Boo-AI — Master Artificial Intelligence by Building from Scratch

Learning Objectives

By the end of this section, you will be able to:

Apply the linearity property to simplify complex transform calculations
Use the first shifting theorem (s-shifting) to handle exponential multipliers
Apply the second shifting theorem (t-shifting) for delayed signals
Transform derivatives using the derivative property to solve differential equations
Understand how integration in time relates to division by s
Visualize how these properties move poles in the complex s-plane
Connect Laplace properties to control systems, signal processing, and machine learning

The Big Picture: Why Properties Matter

"The Laplace transform converts calculus into algebra. Its properties are the rules of this algebraic game." — A control systems engineer's wisdom

In the previous section, we defined the Laplace transform as an integral that converts time-domain functions into s-domain representations. But computing transforms from the definition is tedious. The true power of the Laplace transform lies in its properties — systematic rules that let us build complex transforms from simple ones.

Consider this analogy: computing $\log(a \cdot b)$ directly is hard, but knowing that $\log(a \cdot b) = \log a + \log b$ transforms multiplication into addition. Similarly, Laplace transform properties turn difficult operations (differentiation, integration, convolution) into simple ones (multiplication, division).

The Central Insight

Each Laplace transform property establishes a correspondence between an operation in the time domain and an operation in the s-domain. The s-domain operation is almost always simpler!

Time Domain Operation	s-Domain Operation
Differentiation f'(t)	Multiplication by s
Integration ∫f(τ)dτ	Division by s
Convolution f * g	Multiplication F · G
Multiplication by e^(at)	Shift F(s) → F(s-a)
Delay by a: f(t-a)	Multiply by e^(-as)

Where These Properties Are Used

⚡ Electrical Engineering

Circuit analysis with capacitors and inductors
Transfer function design
Frequency response analysis
Filter design and stability

🔧 Control Systems

PID controller tuning
Stability analysis (pole locations)
System response prediction
Feedback loop design

📡 Signal Processing

Convolution for filtering
System identification
Spectral analysis
Communication systems

🤖 Machine Learning

Neural ODE analysis
Continuous-time models
Dynamical system learning
Optimal control for robotics

Historical Origins: From Operational Calculus to Modern Engineering

The properties of the Laplace transform were developed gradually over nearly two centuries, driven by practical engineering needs.

Oliver Heaviside and Operational Methods

In the 1880s, British engineer Oliver Heaviside (1850–1925) developed "operational calculus" to solve the telegraph equation and analyze electrical circuits. He treated the differential operator $D = d/dt$ as an algebraic quantity, writing equations like:

(D^2 + 3D + 2)y = f(t) \quad \Rightarrow \quad y = \frac{1}{D^2 + 3D + 2} f(t)

Heaviside's methods worked brilliantly in practice, but mathematicians of his era considered them non-rigorous "voodoo." It wasn't until the 20th century that the Laplace transform provided the rigorous foundation for his operational rules.

The Laplace Transform Provides the Foundation

The key insight: Heaviside's operator $D$ corresponds to multiplication by $s$ in the Laplace domain. His algebraic manipulations were implicitly using Laplace transform properties!

Heaviside's Legacy

Despite being dismissed by academics, Heaviside's methods revolutionized electrical engineering. Today, every EE student learns the "proper" Laplace transform version of techniques Heaviside invented through intuition and practical necessity.

The Linearity Property

The most fundamental property is linearity: the Laplace transform respects addition and scalar multiplication.

Linearity Property

If $\\mathcal{L}\\{f(t)\\} = F(s)$ and $\\mathcal{L}\\{g(t)\\} = G(s)$ , then for any constants $a$ and $b$ :

\\mathcal{L}\\{af(t) + bg(t)\\} = aF(s) + bG(s)

Why Linearity Works

This follows directly from the linearity of integration:

\\mathcal{L}\\{af + bg\\} = \\int_0^\\infty e^{-st}[af(t) + bg(t)]\\,dt

$= a\\int_0^\\infty e^{-st}f(t)\\,dt + b\\int_0^\\infty e^{-st}g(t)\\,dt = aF(s) + bG(s)$

Example: Combining Known Transforms

Find $\\mathcal{L}\\{3\\sin(2t) - 5e^{-t}\\}$ .

Solution: Using known transforms and linearity:

$\\mathcal{L}\\{\\sin(2t)\\} = \\frac{2}{s^2 + 4}$

$\\mathcal{L}\\{e^{-t}\\} = \\frac{1}{s + 1}$

Therefore: $\\mathcal{L}\\{3\\sin(2t) - 5e^{-t}\\} = \\frac{6}{s^2 + 4} - \\frac{5}{s + 1}$

Practical Value

Linearity lets us break complex signals into simple components, transform each one, and add the results. This is the foundation of spectral analysis and superposition in physics.

The First Shifting Theorem (s-Shifting)

This property handles multiplication by exponentials in the time domain — one of the most commonly occurring situations in physics and engineering.

First Shifting Theorem (s-Shifting)

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then:

\\mathcal{L}\\{e^{at}f(t)\\} = F(s - a)

Multiplying by $e^{at}$ in time shifts the transform to the right by $a$ in the s-domain.

Proof

\\mathcal{L}\\{e^{at}f(t)\\} = \\int_0^\\infty e^{-st} \\cdot e^{at} f(t)\\,dt

$= \\int_0^\\infty e^{-(s-a)t} f(t)\\,dt = F(s-a)$

The exponential $e^{at}$ combines with $e^{-st}$ from the transform definition, effectively replacing $s$ with $s - a$ .

Example: Damped Oscillation

Find $\\mathcal{L}\\{e^{-2t}\\cos(3t)\\}$ .

Solution:

We know $\\mathcal{L}\\{\\cos(3t)\\} = \\frac{s}{s^2 + 9}$

By the first shifting theorem with $a = -2$ :

$\\mathcal{L}\\{e^{-2t}\\cos(3t)\\} = \\frac{s+2}{(s+2)^2 + 9} = \\frac{s+2}{s^2 + 4s + 13}$

Engineering Interpretation

In the s-domain, the first shifting theorem moves poles horizontally. This is crucial for understanding stability: multiplying a signal by $e^{-at}$ with $a > 0$ shifts poles to the left, making systems more stable.

Pole Shifting in the s-Plane

This visualizes how the first shifting theorem moves poles in the complex s-plane. When poles cross the imaginary axis to the right half-plane, the system becomes unstable.

Shift parameter a: 2.00

Frequency ω: 3.00

The Second Shifting Theorem (t-Shifting)

While the first shifting theorem handles exponential multipliers, the second shifting theorem handles time delays — signals that "turn on" after a delay.

Second Shifting Theorem (t-Shifting)

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then for $a > 0$ :

\\mathcal{L}\\{u(t-a) \\cdot f(t-a)\\} = e^{-as}F(s)

where $u(t-a)$ is the unit step function (Heaviside function), which is 0 for $t < a$ and 1 for $t \\geq a$ .

Understanding the Unit Step Function

The unit step function "switches on" a signal at a specified time:

u(t-a) = \\begin{cases} 0 & \\text{if } t < a \\\\ 1 & \\text{if } t \\geq a \\end{cases}

The combination $u(t-a) \\cdot f(t-a)$ means: "wait until time $a$ , then start the function $f$ from its beginning."

Example: Delayed Pulse

Find the Laplace transform of a unit step delayed by 3 seconds: $\\mathcal{L}\\{u(t-3)\\}$ .

Solution:

Since $\\mathcal{L}\\{u(t)\\} = \\mathcal{L}\\{1\\} = \\frac{1}{s}$ ,

by the second shifting theorem: $\\mathcal{L}\\{u(t-3)\\} = e^{-3s} \\cdot \\frac{1}{s} = \\frac{e^{-3s}}{s}$

Physical Interpretation

In control systems, time delays are common: sensor delays, communication latency, transportation lags. The factor $e^{-as}$ captures this delay in the s-domain. Large delays can destabilize feedback systems!

Transform of Derivatives: The Key to Solving ODEs

This is arguably the most important property of the Laplace transform. It converts differentiation into multiplication, turning differential equations into algebraic equations.

Derivative Property

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then:

\\mathcal{L}\\{f'(t)\\} = sF(s) - f(0)

For the second derivative:

\\mathcal{L}\\{f''(t)\\} = s^2F(s) - sf(0) - f'(0)

The general pattern for the $n$ -th derivative:

\\mathcal{L}\\{f^{(n)}(t)\\} = s^nF(s) - s^{n-1}f(0) - s^{n-2}f'(0) - \\cdots - f^{(n-1)}(0)

Proof of the First Derivative Property

Using integration by parts with $u = f(t)$ and $dv = e^{-st}dt$ :

\\mathcal{L}\\{f'(t)\\} = \\int_0^\\infty e^{-st}f'(t)\\,dt

$= \\left[e^{-st}f(t)\\right]_0^\\infty + s\\int_0^\\infty e^{-st}f(t)\\,dt$

$= (0 - f(0)) + sF(s) = sF(s) - f(0)$

(assuming $f(t)$ doesn't grow faster than $e^{st}$ )

Example: Solving a First-Order ODE

Solve $y' + 3y = e^{-t}$ with $y(0) = 2$ .

Step 1: Take the Laplace transform of both sides:

\\mathcal{L}\\{y'\\} + 3\\mathcal{L}\\{y\\} = \\mathcal{L}\\{e^{-t}\\}

Step 2: Apply the derivative property:

[sY(s) - y(0)] + 3Y(s) = \\frac{1}{s+1}

[sY(s) - 2] + 3Y(s) = \\frac{1}{s+1}

Step 3: Solve for $Y(s)$ :

(s+3)Y(s) = \\frac{1}{s+1} + 2

Y(s) = \\frac{1}{(s+1)(s+3)} + \\frac{2}{s+3}

Step 4: Use partial fractions and inverse transform (covered in Section 3).

The Power of This Property

Notice how the differential equation became an algebraic equation! Initial conditions appear automatically in the transform. This systematic approach works for any order ODE with constant coefficients.

Transform of Integrals

Just as differentiation becomes multiplication by $s$ , integration becomes division by $s$ .

Integration Property

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then:

\\mathcal{L}\\left\\{\\int_0^t f(\\tau)\\,d\\tau\\right\\} = \\frac{F(s)}{s}

Why This Makes Sense

Integration is the inverse of differentiation. Since differentiation multiplies by $s$ , integration should divide by $s$ . More formally:

Let $g(t) = \\int_0^t f(\\tau)\\,d\\tau$ . Then $g'(t) = f(t)$ and $g(0) = 0$ .

By the derivative property: $\\mathcal{L}\\{g'(t)\\} = sG(s) - g(0)$

So: $F(s) = sG(s) - 0$ , giving $G(s) = F(s)/s$

Example: Integral of Sine

Find $\\mathcal{L}\\left\\{\\int_0^t \\sin(\\tau)\\,d\\tau\\right\\}$ .

Solution:

We know $\\mathcal{L}\\{\\sin(t)\\} = \\frac{1}{s^2+1}$

By the integration property:

\\mathcal{L}\\left\\{\\int_0^t \\sin(\\tau)\\,d\\tau\\right\\} = \\frac{1}{s} \\cdot \\frac{1}{s^2+1} = \\frac{1}{s(s^2+1)}

(Note: $\\int_0^t \\sin(\\tau)\\,d\\tau = 1 - \\cos(t)$ , and we can verify this has the same transform!)

The Scaling Property

What happens when we "speed up" or "slow down" a signal by scaling the time variable?

Scaling Property

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then for $a > 0$ :

\\mathcal{L}\\{f(at)\\} = \\frac{1}{a}F\\left(\\frac{s}{a}\\right)

Intuition

Speeding up a signal in time ( $a > 1$ ) compresses it horizontally. In the frequency domain, this spreads out the spectrum and reduces amplitude. The factor $1/a$ preserves the total "energy" of the signal.

Example

Find $\\mathcal{L}\\{\\sin(3t)\\}$ using scaling.

Solution:

Start with $\\mathcal{L}\\{\\sin(t)\\} = \\frac{1}{s^2+1}$

Apply scaling with $a = 3$ :

\\mathcal{L}\\{\\sin(3t)\\} = \\frac{1}{3} \\cdot \\frac{1}{(s/3)^2+1} = \\frac{1}{3} \\cdot \\frac{1}{s^2/9+1} = \\frac{1}{3} \\cdot \\frac{9}{s^2+9} = \\frac{3}{s^2+9}

Connection to Uncertainty Principle

The scaling property embodies a fundamental trade-off: compressing a signal in time spreads it in frequency (and vice versa). This is the mathematical basis of the Heisenberg uncertainty principle in quantum mechanics!

Multiplication by t: Frequency Differentiation

This elegant property relates multiplication by $t$ in the time domain to differentiation in the s-domain.

Multiplication by t Property

If $\\mathcal{L}\\{f(t)\\} = F(s)$ , then:

\\mathcal{L}\\{t \\cdot f(t)\\} = -\\frac{d}{ds}F(s) = -F'(s)

More generally:

\\mathcal{L}\\{t^n \\cdot f(t)\\} = (-1)^n \\frac{d^n}{ds^n}F(s)

Proof

Starting from $F(s) = \\int_0^\\infty e^{-st}f(t)\\,dt$ , differentiate with respect to $s$ :

\\frac{dF}{ds} = \\frac{d}{ds}\\int_0^\\infty e^{-st}f(t)\\,dt = \\int_0^\\infty \\frac{\\partial}{\\partial s}(e^{-st})f(t)\\,dt

$= \\int_0^\\infty (-t)e^{-st}f(t)\\,dt = -\\mathcal{L}\\{tf(t)\\}$

Example: Finding the Transform of t·sin(ωt)

Solution:

Start with $\\mathcal{L}\\{\\sin(\\omega t)\\} = \\frac{\\omega}{s^2+\\omega^2}$

Apply the property:

\\mathcal{L}\\{t\\sin(\\omega t)\\} = -\\frac{d}{ds}\\left(\\frac{\\omega}{s^2+\\omega^2}\\right)

$= -\\omega \\cdot \\frac{-2s}{(s^2+\\omega^2)^2} = \\frac{2\\omega s}{(s^2+\\omega^2)^2}$

Dual Property

There's a dual: dividing by $t$ corresponds to integrating in the s-domain. However, division by $t$ requires $f(0) = 0$ to be well-defined.

Interactive Property Explorer

Use this interactive visualization to explore how Laplace transform properties affect signals in both the time and s-domains.

Laplace Transform Property Visualizer

Shift parameter a: 2.00

Instructions: Select a property to visualize. The left panel shows the time-domain function, and the right panel shows the s-domain (frequency) representation. Adjust the parameter slider to see how changes affect both domains.

Engineering Applications

Circuit Analysis: RC Circuit Step Response

Consider an RC circuit with a capacitor initially uncharged. When a voltage $V_0$ is applied at $t = 0$ , Kirchhoff's voltage law gives:

RC\\frac{dv_C}{dt} + v_C = V_0, \\quad v_C(0) = 0

Taking the Laplace transform and using the derivative property:

RC[sV_C(s) - 0] + V_C(s) = \\frac{V_0}{s}

$V_C(s) = \\frac{V_0}{s(RCs + 1)} = \\frac{V_0}{s(s + 1/RC)}$

The solution (using partial fractions and inverse transform, covered in Section 3) is the familiar exponential charging curve:

v_C(t) = V_0(1 - e^{-t/RC})

Mechanical Vibrations: Damped Spring-Mass System

A mass $m$ on a spring with damping follows:

m\\ddot{x} + c\\dot{x} + kx = F(t)

The Laplace transform converts this to an algebraic equation for $X(s)$ :

(ms^2 + cs + k)X(s) = F_{\\text{ext}}(s) + (ms + c)x(0) + m\\dot{x}(0)

The characteristic polynomial $ms^2 + cs + k$ determines system behavior. The first shifting theorem helps us understand how damping $c$ affects pole locations and stability.

Machine Learning Connections

While discrete methods dominate modern ML, continuous-time models are making a comeback. Laplace transform properties are directly relevant to several cutting-edge areas:

Neural Ordinary Differential Equations (Neural ODEs)

Neural ODEs define neural network layers as continuous dynamics:

\\frac{dh}{dt} = f_\\theta(h(t), t)

The derivative property of Laplace transforms helps analyze the stability and expressiveness of these continuous networks. Poles of the linearized system must lie in the left half-plane for stable training.

Continuous Normalizing Flows

For generative modeling, continuous normalizing flows model probability density transformations via ODEs. The scaling property relates transformations in time to changes in probability density through the instantaneous change of variables formula.

Control-as-Learning

Modern reinforcement learning for robotics often uses optimal control methods. The Laplace transform is fundamental to:

Designing stable controllers (pole placement)
Analyzing system response to learned policies
Understanding latency effects (time delays via second shifting)
Transfer function identification from data

The Continuous-Discrete Bridge

The Z-transform used in discrete-time signal processing is the discrete analog of the Laplace transform. The relationship $z = e^{sT}$ connects continuous and discrete domains, allowing translation of Laplace properties to digital systems.

Python Implementation

Demonstrating Key Properties with SymPy

Laplace Transform Properties in Python

🐍laplace_properties.py

Explanation(5)

Code(147)

1Scientific Computing Stack

We use NumPy for numerical arrays, Matplotlib for visualization, SciPy for signal processing, and SymPy for symbolic mathematics. SymPy can compute exact Laplace transforms algebraically.

15Linearity Property

The Laplace transform is linear, meaning L{af + bg} = aF + bG. This allows us to break complex functions into simpler parts, transform each separately, and combine the results.

EXAMPLE

L{3sin(t) + 2cos(t)} = 3·(1/(s²+1)) + 2·(s/(s²+1))

42First Shifting Theorem

Multiplying f(t) by e^(at) shifts F(s) to the right by 'a' in the s-domain. This is crucial for handling exponentially growing/decaying signals and resonance phenomena.

EXAMPLE

L{e^(3t)sin(2t)} = 2/((s-3)² + 4)

68Derivative Property

This is the most powerful property for solving ODEs. Taking a derivative in time becomes multiplication by s in the s-domain, converting differential equations to algebraic equations.

EXAMPLE

y'' + 3y' + 2y = f(t) becomes (s² + 3s + 2)Y(s) = F(s) + initial conditions

96Integration Property

Integration in time becomes division by s. This is useful for solving integral equations and understanding cumulative effects in systems.

EXAMPLE

L{∫e^(-τ)dτ} = (1/s)·(1/(s+1)) = 1/(s(s+1))

142 lines without explanation

1import numpy as np
2import matplotlib.pyplot as plt
3from scipy import signal
4from sympy import symbols, exp, sin, cos, laplace_transform, inverse_laplace_transform
5from sympy import Heaviside, DiracDelta, simplify
6
7# Define symbolic variables
8t, s, a = symbols('t s a', positive=True, real=True)
9
10# =====================================================
11# PROPERTY 1: LINEARITY
12# L{af(t) + bg(t)} = aF(s) + bG(s)
13# =====================================================
14
15def demonstrate_linearity():
16    """
17    The Laplace transform is linear:
18    - L{c*f(t)} = c*F(s)   (scaling)
19    - L{f(t) + g(t)} = F(s) + G(s)  (additivity)
20    """
21    # Define functions
22    f = sin(t)  # L{sin(t)} = 1/(s^2+1)
23    g = cos(t)  # L{cos(t)} = s/(s^2+1)
24
25    # Compute transforms
26    F, _, _ = laplace_transform(f, t, s)
27    G, _, _ = laplace_transform(g, t, s)
28
29    # Linear combination
30    h = 3*f + 2*g
31    H, _, _ = laplace_transform(h, t, s)
32
33    # Verify: should equal 3*F + 2*G
34    H_expected = simplify(3*F + 2*G)
35
36    print("LINEARITY PROPERTY:")
37    print(f"  L{{sin(t)}} = {F}")
38    print(f"  L{{cos(t)}} = {G}")
39    print(f"  L{{3sin(t) + 2cos(t)}} = {simplify(H)}")
40    print(f"  3F + 2G = {H_expected}")
41    print(f"  Equal: {simplify(H - H_expected) == 0}")
42    print()
43
44# =====================================================
45# PROPERTY 2: FIRST SHIFTING THEOREM (s-shifting)
46# L{e^(at)f(t)} = F(s-a)
47# =====================================================
48
49def demonstrate_s_shifting():
50    """
51    Multiplying by e^(at) in time domain
52    shifts the transform by 'a' in s-domain.
53
54    This is used for solving ODEs with exponential forcing.
55    """
56    # Original function: f(t) = sin(ωt)
57    omega = 2
58    f = sin(omega * t)
59    F, _, _ = laplace_transform(f, t, s)
60
61    # Shifted function: e^(at) * f(t)
62    a_val = 3
63    f_shifted = exp(a_val * t) * f
64    F_shifted, _, _ = laplace_transform(f_shifted, t, s)
65
66    # Verify: should be F(s-a)
67    F_expected = F.subs(s, s - a_val)
68
69    print("FIRST SHIFTING THEOREM (s-shifting):")
70    print(f"  f(t) = sin({omega}t)")
71    print(f"  L{{f(t)}} = {F}")
72    print(f"  L{{e^({a_val}t)f(t)}} = {simplify(F_shifted)}")
73    print(f"  F(s-{a_val}) = {simplify(F_expected)}")
74    print()
75
76# =====================================================
77# PROPERTY 3: DERIVATIVE PROPERTY
78# L{f'(t)} = sF(s) - f(0)
79# =====================================================
80
81def demonstrate_derivative_property():
82    """
83    The derivative in time domain becomes
84    multiplication by s in s-domain (plus initial conditions).
85
86    This is THE key property for solving differential equations!
87    """
88    # f(t) = t^2, so f'(t) = 2t
89    f = t**2
90    f_prime = 2*t
91
92    F, _, _ = laplace_transform(f, t, s)
93    F_prime, _, _ = laplace_transform(f_prime, t, s)
94
95    # Initial condition: f(0) = 0
96    f_0 = 0
97
98    # Verify: L{f'(t)} = sF(s) - f(0)
99    F_derivative_expected = s * F - f_0
100
101    print("DERIVATIVE PROPERTY:")
102    print(f"  f(t) = t^2, f'(t) = 2t, f(0) = 0")
103    print(f"  L{{t^2}} = {F}")
104    print(f"  L{{2t}} = {simplify(F_prime)}")
105    print(f"  sF(s) - f(0) = {simplify(F_derivative_expected)}")
106    print()
107
108    # Second derivative: L{f''(t)} = s^2*F(s) - s*f(0) - f'(0)
109    print("  For L{f''(t)} = s^2*F(s) - s*f(0) - f'(0)")
110    print("  This converts 2nd order ODEs to algebraic equations!")
111    print()
112
113# =====================================================
114# PROPERTY 4: INTEGRATION PROPERTY
115# L{∫f(τ)dτ} = F(s)/s
116# =====================================================
117
118def demonstrate_integration_property():
119    """
120    Integration in time domain becomes
121    division by s in s-domain.
122
123    Useful for solving integral equations.
124    """
125    # f(t) = e^(-t), integral = 1 - e^(-t)
126    f = exp(-t)
127    f_integral = 1 - exp(-t)  # from 0 to t
128
129    F, _, _ = laplace_transform(f, t, s)
130    F_integral, _, _ = laplace_transform(f_integral, t, s)
131
132    # Verify: should equal F(s)/s
133    F_over_s = simplify(F / s)
134
135    print("INTEGRATION PROPERTY:")
136    print(f"  f(t) = e^(-t)")
137    print(f"  ∫f(τ)dτ from 0 to t = 1 - e^(-t)")
138    print(f"  L{{e^(-t)}} = {F}")
139    print(f"  L{{1 - e^(-t)}} = {simplify(F_integral)}")
140    print(f"  F(s)/s = {F_over_s}")
141    print()
142
143# Run demonstrations
144demonstrate_linearity()
145demonstrate_s_shifting()
146demonstrate_derivative_property()
147demonstrate_integration_property()

Control System Analysis: Effect of Pole Shifting

Engineering Application: Pole Shifting and Stability

🐍pole_shifting_analysis.py

Explanation(5)

Code(104)

3System Analysis Setup

We use SciPy's signal processing module to create transfer functions and analyze system behavior. This demonstrates how Laplace transform properties apply to real control systems.

14Pole Shifting Physical Meaning

When we multiply a signal by e^(-at), we're adding damping. In the s-domain, this shifts all poles left by 'a'. This is fundamental to understanding stability in control systems.

EXAMPLE

A pole at s = jω (oscillatory) becomes s = -a + jω (damped oscillation)

30Transfer Function Creation

The transfer function H(s) = ω²/(s² + 2as + (ω² + a²)) represents the damped system. The denominator roots (poles) determine system behavior.

45Step Response Comparison

The step response shows how the system responds to a sudden input. The undamped system oscillates forever, while the damped system settles to a steady state.

55Pole-Zero Analysis

Poles in the left half-plane (negative real part) create stable, decaying responses. The first shifting theorem lets us move poles to achieve desired stability margins.

99 lines without explanation

1import numpy as np
2import matplotlib.pyplot as plt
3from scipy import signal
4
5def analyze_system_with_shifting():
6    """
7    Use the first shifting theorem to analyze
8    how exponential damping affects system response.
9
10    Original system: H(s) = ω²/(s² + ω²)  (undamped oscillator)
11    With damping: multiply by e^(-at), shift poles
12
13    The shifting theorem tells us:
14    L{e^(-at)f(t)} = F(s+a)
15
16    So shifting the transfer function moves poles!
17    """
18    omega = 2  # Natural frequency
19
20    # Undamped system: poles at s = ±jω
21    num_undamped = [omega**2]
22    den_undamped = [1, 0, omega**2]
23
24    # Damped system (a = 0.5): poles shifted to s = -0.5 ± jω
25    a = 0.5
26    # H(s+a) has poles at s+a = ±jω, so s = -a ± jω
27    num_damped = [omega**2]
28    den_damped = [1, 2*a, omega**2 + a**2]
29
30    # Create transfer functions
31    sys_undamped = signal.TransferFunction(num_undamped, den_undamped)
32    sys_damped = signal.TransferFunction(num_damped, den_damped)
33
34    # Step response
35    t = np.linspace(0, 15, 1000)
36    _, y_undamped = signal.step(sys_undamped, T=t)
37    _, y_damped = signal.step(sys_damped, T=t)
38
39    # Plot
40    fig, axes = plt.subplots(2, 2, figsize=(12, 10))
41
42    # Step responses
43    axes[0, 0].plot(t, y_undamped, 'b-', label='Undamped', linewidth=2)
44    axes[0, 0].plot(t, y_damped, 'g-', label=f'Damped (a={a})', linewidth=2)
45    axes[0, 0].set_xlabel('Time (s)')
46    axes[0, 0].set_ylabel('Response')
47    axes[0, 0].set_title('Step Response: Effect of Pole Shifting')
48    axes[0, 0].legend()
49    axes[0, 0].grid(True, alpha=0.3)
50
51    # Pole-zero plot
52    poles_undamped = np.roots(den_undamped)
53    poles_damped = np.roots(den_damped)
54
55    axes[0, 1].scatter(poles_undamped.real, poles_undamped.imag,
56                       marker='x', s=100, c='blue', label='Undamped poles')
57    axes[0, 1].scatter(poles_damped.real, poles_damped.imag,
58                       marker='x', s=100, c='green', label='Damped poles')
59    axes[0, 1].axvline(x=0, color='r', linestyle='--', alpha=0.5, label='Stability boundary')
60    axes[0, 1].set_xlabel('Real(s)')
61    axes[0, 1].set_ylabel('Imag(s)')
62    axes[0, 1].set_title('Pole Locations in s-Plane')
63    axes[0, 1].legend()
64    axes[0, 1].grid(True, alpha=0.3)
65    axes[0, 1].set_aspect('equal')
66
67    # Bode magnitude plot
68    w = np.logspace(-1, 2, 1000)
69    _, mag_undamped, _ = signal.bode(sys_undamped, w)
70    _, mag_damped, _ = signal.bode(sys_damped, w)
71
72    axes[1, 0].semilogx(w, mag_undamped, 'b-', label='Undamped', linewidth=2)
73    axes[1, 0].semilogx(w, mag_damped, 'g-', label='Damped', linewidth=2)
74    axes[1, 0].set_xlabel('Frequency (rad/s)')
75    axes[1, 0].set_ylabel('Magnitude (dB)')
76    axes[1, 0].set_title('Frequency Response (Bode Plot)')
77    axes[1, 0].legend()
78    axes[1, 0].grid(True, alpha=0.3)
79
80    # Show how shifting affects frequency response
81    # Impulse response
82    _, y_imp_undamped = signal.impulse(sys_undamped, T=t)
83    _, y_imp_damped = signal.impulse(sys_damped, T=t)
84
85    axes[1, 1].plot(t, y_imp_undamped, 'b-', label='Undamped', linewidth=2)
86    axes[1, 1].plot(t, y_imp_damped, 'g-', label='Damped', linewidth=2)
87    axes[1, 1].set_xlabel('Time (s)')
88    axes[1, 1].set_ylabel('Response')
89    axes[1, 1].set_title('Impulse Response')
90    axes[1, 1].legend()
91    axes[1, 1].grid(True, alpha=0.3)
92
93    plt.tight_layout()
94    plt.show()
95
96    print("ENGINEERING INSIGHT:")
97    print(f"  Undamped poles: {poles_undamped}")
98    print(f"  Damped poles:   {poles_damped}")
99    print()
100    print("  The first shifting theorem tells us that multiplying")
101    print("  by e^(-at) shifts all poles LEFT by 'a' units.")
102    print("  This moves poles away from the instability region!")
103
104analyze_system_with_shifting()

Common Pitfalls

Pitfall 1: Forgetting Initial Conditions

The derivative property is $\\mathcal{L}\\{f'\\} = sF(s) - f(0)$ , not just $sF(s)$ . Forgetting initial conditions gives wrong answers for differential equations!

Pitfall 2: Confusing the Two Shifting Theorems

First shifting (s-shift): Multiplying by $e^{at}$ shifts $F(s)$ .

Second shifting (t-shift): Delaying by $a$ multiplies by $e^{-as}$ .

The directions are opposite! Time shift → exponential multiplier. Exponential multiplier → frequency shift.

Pitfall 3: Domain of Convergence

The first shifting theorem $\\mathcal{L}\\{e^{at}f(t)\\} = F(s-a)$ requires $\\text{Re}(s) > \\text{Re}(a) + \\sigma_0$ where $\\sigma_0$ is the original convergence abscissa. Multiplying by a growing exponential narrows the region of convergence.

Sign Convention

Some books define the first shifting theorem as $\\mathcal{L}\\{e^{-at}f(t)\\} = F(s+a)$ (note the signs). Always check which convention your textbook uses!

Complete Property Reference

Property	Time Domain	s-Domain
Linearity	af(t) + bg(t)	aF(s) + bG(s)
First Shifting	e^(at)f(t)	F(s - a)
Second Shifting	u(t-a)f(t-a)	e^(-as)F(s)
First Derivative	f'(t)	sF(s) - f(0)
Second Derivative	f''(t)	s²F(s) - sf(0) - f'(0)
n-th Derivative	f^(n)(t)	s^n F(s) - (initial terms)
Integration	∫₀ᵗ f(τ)dτ	F(s)/s
Scaling	f(at), a > 0	(1/a)F(s/a)
Multiplication by t	tf(t)	-F'(s)
Multiplication by t^n	t^n f(t)	(-1)^n F^(n)(s)
Division by t	f(t)/t	∫ₛ^∞ F(u)du (if exists)
Convolution	(f * g)(t)	F(s) · G(s)
Initial Value	f(0+)	lim_{s→∞} sF(s)
Final Value	lim_{t→∞} f(t)	lim_{s→0} sF(s) (if exists)

Summary

The properties of the Laplace transform form a powerful toolkit for analyzing differential equations, control systems, and signal processing problems.

Key Concepts

Linearity allows decomposing complex signals into simple components and transforming each separately.
First shifting theorem: Multiplication by $e^{at}$ shifts $F(s)$ by $a$ — essential for damped oscillations and stability analysis.
Second shifting theorem: Time delays introduce exponential factors $e^{-as}$ — crucial for systems with latency.
Derivative property: Differentiation becomes multiplication by $s$ — the key to solving ODEs algebraically.
Integration property: Integration becomes division by $s$ .
Scaling property: Time compression spreads the frequency spectrum, embodying the uncertainty principle.
Multiplication by t: Connects time-domain weighting to s-domain differentiation.

Engineering Significance

For Control Engineers

The shifting theorems reveal how exponential damping and time delays affect pole locations and system stability.

For Signal Processors

Properties like scaling and convolution are fundamental to filter design and spectral analysis.

For ML Researchers

Neural ODEs and continuous models rely on understanding how transformations in time affect system dynamics.

For Applied Mathematicians

These properties transform calculus problems into algebra, providing systematic solution methods for differential and integral equations.

The Essence of Laplace Properties:

"The Laplace transform properties establish a dictionary between calculus operations and algebra. Differentiation becomes multiplication; integration becomes division; convolution becomes multiplication. Complex dynamic systems become polynomial equations."

Coming Next: In the next section, we'll explore Inverse Laplace Transforms — how to recover time-domain functions from their s-domain representations using partial fractions, convolution, and the Bromwich integral.