Formula sheet

Signals and Systems Formula Sheet

From computing the Fourier transform of a rectangular pulse in a GATE 2023 problem to designing a causal IIR filter in a DSP lab, signals and systems formulas are the backbone of every frequency-domain analysis task. These formulas cover continuous-time and discrete-time signals, Fourier series and transforms, Laplace and Z-transforms, convolution, correlation, sampling, and LTI system properties — fully aligned with GATE ECE/EEE, ESE, and BITS Pilani semester syllabi.

EEE, ECE, EI

Fourier Series

Exponential Fourier Series Coefficients

c_n = \frac{1}{T} \int_{0}^{T} x(t)\, e^{-jn\omega_0 t}\, dt

SymbolDescriptionUnit
c_nComplex Fourier coefficientV (or signal units)
TFundamental periods
\omega_0Fundamental angular frequency = 2π/Trad/s

Worked example

Find c_0 (DC component) of x(t) = 2 + 3cos(2πt) + sin(4πt).

Given: x(t) = 2 + 3cos(2πt) + sin(4πt)

  1. c_0 = (1/T) ∫₀ᵀ x(t) dt
  2. Average of cos and sin over full period = 0
  3. c_0 = (1/T) ∫₀ᵀ 2 dt = 2

Answer: c_0 = 2 (DC component)

Parseval's Theorem (Power, Fourier Series)

P = \frac{1}{T} \int_0^T |x(t)|^2 dt = \sum_{n=-\infty}^{\infty} |c_n|^2

SymbolDescriptionUnit
PAverage powerW

Worked example

Find average power of x(t) = 3cos(2πt) + 4sin(4πt).

Given: Amplitudes: 3 and 4

  1. For cosine: c_1 = c_{-1} = 3/2, |c_±1|² = 9/4 each → contribution = 9/4 + 9/4 = 4.5
  2. For sine: c_2 = -j4/2 = -j2, |c_±2|² = 4 each → contribution = 4 + 4 = 8
  3. P = 4.5 + 8 = 12.5 W
  4. Check: P = (3²/2) + (4²/2) = 4.5 + 8 = 12.5 ✓

Answer: P = 12.5 W

Trigonometric Fourier Coefficients

a_n = \frac{2}{T}\int_0^T x(t)\cos(n\omega_0 t)\,dt,\quad b_n = \frac{2}{T}\int_0^T x(t)\sin(n\omega_0 t)\,dt

SymbolDescriptionUnit
a_nCosine coefficientsignal units
b_nSine coefficientsignal units

Worked example

For a square wave x(t) with amplitude 1 and period T, find a_1.

Given: x(t) = +1 for 0 < t < T/2, -1 for T/2 < t < T

  1. a_1 = (2/T)[∫₀^{T/2} 1·cos(2πt/T)dt + ∫_{T/2}^T (-1)·cos(2πt/T)dt]
  2. = (2/T)[T/(2π)·sin(2πt/T)]₀^{T/2} - (2/T)[T/(2π)·sin(2πt/T)]_{T/2}^T
  3. = (1/π)[sin(π) - sin(0)] - (1/π)[sin(2π) - sin(π)]
  4. = (1/π)[0] - (1/π)[0] = 0

Answer: a_1 = 0 (square wave has no cosine terms — it is an odd function)

Continuous-Time Fourier Transform (CTFT)

Fourier Transform Definition

X(j\omega) = \int_{-\infty}^{\infty} x(t)\,e^{-j\omega t}\,dt

SymbolDescriptionUnit
X(j\omega)Frequency-domain representationV·s
\omegaAngular frequencyrad/s

Worked example

Find FT of x(t) = e^{-2t}u(t).

Given: x(t) = e^{-2t}u(t), a = 2

  1. X(jω) = ∫₀^∞ e^{-2t} e^{-jωt} dt
  2. = ∫₀^∞ e^{-(2+jω)t} dt
  3. = 1/(2+jω)

Answer: X(jω) = 1/(2 + jω)

Inverse Fourier Transform

x(t) = \frac{1}{2\pi}\int_{-\infty}^{\infty} X(j\omega)\,e^{j\omega t}\,d\omega

Worked example

Verify that IFT of X(jω) = 2/(1+jω) gives x(t) = 2e^{-t}u(t).

Given: X(jω) = 2/(1+jω)

  1. Standard pair: 1/(a+jω) ↔ e^{-at}u(t)
  2. 2/(1+jω) = 2 × [1/(1+jω)]
  3. Linearity: x(t) = 2 × e^{-t}u(t)

Answer: x(t) = 2e^{-t}u(t) ✓

FT Convolution Property

y(t) = x(t)*h(t) \xrightarrow{\mathcal{F}} Y(j\omega) = X(j\omega)\cdot H(j\omega)

SymbolDescriptionUnit
H(j\omega)System frequency responsedimensionless

Worked example

System with H(jω) = 1/(1+jω). Input x(t) = e^{-2t}u(t). Find Y(jω).

Given: X(jω) = 1/(2+jω), H(jω) = 1/(1+jω)

  1. Y(jω) = X(jω) × H(jω)
  2. = 1/(2+jω) × 1/(1+jω)
  3. = 1/[(1+jω)(2+jω)]

Answer: Y(jω) = 1/[(1+jω)(2+jω)]

Parseval's Theorem (Energy, CTFT)

E = \int_{-\infty}^{\infty} |x(t)|^2 dt = \frac{1}{2\pi}\int_{-\infty}^{\infty} |X(j\omega)|^2 d\omega

SymbolDescriptionUnit
ESignal energyJ

Worked example

Find energy of x(t) = e^{-3t}u(t).

Given: x(t) = e^{-3t}u(t)

  1. E = ∫₀^∞ (e^{-3t})² dt = ∫₀^∞ e^{-6t} dt
  2. = [-1/6 · e^{-6t}]₀^∞
  3. = 0 - (-1/6) = 1/6

Answer: E = 1/6 J

Laplace Transform

Unilateral Laplace Transform

X(s) = \int_0^{\infty} x(t)\,e^{-st}\,dt

SymbolDescriptionUnit
X(s)Laplace transformV·s
sComplex frequency s = σ + jωrad/s

Worked example

Find L{e^{-at}u(t)}.

Given: x(t) = e^{-at}u(t)

  1. X(s) = ∫₀^∞ e^{-at} e^{-st} dt = ∫₀^∞ e^{-(s+a)t} dt
  2. = [e^{-(s+a)t} / (-(s+a))]₀^∞
  3. = 0 - (-1/(s+a)) = 1/(s+a)

Answer: L{e^{-at}u(t)} = 1/(s+a), ROC: Re(s) > -a

Initial Value Theorem

x(0^+) = \lim_{s \to \infty} s\,X(s)

Worked example

Find x(0+) from X(s) = (s+3)/[(s+1)(s+2)].

Given: X(s) = (s+3)/[(s+1)(s+2)]

  1. x(0+) = lim_{s→∞} s·(s+3)/[(s+1)(s+2)]
  2. = lim_{s→∞} (s²+3s)/(s²+3s+2)
  3. = lim_{s→∞} (1 + 3/s)/(1 + 3/s + 2/s²)
  4. = 1/1 = 1

Answer: x(0+) = 1

Final Value Theorem

x(\infty) = \lim_{s \to 0} s\,X(s)

Worked example

Find x(∞) from X(s) = (s+3)/[(s+1)(s+2)].

Given: X(s) = (s+3)/[(s+1)(s+2)]

  1. Check: all poles (s = -1, s = -2) in LHP — FVT applies
  2. x(∞) = lim_{s→0} s·(s+3)/[(s+1)(s+2)]
  3. = 0·(3)/(1·2) = 0

Answer: x(∞) = 0

Transfer Function

H(s) = \frac{Y(s)}{X(s)} = \frac{b_m s^m + \ldots + b_0}{a_n s^n + \ldots + a_0}

SymbolDescriptionUnit
H(s)Transfer functiondimensionless
Y(s), X(s)Output and input Laplace transformsV·s

Worked example

System: dy/dt + 3y = 2x. Find H(s) and pole location.

Given: ODE: dy/dt + 3y = 2x

  1. Take Laplace: sY(s) + 3Y(s) = 2X(s)
  2. Y(s)(s+3) = 2X(s)
  3. H(s) = Y(s)/X(s) = 2/(s+3)
  4. Pole at s = -3 (LHP, so stable)

Answer: H(s) = 2/(s+3), pole at s = -3

Z-Transform

Z-Transform Definition

X(z) = \sum_{n=-\infty}^{\infty} x[n]\, z^{-n}

SymbolDescriptionUnit
X(z)Z-transformdimensionless
zComplex variabledimensionless

Worked example

Find Z{a^n u[n]}.

Given: x[n] = a^n u[n]

  1. X(z) = Σ_{n=0}^∞ a^n z^{-n} = Σ_{n=0}^∞ (a/z)^n
  2. Geometric series converges for |a/z| < 1, i.e. |z| > |a|
  3. X(z) = 1/(1 - a·z^{-1}) = z/(z-a)

Answer: X(z) = z/(z-a), ROC: |z| > |a|

Z-Transform Time Shifting

x[n-k] \xrightarrow{\mathcal{Z}} z^{-k}X(z)

SymbolDescriptionUnit
kDelay in samplessamples

Worked example

If X(z) = z/(z-0.5), find Z{x[n-2]}.

Given: X(z) = z/(z-0.5), delay k = 2

  1. Z{x[n-2]} = z^{-2} × X(z)
  2. = z^{-2} × z/(z-0.5)
  3. = z^{-1} / (z-0.5)
  4. = 1 / [z(z-0.5)]

Answer: Z{x[n-2]} = 1/[z(z-0.5)]

Inverse Z-Transform (Partial Fractions)

X(z)/z = \sum_k \frac{A_k}{z - p_k} \Rightarrow x[n] = \sum_k A_k p_k^n u[n]

SymbolDescriptionUnit
p_kPole of X(z)dimensionless
A_kResidue at pole kdimensionless

Worked example

Find inverse Z-transform of X(z) = z/[(z-1)(z-0.5)], ROC: |z|>1.

Given: X(z) = z/[(z-1)(z-0.5)]

  1. X(z)/z = 1/[(z-1)(z-0.5)]
  2. Partial fractions: A/(z-1) + B/(z-0.5)
  3. A = 1/(1-0.5) = 2; B = 1/(0.5-1) = -2
  4. X(z) = 2z/(z-1) - 2z/(z-0.5)
  5. x[n] = 2(1)^n u[n] - 2(0.5)^n u[n]

Answer: x[n] = [2 - 2(0.5)^n] u[n]

Sampling and Reconstruction

Nyquist Sampling Theorem

f_s \geq 2 f_{max} = 2B

SymbolDescriptionUnit
f_sSampling frequencyHz
f_{max}Maximum signal frequency (bandwidth B)Hz

Worked example

Find minimum sampling rate for a signal with bandwidth 4 kHz.

Given: f_max = 4 kHz

  1. f_s(min) = 2 × f_max
  2. = 2 × 4000
  3. = 8000 Hz

Answer: Minimum sampling rate = 8 kHz

Aliasing Condition

f_{alias} = |f_s \cdot k - f|, \quad k \in \mathbb{Z}

SymbolDescriptionUnit
f_{alias}Alias frequencyHz
fOriginal frequencyHz
kIntegerdimensionless

Worked example

A 1.5 kHz sinusoid is sampled at 2 kHz. Find the alias frequency.

Given: f = 1500 Hz, f_s = 2000 Hz

  1. Alias = |f_s - f| = |2000 - 1500|
  2. = 500 Hz

Answer: Alias frequency = 500 Hz

Convolution and Correlation

Continuous-Time Convolution

y(t) = x(t) * h(t) = \int_{-\infty}^{\infty} x(\tau)\,h(t-\tau)\,d\tau

SymbolDescriptionUnit
h(t)Impulse response of LTI system1/s or dimensionless
y(t)Output signalV

Worked example

Find y(t) = u(t) * u(t) for t ≥ 0.

Given: x(t) = u(t), h(t) = u(t)

  1. y(t) = ∫₀^t 1·1 dτ (both u(τ) and u(t-τ) are 1 for 0 ≤ τ ≤ t)
  2. y(t) = ∫₀^t dτ = t for t ≥ 0

Answer: y(t) = t·u(t) (ramp function)

Cross-Correlation

R_{xy}(\tau) = \int_{-\infty}^{\infty} x(t)\,y(t+\tau)\,dt

SymbolDescriptionUnit
R_{xy}Cross-correlation functionsignal units squared × s
\tauTime lags

Worked example

Find R_xx(0) for x(t) = e^{-2t}u(t). What does this equal?

Given: x(t) = e^{-2t}u(t)

  1. R_xx(0) = ∫_{-∞}^{∞} x(t)·x(t) dt = ∫₀^∞ e^{-4t} dt
  2. = [-1/4 · e^{-4t}]₀^∞ = 1/4

Answer: R_xx(0) = 1/4 = signal energy (auto-correlation at zero lag equals energy)

Discrete-Time Convolution

y[n] = x[n] * h[n] = \sum_{k=-\infty}^{\infty} x[k]\,h[n-k]

Worked example

Convolve x[n] = {1,2,3} with h[n] = {1,1}.

Given: x = [1,2,3], h = [1,1]

  1. y[0] = x[0]·h[0] = 1·1 = 1
  2. y[1] = x[0]·h[1] + x[1]·h[0] = 1·1 + 2·1 = 3
  3. y[2] = x[1]·h[1] + x[2]·h[0] = 2·1 + 3·1 = 5
  4. y[3] = x[2]·h[1] = 3·1 = 3

Answer: y[n] = {1, 3, 5, 3}

Quick reference

FormulaExpression
Fourier TransformX(jω) = ∫ x(t) e^{-jωt} dt
Inverse FTx(t) = (1/2π) ∫ X(jω) e^{jωt} dω
Parseval (Energy)E = ∫|x(t)|² dt = (1/2π) ∫|X(jω)|² dω
Parseval (Power, FS)P = (1/T) ∫|x(t)|² dt = Σ|c_n|²
Convolution (CT)y(t) = ∫ x(τ) h(t-τ) dτ
Convolution (DT)y[n] = Σ x[k] h[n-k]
Laplace TransformX(s) = ∫₀^∞ x(t) e^{-st} dt
Initial Value Theoremx(0+) = lim_{s→∞} sX(s)
Final Value Theoremx(∞) = lim_{s→0} sX(s)
Z-Transform of a^n u[n]X(z) = z/(z-a), |z|>|a|
Z Time Shiftx[n-k] → z^{-k} X(z)
Nyquist Ratef_s ≥ 2 f_max
Alias Frequencyf_alias = |f_s·k - f|
FT Convolution Propertyx(t)*h(t) ↔ X(jω)·H(jω)
Transfer FunctionH(s) = Y(s)/X(s)

Exam tips

  • GATE frequently uses the duality property of FT — if x(t) ↔ X(jω), then X(jt) ↔ 2π x(-ω) — verify the sign convention in the question before applying.
  • For ROC questions on Z-transforms, right-sided sequences have ROC of the form |z| > r; left-sided sequences have |z| < r; always state the ROC alongside the transform.
  • Final Value Theorem is only valid when all poles of sX(s) are in the open left-half s-plane — applying it when a pole is at s = 0 or in the RHP gives wrong results.
  • In convolution length questions, if x has M samples and h has N samples, the output y has M+N-1 samples — examiners test both the result and the length.
  • GATE energy vs power classification: finite energy signals have zero average power; periodic signals have infinite energy but finite power.
  • For sampling questions, always check whether the question asks for Nyquist rate (2B) or Nyquist interval (1/2B) — they are reciprocals and confusion costs marks.