Lecture 2

Introduction to Systems
 Introduction
▪ A continuous-time (CT) / discrete-time (DT) system
transforms CT/DT input signals into CT/DT output signals.
CT
𝑥(𝑡) 𝑦(𝑡)
System

DT
𝑥𝑛 𝑦𝑛
System

▪ Systems can be classified as

1) Linear vs. Nonlinear Systems
2) Time-Invariant vs. Time-Variant Systems
3) Causal Systems
4) Stable Systems
DT Linear Systems
▪ Definition: A DT system is said to be linear iff it satisfies.
1. the response to 𝑥1 𝑛 + 𝑥2 𝑛 is 𝑦1 𝑛 + 𝑦2 𝑛
2. the response to 𝛼𝑥1 𝑛 is 𝛼𝑦1 𝑛 ,  is a constant

▪ If 𝑥1 𝑛 𝓛 𝑦1 𝑛
A DT linear system
is a linear
𝑥2 𝑛 𝓛 𝑦2 𝑛 transformation

▪ Then,
𝑥1 𝑛 + 𝑥2 𝑛 𝑦1 𝑛 + 𝑦2 𝑛
𝓛
Linear transformations
between finite dimensional
𝛼𝑥1 𝑛 𝓛 𝛼𝑦1 𝑛 spaces can be represented
by matrices.
CT Linear Systems
▪ Definition: A CT system is said to be linear iff it satisfies.
1. the response to 𝑥1 (𝑡) + 𝑥2 (𝑡) is 𝑦1 (𝑡) + 𝑦2 (𝑡)
2. the response to 𝛼𝑥1 (𝑡) is 𝛼𝑦1 (𝑡),  is a constant

▪ If 𝑥1 (𝑡) 𝓛 𝑦1 (𝑡)
A CT linear system
is a linear
𝑥2 (𝑡) 𝓛 𝑦2 (𝑡) transformation
between infinite
▪ Then, dimensional spaces.
𝑥1 (𝑡) + 𝑥2 (𝑡) 𝑦1 (𝑡) + 𝑦2 (𝑡)
𝓛

𝛼𝑥1 (𝑡) 𝓛 𝛼𝑦1 (𝑡)

Principle of Superposition
▪ The principle of superposition: If the response to 𝑥(𝑡) is 𝑦(𝑡), then
the response of the system to 𝛼1 𝑥1 (𝑡) + 𝛼2 𝑥2 (𝑡) is
𝛼1 𝑦1 (𝑡) + 𝛼2 𝑦2 (𝑡)

▪ If 𝑥1 (𝑡) 𝓛 𝑦1 (𝑡)

𝑥2 (𝑡) 𝓛 𝑦2 (𝑡)

▪ Then,
𝛼1 𝑥1 (𝑡) + 𝛼2 𝑥2 (𝑡) 𝛼1 𝑦1 (𝑡) + 𝛼2 𝑦2 (𝑡)
𝓛
 Examples
▪ Example 1:
▪ Is the system defined by 𝑦(𝑡) = 𝑘𝑥(𝑡) + 𝛽, where k and 𝛽 are constants,
linear?
▪ Given 𝑦1 (𝑡) = 𝑘𝑥1 (𝑡) + 𝛽 and 𝑦2 (𝑡) = 𝑘𝑥2 (𝑡) + 𝛽,
the response to 𝑥1 (𝑡) + 𝑥2 (𝑡) is
𝑦3 (𝑡) = 𝑘 𝑥1 (𝑡) + 𝑥2 (𝑡) + 𝛽 ≠ 𝑦1 (𝑡) + 𝑦2 (𝑡) = 𝑘𝑥1 (𝑡) + 𝑘𝑥2 (𝑡) + 2𝛽
∴ the system is nonlinear if 𝛽 ≠ 0.

▪ Remark: Linear systems must satisfy the property

𝑥(𝑡) ≡ 0 ∀t ⇒ y(𝑡) ≡ 0 ∀𝑡
▪ In the above system, 𝑥(𝑡) ≡ 0 ⇒ 𝑦(𝑡) ≡ 𝛽
Therefore, it is not linear.
Examples (cont’d)
▪ Example 2:
▪ Is the system given by 𝑦(𝑡) = 𝑘𝑥(𝑡), where k is a constant, linear?
▪ Given 𝑦1 (𝑡) = 𝑘𝑥1 (𝑡) and 𝑦2 (𝑡) = 𝑘𝑥2 (𝑡),
the response to 𝑥1 (𝑡) + 𝑥2 (𝑡) is
▪ 𝑦3 (𝑡) = 𝑘 𝑥1 (𝑡) + 𝑥2 (𝑡) = 𝑘𝑥1 (𝑡) + 𝑘𝑥2 (𝑡) = 𝑦1 (𝑡) + 𝑦2 (𝑡), and
▪ 𝛼𝑦(𝑡) = 𝛼𝑘𝑥(𝑡)
▪  the system is linear.

▪ Example 3:
▪ Is the system defined by 𝑦(𝑡) = 𝑒𝑥𝑝 𝑥(𝑡) linear?
▪ Given 𝑦1 (𝑡) = 𝑒𝑥𝑝 𝑥1 (𝑡) and 𝑦2 (𝑡) = 𝑒𝑥𝑝 𝑥2 (𝑡) ,
the response to 𝑥1 (𝑡) + 𝑥2 (𝑡) is
𝑦3 (𝑡) = 𝑒𝑥𝑝 𝑥1 (𝑡) + 𝑥2 (𝑡) = 𝑒𝑥𝑝 𝑥1 (𝑡) 𝑒𝑥𝑝 𝑥2 (𝑡) ≠ 𝑦1 (𝑡) + 𝑦2 (𝑡)
▪ ∴ the system is nonlinear
Time-Invariant Systems
▪ A system is called time-invariant if a time shift in the input signal causes only
an identical time shift in the output signal; i.e., the system’s response does not
change depending on the time at which the input is applied.

𝑥(𝑡) 𝓣𝓘 𝑦(𝑡)

𝑥(𝑡 − 𝑡0 ) 𝓣𝓘 𝑦(𝑡 − 𝑡0 )

▪ The procedure for testing whether a system is time-invariant is as follows:

1) Let 𝑦1(𝑡) be the output corresponding to 𝑥1 (𝑡).
2) Consider a second input, 𝑥2(𝑡) obtained by shifting 𝑥1 (𝑡); i.e.,
𝑥2 (𝑡) = 𝑥1 (𝑡 − 𝑡0 )
and find the output 𝑦2 (𝑡) corresponding to 𝑥2 (𝑡).
3) From step 1, shift 𝑦1(𝑡) to obtain 𝑦1 (𝑡 − 𝑡0 )
4) If 𝑦2(𝑡) = 𝑦1(𝑡 − 𝑡0), the system is time-invariant; otherwise it is time-varying.
Example
▪ Example 1:
Is the system 𝑦(𝑡) = 𝑐𝑜𝑠 𝑥(𝑡) time-invariant?
▪ Applying the procedure
▪ i) 𝑦1 (𝑡) = 𝑐𝑜𝑠 𝑥1 (𝑡) ;
▪ ii) Shift the input: 𝑥2 (𝑡) = 𝑥1 (𝑡 − 𝑡0 )
▪ iii) Compute output for 𝑥2 𝑡 ∶ 𝑦2 𝑡 = 𝑐𝑜𝑠 𝑥2 𝑡 = 𝑐𝑜𝑠 𝑥1 (𝑡 − 𝑡0 )
▪ iv) Is it equal to shifted output: 𝑦2 (𝑡) = 𝑦1 (𝑡 − 𝑡0 ) = 𝑐𝑜𝑠 𝑥1 (𝑡 − 𝑡0 )
▪ ∴ it is time-invariant.
▪ Example 2:
▪ Is the system 𝑦(𝑡) = 𝑡 𝑥(𝑡) time-invariant?
▪ Applying the procedure
▪ i) 𝑦1 (𝑡) = 𝑡𝑥1 (𝑡)
▪ ii) 𝑥2 (𝑡) = 𝑥1 (𝑡 − 𝑡0 )
▪ iii) 𝑦2 (𝑡) = 𝑡𝑥2 (𝑡) = 𝑡𝑥1 (𝑡 − 𝑡0 )
▪ iv) 𝑦1 (𝑡 − 𝑡0 ) = (𝑡 − 𝑡0 )𝑥1 (𝑡 − 𝑡0 ) ≠ 𝑦2 (𝑡)
▪ ∴ the system is time-variant.
Try Yourself
▪ Are these systems linear? Time invariant?

▪ 𝑦(𝑡) = 𝑥(𝑡)

▪ 𝑦 𝑡 = 𝑥 𝑡 cos 2𝜋𝑓0 𝑡

▪ 𝑦(𝑡) = 𝑥(𝑎𝑡)
Linear, Time-Invariant Systems
▪ They are linear

▪ They are time invariant

𝑥(𝑡) 𝓛𝓣𝓘 𝑦(𝑡)

 Stable Systems
▪ Definition: A signal 𝑥(𝑡) is said to be bounded if its magnitude
does not grow without a bound, i.e.,
𝑥(𝑡) < 𝐵 < ∞ for all t.

𝑥(𝑡) = 𝑢(𝑡) 𝑥(𝑡) = 𝑡

𝑡 𝑡
bounded unbounded

▪ Definition: A system is BIBO stable if for any bounded input

x(t), the response y(t) is also bounded, i.e.,
𝑥(𝑡) < 𝐵1 < ∞ ⇒ 𝑦(𝑡) < 𝐵2 < ∞
Example
▪ Example 1:
𝑥 𝑡 = 𝑒𝑡 − ∞ ≤ 𝑡 ≤ ∞ is an unbounded signal.

▪ Example 2:
▪ Is the system 𝑦(𝑡) = 𝑒 𝑥(𝑡) stable?

▪ We start by noting that 𝑦(𝑡) = 𝑒 𝑥(𝑡) ≤ 𝑒 𝑥(𝑡) ; e.g., |e-2| = e-2 < e2
▪ Given a bounded input 𝑥(𝑡) < 𝐵1 < ∞, the output
𝑦(𝑡) = 𝑒 𝑥(𝑡) ≤ 𝑒 𝑥(𝑡) < 𝑒 𝐵1 = 𝐵2 < ∞ is bounded.
∴ the system is stable.
Example (cont’d)
▪ Example 3:
𝑡
▪ Is the integrator system defined by 𝑦 𝑡 = ‫׬‬−∞ 𝑥 𝜏 𝑑𝜏 stable?
▪ Let 𝑥(𝑡) < 𝐵 < ∞ be a bounded input.
𝑡 𝑡 𝑡
𝑦(𝑡) = න 𝑥(𝜏)𝑑𝜏 ≤ න 𝑥 𝜏 𝑑𝜏 < න 𝐵𝑑𝜏
−∞ −∞ −∞
as 𝑡 → ∞, 𝑦(𝑡) → ∞.
▪ Thus, 𝑦(𝑡) is not bounded; hence, the system is unstable.

▪ Alternative Solution: There exists a bounded input, for which

the output is unbounded. Let 𝑥(𝑡) = 𝑢(𝑡)
𝑥(𝑡) = 𝑢(𝑡) 𝑦(𝑡) = 𝑡

𝑡 𝑡
Systems with Memory
▪ A system is memoryless if the output depends only on
the present input.
▪ Example 1: Resistor

𝑦(𝑡) = 𝑅𝑖(𝑡) is memoryless

▪ A system with memory has an output signal that

depends on inputs in the past or future.
▪ Example 2: Energy storage elements such as capacitors and inductors

1 𝑡
𝑦 𝑡 = ‫׬‬−∞ 𝑥 𝜏 𝑑𝜏 has memory.
𝐶

▪ Other Examples: Springs or moving masses in mechanical systems

Causal Systems
▪ A system is causal if its output at any time t0 depends only
on values of the input for 𝑡 ≤ 𝑡0

▪ Example 1:
𝑦(𝑡) = 𝑥(𝑡 − 2) is causal.

▪ Example 2:
𝑦(𝑡) = 𝑥(𝑡 + 2) 2
is noncausal, because its output depends on inputs applied at a later
time.
CT Systems described by LCCDE
▪ The input-output relationship for many CT systems are described by a
linear constant coefficient differential equation (LCCDE):

𝑎𝑛 𝑦 (𝑛) 𝑡 + ⋯ + 𝑎1 𝑦′ (𝑡) + 𝑎0 𝑦(𝑡) = 𝑏𝑚 𝑥 (𝑚) (𝑡) + ⋯ + 𝑏1 𝑥′(𝑡) + 𝑏0 𝑥(𝑡)

with given initial condition

𝑦 (𝑛−1) (0),…, y’(0); y(0)
▪ n is called the order of the system
▪ 𝑏0 ,⋯ , 𝑏𝑚 ; 𝑎0 ,⋯ , 𝑎𝑛 are the coefficients of the system

▪ Systems defined by LCCDE are linear iff all initial conditions are zero.

▪ An LCCDE describes how 𝑥(𝑡), 𝑦(𝑡), and their derivatives are interrelated.
▪ You need to solve the LCCDE to get an explicit solution for 𝑦(𝑡) in terms of 𝑥(𝑡)
Example: Series RLC circuit

▪ Kirchoff's voltage law

𝑥(𝑡) – 𝐿𝑖’(𝑡) − 𝑅𝑖(𝑡) − 𝑦(𝑡) = 0
▪ Using 𝑖(𝑡) = 𝐶𝑦’(𝑡),
𝑥(𝑡) – 𝐿𝐶𝑦’’(𝑡) – 𝑅𝐶𝑦’(𝑡) − 𝑦(𝑡) = 0
or
𝑅 1 1
𝑦’′(𝑡) + 𝑦’(𝑡) + 𝑦(𝑡) = 𝑥(𝑡)
𝐿 𝐿𝐶 𝐿𝐶
is a second-order LCCDE.
▪ This is a linear system assuming zero initial conditions.
DT Systems: Difference equations
▪ The system equations are difference equations
𝑎0 𝑦 𝑛 + 𝑎1 𝑦 𝑛 − 1 + ⋯ + 𝑎𝑁 𝑦 𝑛 − 𝑁
= 𝑏0 𝑥[𝑛] + 𝑏1 𝑥[𝑛 − 1] + ⋯ + 𝑏𝑀 𝑥[𝑛 − 𝑀]
where 𝑥[𝑛] is the input, and 𝑦[𝑛] is the output.
▪ Multiplier

𝑥[𝑛] 𝛼 𝛼 𝑥[𝑛]

▪ Delay
𝑥[𝑛] Delay 𝑥[𝑛 − 1]
Example: First-order DT System

▪ The input into the delay is

𝑒 𝑛 = 𝑥 𝑛 − 𝑎𝑦[𝑛]
▪ The output is 𝑦 𝑛 = 𝑒[𝑛 − 1], so
𝑦 𝑛 = 𝑥 𝑛 − 1 − 𝑎𝑦[𝑛 − 1]