Laplace transforms can be used to avoid these
complicated calculations.

Theorem 27

Suppose that exists for s > b ≥ 0. If a is a positive constant, then (8.31)

Example 8.4.2

Find .

SolutionIn this case, a = 3 and f(t) = t5. Thus, Equivalent results are obtained with
Mathematica. Here is the unsimplified the resultLaplaceTransform[(t−3)5UnitStep[-
t−3], t, s]Notice that FullSimplify simplifies the result to that you would obtain doing
the calculation by hand.LaplaceTransform[(t−3)5UnitStep[t−3], t, s]//FullSimplifyIn
most cases, we must calculate instead of . To solve this problem, we let g(t) = f(t −
a), so f(t) = g(t + a). Therefore,

(8.32)

Example 8.4.3

Calculate .

Solution

In this case, and a = π. Thus, The same result is obtained using LaplaceTransform.

LaplaceTransform[Sin[t]UnitStep[t−π], t, s]

Theorem 28

Suppose that exists for s > b ≥ 0. If a is a positive constant and y = f(t) is continuous
on , then (8.33)

Example 8.4.4

Find (a) and (b) .

Solution

(a) If we write the expression in the form e−asF(s), we see that a = 4 and F(s) = s−3.
Hence, and (b)In this case, a = π/2 and . Then, and For each of (a) and (b), the same
results are obtained usingInverseLaplaceTransform, although we must use Simplify
to simplify the result obtained for (b). Observe that several of are nearly identical in
their utility and scope HeavisideTheta function. Observe that HeavisideTheta

and UnitStep

are nearly identical in their utility and scope as mentioned previously.


Simplify[step1]

> Read full chapter

Time-invariant Linear Systems


R.F. Hoskins, in Delta Functions (Second Edition), 2011

4.2.1 Definition of impulse response


Given a T.I.L.S. represented by an operator T, suppose that the unit step function, u,
is an admissible input. That is to say, assume that in response to the unit step applied
as an input there exists a well-defined function  = T[u]. The function sketched in the
second graph of Fig. 4.2 is intended to illustrate just such a response to the unit step
function input shown in the first graph. This function is usually referred to as the
step-response of the system.

Figure 4.2. Unit step input and step response

Now consider the system response to the function

(4.6)

which represents a rectangular pulse of duration a and of amplitude k, centred about


the origin. Since the system concerned is both linear and time-invariant the response
to this signal applied as an input can be expressed as a linear combination of suitably
translated versions of the step-response. (See Fig. 4.3 for a sketch of the rectangular
pulse input, expressed as a combination of step functions, and of the corresponding
output.)
Figure 4.3.

That is to say we will have

(4.7)

Now suppose further that the function has a continuous classical derivative, h =  .
If the pulse width a is sufficiently small, then we can write

(4.8)

(For, if is continuously differentiable, then the first mean value theorem of the
differential calculus gives

where is some point such that –a/2 <   < + a/2. If a is small then by the continuity
of we have

Taking the particular case when k = 1/a we can say that the function h, defined here as
the derivative of the step response , represents to a first approximation the system
response to a narrow pulse of unit area located at the origin. This function h is called
the impulse response of the system.

> Read full chapter

Integration Theory
R.F. Hoskins, in Delta Functions (Second Edition), 2011
9.3.4
By contrast consider the case of the Riemann-Stieltjes integral in which the integra-
tor is a unit step function (t) = uc(t).We know that for any function f continuous on
some neighbourhood of the origin (and therefore certainly for any function belong-
ing to C0) this integral defines a possible model for the delta function described by
Dirac. That is, it defines a positive linear functional on C0 given by the mapping

and we recall that this result is quite independent of the number c. If we now carry
out the extension of the functional , as described above, the restriction to functions
continuous at the origin no longer applies. Consider, for example, the step function
u0(t). This is the characteristic function of the open interval (0, +∞) and has a jump
discontinuity at the origin. But it is easy to see that it is the limit of a monotone
increasing sequence of functions fm belonging to C0 such that for each m we have
(fm) = 0 (see, for example, Fig. 9.1). Hence we must have

Figure 9.1.

independently of c.

Similarly we can show that

The extended -functional a (or the distribution at a) yields a measure on called


the Dirac measure located at a.

For an arbitrary set A we have

and

> Read full chapter

Introduction to distributions
R.F. Hoskins, in Delta Functions (Second Edition), 2011
8.4.3 Distributions and ultradistribution
In Chapter 6 we were able to derive some ad hoc extensions of the classical Fourier
transform which applied to the unit step function, delta functions, end even to
infinite series of delta functions. It is clearly desirable that there should be a canonical
definition of the Fourier Transform, consistent with classical definitions, which is
applicable to all distributions - or, at least to some large and well-defined subset of
distributions. Now for the classical Fourier transform of sufficiently well-behaved
functions we have

where, for convenience, we now use the well-known alternative notation for the
transform of the function f. Further, for such well behaved functions f and g the
Parseval relation holds:

To extend the definition of Fourier transform to distributions Schwartz used a


generalised form of the Parseval relation to define the transform of a distribution μ,
as the functional satisfying

But there is a difficulty here: if then its well-defined classical Fourier Transform will
certainly be an infinitely differentiable function, but it will not be a function which
vanishes outside some finite interval. That is to say, will not be a member of the
space . It belongs to another, quite different function space . Dually, if does belong
to then its inverse Fourier transform will be a member of . It can be shown, in fact,
that and have no members in common: is the empty set.

It follows that if μ is an arbitrary Schwartz distribution then we can certainly define


its Fourier transform as a linear continuous functional on the space , but it will
not necessarily be defined as a linear continuous functional on . It may therefore
not be a Schwartz distribution at all. The linear continuous functionals on in fact
constitute a new space of generalised functions called ultradistributions. Some
ultradistributions do turn out to be distributions as well, and an extension of the
Fourier transform can be successfully defined for them - as the examples discussed
in Chapter 6 illustrate. These constitute an important subspace of and are called
tempered distributions. But there are ultradistributions which are not distributions
and there also exist distributions which are not ultradistributions. Further discussion
of this matter is outside the scope of this text, but a very clear account of the
distributional transform and of the definition and properties of ultradistributions
is to be found in the book by Zemanian referred to below.

> Read full chapter


Multidimensional Problems
Frank E. Harris, in Mathematics for Physical Science and Engineering, 2014

Unit Step Function


In some contexts, particularly in discussions of Laplace transforms, one encounters
another generalized function, the Heaviside function, also more descriptively called
the unit step function. The Heaviside function is, like the Dirac delta function, a
generalized function that has a clear meaning when it occurs within an integral of
the type shown here. Its defining property is that

(6.61)

applicable when the integral on the right-hand side converges. We see that the unit
step function has the basic property

(6.62)

and can be specified more formally as the limit of a sequence of functions as shown
in Fig. 6.16. We now observe that is related to the Dirac delta function according to

Figure 6.16. Sequence for unit step function.

(6.63)

Equations like Eq. (6.63) can be checked by verifying that the value of an appropriate
integral (for an arbitrary well-behaved function) is not changed when one side of this
equation is replaced by the other. The verification is the topic of Exercise 6.7.5.

> Read full chapter

INTRODUCTION AND OVERVIEW


BISWA NATH DATTA, in Numerical Methods for Linear Control Systems, 2004
1.2 SYSTEM RESPONSES (CHAPTER 5)
For the continuous-time system (1.0.1)–(1.0.2), the dynamical system responses x(t)
and y(t) for t t0 can be determined from the following formulas:

(1.2.1)

(1.2.2)

In order to study the behavior of a dynamical system, it is customary to determine


the responses of the system due to different inputs. Two most common inputs are
the unit step function and the unit impulse.

Thus, the unit step response of a system is the output that occurs when the input
is the unit step function (it is assumed that x(0) = 0). Similarly, the unit impulse
response is the output that occurs when the input is the unit impulse.

The impulse response matrix of the system (1.0.1) and (1.0.2) is defined by

where (t) is the Dirac delta function. The impulse response is the response of the
system to a Dirac input (t).

Thus, to obtain different responses, one needs to compute the matrix exponential
eAt = I + At + (A2t2/2) + … and the integrals involving this matrix. The computational
challenge here is how to determine eAt without explicitly computing the matrix powers.
Finding higher powers of a matrix is computationally intensive and is a source of
instability for the algorithm that requires such computations.

An obvious way to compute eA is to use some simple canonical forms of A such as


the JCF or a companion form of A. It is shown in Chapter 5 by simple examples how
such computations can lead to inaccurate results. Computations using truncated Taylor
series might also give erroneous result (see Example 5.3.3).

The method of choice here is either Padé approximation with scaling and squaring
(Algorithm 5.3.1) or the method based on reduction of A to real Schur form (Algo-
rithm 5.3.2).

A method (Algorithm 5.3.3) due to Van Loan (1978) for computing an integral
involving an matrix exponentials is also described in Section 5.3.5.

Frequency Response Computations


The frequency response plot for many different values of the frequency is impor-
tant in the study of various important properties of linear systems. The frequency
response curves indicate how the magnitude and angle of the sinusoidal steady-state
response change as the frequency of the input is changed. For this, the frequency-re-
sponse matrix G(j ) = C(j I – A)−1 B + D( 0) needs to be computed. Computing
G(j ) using the LU decomposition of A would require O(n3) operations per and is,
therefore, not practical when this computation has to be done for a large number
of values of . An efficient and practical method due to Laub (1981), based on
reduction of A to a Hessenberg matrix, is presented in Algorithm 5.5.1, and short
discussions on some other recent methods for efficient computations of the
frequency-response matrix is included in Section 5.5.2.

> Read full chapter

Definitions and General Properties


Vladimir Britanak, ... K.R. Rao, in Discrete Cosine and Sine Transforms, 2007

2.5 Some examples of the FST


(a) The unit rectangular pulse:x(t) = U(t) − U(t − 1), where U(t) is the Heaviside unit
step function.The FST is given by(2.39)
(b) The inverse quadratic function:By using the techniques of contour integration,
its FST is obtained as(2.40)Here and Ei are special functions called the ex-
ponential integral functions defined byand(2.41)The result here is somewhat
more complicated than that for the FCT.
(c) The exponential function:It is easily seen that the FST here is just the Laplace
transform of the sine function up to a scale factor,(2.42)
(d) The sinc function:Its FST is given by(2.43)It is interesting to note here that since
Fs−1 = Fs, the FST of the resulting logarithmic function is immediately seen to
be the sinc function up to a scale factor.
(e) The decaying sine function:Its FST may be recognized as the Laplace transform
of the function sin(at) sin( t) up to a scale factor. Thus,(2.44)
(f ) Bessel function of the first kind:The result of the FST is very similar to the FCT
result, giving,(2.45)

> Read full chapter

ScienceDirect is Elsevier’s leading information solution for researchers.


Copyright © 2018 Elsevier B.V. or its licensors or contributors. ScienceDirect ® is a registered trademark of Elsevier B.V. Terms and conditions apply.

You might also like