# Continuous Signals¶

This Jupyter notebook is part of a collection of notebooks in the bachelors module Signals and Systems, Communications Engineering, Universität Rostock. Please direct questions and suggestions to [email protected].

## Standard Signals¶

Certain signals play an important role in the theory and practical application of signal processing. They emerge from the theory of signals and systems, are used to characterize the properties of linear time-invariant (LTI) systems or frequently occur in practical applications. These standard signals are introduced and illustrated in the following. The treatise is limited to one-dimensional deterministic time- and amplitude-continuous signals.

### Complex Exponential Signal¶

The complex exponential signal over time $t$ is defined by the complex exponential function

$$x(t) = e^{s t}$$

where $s = \sigma + j \omega$ denotes the complex frequency with $\sigma, \omega \in \mathbb{R}$ and $j$ the imaginary unit $(j^2=-1)$. The signal is often used as a generalized representation of harmonic signals. Using Euler's formula above definition can be reformulated as

$$x(t) = e^{(\sigma + j \omega) t} = e^{\sigma t} \cos(\omega t) + j e^{\sigma t} \sin(\omega t)$$

The real/imaginary part of the exponential signal is given by a weighted cosine/sine with angular frequency $\omega = 2 \pi f$. For $t>0$, the time-dependent weight $e^{\sigma t}$ is

• exponentially decaying over time for $\sigma < 0$,
• constantly one for $\sigma = 0$,
• exponentially growing over time for $\sigma > 0$,

and vice-versa for $t<0$. The complex exponential signal is used to model harmonic signals with constant or exponentially decreasing/increasing amplitude.

Example

The following example illustrates the complex exponential signal and its parameters. The Python module SymPy is used for this purpose. It provides functionality for symbolic variables and functions, as well as their calculus. The required symbolic variables need to be defined explicitly before usage. In the example $t$, $\omega$ and $\sigma$ are defined as real-valued symbolic variables, followed by the definition of the exponential signal.

In [1]:
import sympy as sym
sym.init_printing()

t, sigma, omega = sym.symbols('t sigma omega', real=True)
s = sigma + 1j*omega
x = sym.exp(s*t)
x

Out[1]:
$\displaystyle e^{t \left(1.0 i \omega + \sigma\right)}$

Now specific values for the complex frequency $s = \sigma + j \omega$ are considered for illustration. For this purpose a new signal is defined by substituting both $\sigma$ and $\omega$ with specific values. The real and imaginary part of the signal is plotted for illustration.

In [2]:
y = x.subs({omega: 10, sigma: -.1})

sym.plot(sym.re(y), (t, 0, 2*sym.pi), ylabel=r'Re{$e^{st}$}')
sym.plot(sym.im(y), (t, 0, 2*sym.pi), ylabel=r'Im{$e^{st}$}');


Exercise

• Try other values for omega and sigma to create signals with increasing/constant/decreasing amplitudes and different angular frequencies.

### Dirac Impulse¶

The Dirac impulse is one of the most important signals in the theory of signals and systems. It is used for the characterization of LTI systems and the modeling of impulse-like signals. The Dirac impulse is defined by way of the Dirac delta function which is not a function in the conventional sense. It is a generalized function or distribution. The Dirac impulse is denoted as $\delta(t)$. The Dirac delta function is defined by its effect on other functions. A rigorous treatment is beyond the scope of this course material. Please refer to the literature for a detailed discussion of the mathematical foundations of the Dirac delta distribution. Fortunately it is suitable to consider only certain properties for its application in signal processing. The most relevant ones are

1. Sifting property $$\int_{-\infty}^{\infty} \delta(t) \cdot x(t) = x(0)$$ where $x(t)$ needs to be differentiable at $t=0$. The sifting property implies $\int_{-\infty}^{\infty} \delta(t) = 1$.

2. Multiplication $$x(t) \cdot \delta(t) = x(0) \cdot \delta(t)$$ where $x(t)$ needs to be differentiable at $t=0$.

3. Linearity $$a \cdot \delta(t) + b \cdot \delta(t) = (a+b) \cdot \delta(t)$$

4. Scaling $$\delta(a t) = \frac{1}{|a|} \delta(t)$$ where $a \in \mathbb{R} \setminus 0$. This implies that the Dirac impulse is a function with even symmetry.

5. Derivation $$\int_{-\infty}^{\infty} \frac{d \delta(t)}{dt} \cdot x(t) \; dt = - \frac{d x(t)}{dt} \bigg\vert_{t = 0}$$

6. Convolution

Generalization of the sifting property yields $$\int_{-\infty}^{\infty} \delta(\tau) \cdot x(t - \tau) \, d\tau = x(t)$$

This operation is known as convolution and will be introduced later in more detail. It may be concluded already here that the Dirac delta function constitutes the neutral element of the convolution.

It is important to note that the product $\delta(t) \cdot \delta(t)$ of two Dirac impulses is not defined.

Example

This example illustrates some of the basic properties of the Dirac impulse. Let's first define a Dirac impulse by way of the Dirac delta function

In [3]:
delta = sym.DiracDelta(t)
delta

Out[3]:
$\displaystyle \delta\left(t\right)$

Now let's check the sifting property by defining an arbitrary signal (function) $f(t)$ and integrating over its product with the Delta impulse

In [4]:
f = sym.Function('f')(t)
sym.integrate(delta*f, (t, -sym.oo, sym.oo))

Out[4]:
$\displaystyle f{\left(0 \right)}$

Exercise

• Derive the sifting property for a shifted Dirac impulse $\delta(t-\tau)$ and check your results by modifying above example.

### Heaviside Signal¶

The Heaviside signal is defined by the Heaviside step function

$$\epsilon(t) = \begin{cases} 0 & t<0 \\ \frac{1}{2} & t=0 \\ 1 & t > 0 \end{cases}$$

Note that alternative definitions exist, which differ with respect to the value of $\epsilon(t)$ at $t=0$. The Heaviside signal may be used to represent a signal that switches on at a specified time and stays switched on indefinitely. The Heaviside signal can be related to the Dirac impulse by

$$\epsilon(t) = \int_{-\infty}^{t} \delta(\tau) \; d\tau$$

Example

In the following, a Heaviside signal $\epsilon(t)$ is defined and plotted. Note that Sympy denotes the Heaviside function by $\theta(t)$.

In [5]:
step = sym.Heaviside(t)
step

Out[5]:
$\displaystyle \theta\left(t\right)$
In [6]:
sym.plot(step, (t, -2, 2), ylim=[-0.2, 1.2], ylabel=r'$\epsilon(t)$');


Let's construct a harmonic signal $\cos(\omega t)$ with $\omega=2$ which is switched on at $t=0$. Considering the definition of the Heaviside function, the desired signal is given as

$$x(t) = \cos(\omega t) \cdot \epsilon(t)$$
In [7]:
x = sym.cos(omega*t) * sym.Heaviside(t)
sym.plot(x.subs(omega, 2), (t, -2, 10), ylim=[-1.2, 1.2], ylabel=r'$x(t)$');


### Rectangular Signal¶

The rectangular signal is defined by the rectangular function

$$\text{rect}(t) = \begin{cases} 1 & |t| < \frac{1}{2} \\ \frac{1}{2} & |t| = \frac{1}{2} \\ 0 & |t| > \frac{1}{2} \end{cases}$$

Its time limits and amplitude are chosen such that the area under the function is $1$.

Note that alternative definitions exist, which differ with respect to the value of $\text{rect}(t)$ at $t = \pm \frac{1}{2}$. The rectangular signal is used to represent a signal which has finite duration, respectively is switched on for a limited period of time. The rectangular signal can be related to the Heaviside signal by

$$\text{rect}(t) = \epsilon \left(t + \frac{1}{2} \right) - \epsilon \left(t - \frac{1}{2} \right)$$

Example

The Heaviside function is used to define a rectangular function in Sympy. This function is then used as rectangular signal.

In [8]:
class rect(sym.Function):

@classmethod
def eval(cls, arg):
return sym.Heaviside(arg + sym.S.Half) - sym.Heaviside(arg - sym.S.Half)

In [9]:
sym.plot(rect(t), (t, -1, 1), ylim=[-0.2, 1.2], ylabel=r'rect$(t)$');


Exercise

• Use $\text{rect}(t)$ to construct a harmonic signal $\cos(\omega t)$ with $\omega=2$ which is switched on at $t=-\frac{1}{2}$ and switched off at $t=+\frac{1}{2}$.

### Sign Signal¶

The sign signal is defined by the sign/signum function which evaluates the sign of its argument

$$\text{sgn}(t) = \begin{cases} 1 & t>0 \\ 0 & t=0 \\ -1 & t < 0 \end{cases}$$

The sign signal is useful to represent the absolute value of a real-valued signal $x(t) \in \mathbb{R}$ by a multiplication

$$|x(t)| = x(t) \cdot \text{sgn}(x(t))$$

It is related to the Heaviside signal by

$$\text{sgn}(t) = 2 \cdot \epsilon(t) - 1$$

when following above definition with $\epsilon(0)=\frac{1}{2}$.

Example

The following example illustrates the sign signal $\text{sgn}(t)$. Note that the sign function is represented as $\text{sign}(t)$ in Sympy.

In [10]:
sgn = sym.sign(t)
sgn

Out[10]:
$\displaystyle \operatorname{sign}{\left(t \right)}$
In [11]:
sym.plot(sgn, (t, -2, 2), ylim=[-1.2, 1.2], ylabel=r'sgn$(t)$');


Exercise

• Check the values of $\text{sgn}(t)$ for $t \to 0^-$, $t = 0$ and $t \to 0^+$ as implemented in SymPy. Do they conform to above definition?