INS2094 SIGNALS AND SYSTEMS

Lecture 05: Introduction to Systems

(Topic 2- Part 2)

❖Lecturer: Dr. Pham Ngoc Thanh

❖Faculty of Applied Sciences, International School, Vietnam National
University VNUIS

1
Differential equations
❖ Differential equation Modelling is the most popular type of mathematical modelling of
dynamic systems in various fields.
❖ For physical systems, their differential equation models are based on the equations of
physical laws governing the operation of those systems.
❖ Continuous-time linear time-invariant systems are represented by constant-coefficient
linear differential equations

General form of differential equations

General form of constant-coefficient linear differential equations representing continuous-
time LTI systems
𝑑𝑦 𝑡 𝑑𝑁 𝑦 𝑡 𝑑𝑥 𝑡 𝑑𝑀 𝑥 𝑡
𝑎0 𝑦 𝑡 + 𝑎1 + ⋯ + 𝑎𝑁 𝑁
= 𝑏0 𝑥 𝑡 + 𝑏1 + ⋯ + 𝑏𝑀
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡𝑀
𝑑𝑘 𝑦 𝑡
with N initial conditions: 𝑦 0 , ȁ for 𝑘 = 1, · · · , 𝑁 − 1 and input 𝑥(𝑡) = 0 for
𝑑𝑡 𝑘 𝑡=0
𝑡 < 0, 2
RC CIRCUITS
Example 2.14. Consider a RC circuit in series with
a constant voltage source 𝑣𝑖 𝑡 input and R = 1Ω,
C = 1F, with initial voltage 𝑣𝑐 0 = 𝑉 across
capacitor.
Solution 2.14.
𝑑𝑞 𝑑
𝑣𝑅 𝑡 = 𝑅 𝑖(𝑡), 𝑖 𝑡 = =𝐶 𝑣 (𝑡) ,
𝑑𝑡 𝑑𝑡 𝑐
𝑣𝑖 𝑡 = 𝑣𝑐 𝑡 + 𝑣𝑅 𝑡
𝑑
→ 𝑣𝑖 𝑡 = 𝑣𝑐 𝑡 + 𝑅𝐶 𝑣 (𝑡)
𝑑𝑡 𝑐
Denote 𝑥 𝑡 = 𝑣𝑖 𝑡 , 𝑦 𝑡 = 𝑣𝑐 𝑡 ,
By solving the above differential
𝑅𝐶 𝑑𝑦 𝑡
y(𝑡) + =𝑥 𝑡 equation, the output signal y(t) is
𝑑𝑡
For 𝑅 = 1 Ω and 𝐶 = 1𝐹 determined when the input signal x(t)
𝑑 has been known. 3
→ 𝑣𝑖 𝑡 = 𝑣𝑐 𝑡 + 𝑣 (𝑡)
𝑑𝑡 𝑐
Solving Differential equations
The general solution of a constant- The homogeneous equation
coefficient linear differential equation has 𝑁 𝑑 𝑖 (𝑦 𝑡 )
σ𝑖=0 𝑎𝑖 =0
the following form: 𝑑𝑡 𝑖
𝑦(𝑡) = 𝑦0 (𝑡) + 𝑦𝑠 (𝑡) has a solution 𝑦0 (𝑡) in the form 𝑒 𝑠𝑡 in which s
𝒚𝟎 (𝒕): the initial response, which is the is a complex variable (𝑠 = 𝑠𝑅𝑒 + 𝑗𝑠𝐼𝑚 ).
solution of the following homogeneous Replacing 𝑦0 𝑡 = 𝑒 𝑠𝑡 in the equation by it
𝑑 𝑖 (𝑦 𝑡 )
yields:
𝑁
σ𝑖=0 𝑎𝑖
equation: =0 σ𝑁 𝑖 𝑠𝑡
𝑑𝑡 𝑖 𝑖=0 𝑎𝑖 𝑠 𝑒 = 0
𝒚𝒔 (𝒕): the zero-state response, which is the s is the solution of the following linear algebra
particular solution of the equation for an equation with the order of N:
input signal x(t). σ𝑁𝑖=0 𝑎 𝑖 𝑠 𝑖 =0
Now we find 𝒚𝟎 𝒕 The equation is called the characteristic
𝑦0 (𝑡) is the response of the system to the equation of the system. Denote 𝑠𝑘 , k=1,…N,
system conditions at the initial point (t = be the roots of the equation
0), without input signal x(t).
4
Solving Differential equations
The general solution of the homogeneous equation representing the system.
𝑵 𝒅𝒊 (𝒚 𝒕 ) ✓To determine 𝑦𝑠 (𝑡) it is usually assumed that
equation σ𝒊=𝟎 𝒂𝒊 = 𝟎 has the 𝑦𝑠 (𝑡) has a form similar to the input signal
𝒅𝒕𝒊
following form if all {𝑠𝑘 } are distinct roots: x(t) with some unknown coefficients which
𝑦0 (𝑡) = σ𝑁 𝑘=1 𝑐 𝑘 𝑒 𝑠𝑘 𝑡
will be identified later from the equation.
In case the equation σ𝑁 𝑎
𝑖=0 𝑖 𝑠 𝑖
= 0 has 𝑦𝑠 (𝑡) must be independent of all terms of the
repeated roots, the general form of the homogeneous solution 𝑦0 𝑡 .
homogeneous solution will be: • if 𝑥(𝑡) = 𝑒 𝛼𝑡 we may have to consider the
𝑁 𝑠 𝑡 𝑝𝑘 −1 𝑖
σ
𝑦0 𝑡 = 𝑘=1 𝑐𝑘 𝑒 𝑘 σ 𝑖=0 𝑡
following cases:
in which each root 𝑠𝑘 is repeated 𝑝𝑘 times. ✓If 𝑒 𝛼𝑡 is not a component of 𝑦0(𝑡), then
Now we find 𝐲𝐬 (𝐭) suggest 𝑦𝑠 𝑡 = 𝑐𝑒 𝛼𝑡 .
𝒚𝒔 𝒕 : system response to the input signal ✓If 𝑒 𝛼𝑡 is a component of 𝑦0(𝑡), → 𝑦𝑠 𝑡 =
x(t) when all initial conditions are zeros. 𝑡𝑐𝑒 𝛼𝑡 .
𝒚𝒔 𝒕 is also called the particular solution of ✓If 𝑒 𝛼𝑡 , 𝑡𝑒 𝛼𝑡 , … , 𝑡 𝑝−1𝑒 𝛼𝑡 are components of
the constant-coefficient linear differential 𝑦0 (𝑡), → 𝑦𝑠 (𝑡) = 𝑐𝑡 𝑝 𝑒 𝛼𝑡
5
Solving Differential equations
Example 2.15. We reconsider the Find the solution of the homogeneous
example 2.14 RC circuit. Determine the equation
output signal 𝑦 𝑡 = 𝑣𝑐 (𝑡) in respect to 𝑦 𝑡 + 𝑅𝐶
𝑑𝑦(𝑡)
𝑒 −2𝑡 =
𝑑𝑡
the input 𝑥 𝑡 = 𝑣𝑖 (𝑡) = 𝑒 −2𝑡 with initial
We suggest 𝑦𝑠 𝑡 = 𝐷𝑒 −2𝑡
voltage 𝑣𝑐 0 = 1V across capacitor
→ 𝐷𝑒 −2𝑡 + 𝑅𝐶𝐷 −2 𝑒 −2𝑡 = 𝑒 −2𝑡
Find the general solution 𝑦0 (𝑡) of → 𝐷 1 − 2𝑅𝐶 = 1
1 1
𝑦 𝑡 + 𝑅𝐶
𝑑𝑦(𝑡)
= 0. →𝐷= → 𝑦𝑠 𝑡 = 𝑒 −2𝑡
𝑑𝑡 1−2𝑅𝐶 1−2𝑅𝐶
characteristic equation of the system is Hence → 𝒚 𝒕 = 𝒚𝟎 𝒕 + 𝒚𝒔 𝒕
1 + 𝑅𝐶𝑠 = 0 1
𝒚 𝒕 = 𝑪𝟎 𝒆−𝒕 + 𝑒 −2𝑡
1−2𝑅𝐶
Due to 1+𝑅𝐶s=0 has only one root 𝒔 =
For 𝑡 = 0, 𝑦 𝑡 = 𝑣𝑐 0 = 1 → 𝟏 = 𝑪𝟎 − 𝟏
𝟏
− , general solution 𝑦0 (𝑡) is Note 𝑹𝑪 = 𝟏.
𝑹𝑪
𝟏
𝑦0 𝑡 = 𝐶0 𝑒 −𝑹𝑪𝑡
= 𝐶0 𝑒 −𝑡 −𝒕
1
𝒚 𝒕 = 𝟐𝒆 + 𝑒 −2𝑡
1 − 2𝑅𝐶 6
Impulse response of a system
The impulse response of a continuous-time LTI Properties
system, h(t), is the output of the system ✓Commutativity: f(t) ∗ g(t) = g(t) ∗ f(t)
corresponding to an impulse δ(t) and initial ✓Associativity:
conditions equal to zero. [f(t) ∗ g(t)] ∗ h(t) = f(t) ∗ [g(t) ∗ h(t)]
✓Distributivity: [f(t) + g(t)] ∗ h(t) = f(t) ∗
The response of a LTI system represented by its
h(t) + g(t) ∗ h(t)
impulse response ℎ(𝑡), to any signal 𝑥(𝑡) is the
✓Time shift: if 𝑥 𝑡 = 𝑓 𝑡 ∗ 𝑔 𝑡 ,
convolution integral:
𝑡ℎ𝑒𝑛 𝑥 𝑡 − 𝑡0 = 𝑓 𝑡 − 𝑡0 ∗ 𝑔 𝑡
∞
𝑦 𝑡 = න 𝑥 𝜏 ℎ 𝑡 − 𝜏 𝑑𝜏 = [𝑥 ∗ ℎ](𝑡) = 𝑓(𝑡) ∗ 𝑔(𝑡 − 𝑡0 )
−∞ ✓Convolution of a signal with the unit
impulse: f(t) ∗ δ(t) = f(t)
The convolution of two signals f(t) and g(t),
denoted f(t) ∗ g(t), is defined as:
✓Causality: if f(t) and g(t) are causal
∞ signals then f(t) ∗ g(t) is also causal.
𝒇 𝒕 ∗ 𝒈 𝒕 = ‫׬‬−∞ 𝒇 𝝉 𝒈 𝒕 − 𝝉 𝒅𝝉

7
Impulse response: Example
Example 2.16: Let the impulse response
for an LTI system be ℎ 𝑡 = 𝑒 −𝑎𝑡 𝑢 𝑡
where a is a positive constant. If the input
to the system is 𝑥(𝑡) = 𝑢(𝑡), (unit step
signal). Find the output y(t) using the
convolution integral.
Solution (cont)
∞
𝑦 𝑡 = ‫׬‬−∞ 𝑢 𝜏 𝑒 −𝑎 𝑡−𝜏 𝑢(𝑡 − 𝜏)𝑑𝜏
∞ −𝑎 𝑡−𝜏
𝑦 𝑡 = ‫׬‬0 𝑒 𝑑𝜏
∞
𝑦 𝑡 = ‫׬‬0 𝑒 −𝑎𝑡 𝑒 𝑎𝜏 𝑑𝜏 . Note that this is a Solution 2.16: the output y(t) is the result of
integral in respect to 𝜏. Hence 𝑒 −𝑎𝑡 can be convolution of h(t) and x(t),
∞
taken out, then 𝒚 𝒕 = 𝒙 𝒕 ∗ 𝒉 𝒕 = ‫׬‬−∞ 𝒙 𝝉 𝒉 𝒕 − 𝝉 𝒅𝝉
∞ 𝑎𝜏 1−𝑒 −𝑎𝑡
𝑦 𝑡 = 𝑒 −𝑎𝑡 ‫׬‬0 𝑒 𝑑𝜏 = for 𝑡 > 0. Note that 𝑥 𝜏 = 𝑢 𝜏 ,
𝑎
ℎ 𝑡 − 𝜏 = 𝑒 −𝑎 𝑡−𝜏 𝑢(𝑡 − 𝜏).
8
System’s Property with impulse response
➢A LTI system is determined when Example 2.17: a microphone close to a set of speakers
its impulse response is determined. that are putting out an amplified acoustic signal.
❖ The microphone picks up the input signal x(t) and
➢Memoryless: impulse response has the amplified and delayed signal βy(t −τ), where β ≥
only one non-zero value at t = 0. 1 is the gain of the amplifier.
➢Causal systems: impulse response ❖ The speakers provide the feedback.
is a causal signal.
➢ Find the equation that connects the input x(t) and
➢Stable systems: if and only if the the output y(t)
following condition is satisfied ➢ Determine if the system is BIBO stable or not; use
∞
x(t) = u(t), β = 2 and τ = 1.
න ȁ𝒉 𝒕 ȁ𝒅𝒕 < ∞
−∞ ➢ Solution 2.17: Find the equation that connects the
Indeed
∞
input x(t) and the output y(t)
𝑦 𝑡 = ‫׬‬−∞ 𝑥 𝑡 − 𝜏 ℎ 𝜏 𝑑𝜏 𝑦 𝑡 = 𝑥 𝑡 + 𝛽𝑦(𝑡 − 𝜏)
∞ We need to find out 𝑦(𝑡 − 𝜏)
≤ 𝑀 ‫׬‬−∞ ℎ 𝜏 𝑑𝜏 ≤ 𝑀𝐾 < ∞
9
Microphone - Speaker
▪ 𝑦 𝑡 = 𝑥 𝑡 + 𝛽𝑦 𝑡 − 𝜏 . 𝑦 𝑡 = 𝑢 𝑡 + 𝛽𝑢 𝑡 − 𝜏 + 𝛽𝑢 𝑡 − 𝜏
▪ 𝑦 𝑡 − 𝜏 = 𝑥 𝑡 − 𝜏 + 𝛽𝑦 𝑡 − 2𝜏 . +𝛽2 𝑢 𝑡 − 2𝜏 + ⋯
▪ 𝑦 𝑡 − 2𝜏 = 𝑥 𝑡 − 2𝜏 + 𝛽𝑦 𝑡 − 3𝜏 . ❖ The input–output equation is
Hence 𝑦 𝑡 − 𝑛𝜏 = 𝑥 𝑡 − 𝑛𝜏 + 𝛽𝑦 𝑡 − 𝑛𝜏 . continuously grows as time increases.
❖ From here, we obtain 𝑦 𝑡 as follows ❖ The output is clearly not a bounded
𝑦 𝑡 = 𝑥 𝑡 + 𝛽𝑥 𝑡 − 𝜏 + 𝛽2 𝑦 𝑡 − 2𝜏 signal. Thus the systems is unstable
= 𝑥 𝑡 + 𝛽𝑥 𝑡 − 𝜏 + 𝛽2 𝑥 𝑡 − 2𝜏
+ 𝛽3 𝑦 𝑡 − 3𝜏
❖ Therefore
𝑦 𝑡 = 𝑥 𝑡 +𝛽𝑥 𝑡 − 𝜏 + 𝛽𝑥 𝑡 − 𝜏
+𝛽2 𝑥 𝑡 − 2𝜏 + ⋯ .
❖ Determine if the system is BIBO stable or
not; use x(t) = u(t), β = 2 and τ = 1, where
10
𝑢 𝑡 is the unit step. We obtain → → →
Positive feedback system: microphone. 10
RL circuit
Example 2.18: A series RL circuit where R = 1Ω
and L = 1 H, and a voltage source 𝑣𝑠 (𝑡) which is
bounded. Discus why such a system is stable.

According to Kirchhoff’s voltage law, a circuit

represented by a first order ordinary differential
equation
𝐿𝑑𝑖 𝑡 𝑑𝑖 𝑡 di 𝑡
𝑣𝑠 𝑡 = 𝑖 𝑡 𝑅 + =𝑖 𝑡 + Hence = (𝑒 −𝑡 𝛿 𝑡 − 𝑒 −𝑡 𝑢 𝑡 )
𝑑𝑡 𝑑𝑡 𝑑𝑡
Let 𝑥 𝑡 = 𝑣𝑠 (𝑡) = 𝜹(𝒕), with zero initial 𝛿 𝑡 = 𝑒 −𝑡 𝑢 𝑡 + (𝑒 −𝑡 𝛿 𝑡 − 𝑒 −𝑡 𝑢 𝑡 )
conditions, we can obtain the impulse response = 𝑒 0 𝛿 𝑡 = 𝛿(𝑡)
ℎ 𝑡 = 𝑦 𝑡 = 𝑖 𝑡 = 𝑒 −𝑡 𝑢(𝑡) ∞ ∞
Let we check to confirm the above result න ȁℎ 𝑡 ȁ𝑑𝑡 = න 𝑒 −𝑡 𝑢 𝑡 𝑑𝑡
−𝑡 −∞ −∞
di 𝑡 𝑑 𝑢 𝑡 𝑑 𝑒 ∞
= 𝑒 −𝑡 + 𝑢 𝑡
𝑑𝑡 𝑑𝑡
−𝑡
𝑑𝑡 = න 𝑒 −𝑡 𝑑𝑡 = 𝑒 −𝑡 < ∞ 11
𝑑 𝑒 0
Note that 𝑢ሶ 𝑡 = 𝛿 𝑡 , = 𝑒 −𝑡
𝑑𝑡 11