Scribd
Upload
6 views8 pages

Signal Space Representation

representación espacio señal

Uploaded by

Javier Vía
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views8 pages

Signal Space Representation

representación espacio señal

Uploaded by

Javier Vía
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

Signal Space Representation

Procesado de Señal en Comunicaciones Inalámbricas

Curso 2023-2024
Contents

Geometric Representation of Signals

Optimal Receiver in AWGN

Analysis and Simulation of Communication Signals


Signal Space

▶ Internal product: ⟨s1 (t), s2 (t)⟩ = ab s1(t)s2∗ (t)dt


R

▶ Orthogonality: s1 (t) ⊥ s2 (t) ⇔ ⟨s1 (t), s2 (t)⟩ = 0


√ R 1/2
▶ Norm ( Energy): ∥s(t)∥ = ⟨s(t), s(t)⟩ = ab |s(t)|2 dt
p

▶ Orthonormal Basis: {Ψ1 (t), Ψ2 (t), . . . , ΨN (t)} orthonormal iff


▶ Ortho: ⟨Ψi (t), Ψj (t)⟩ = 0, i ̸= j
▶ Normal: ∥Ψi (t)∥ = 1, ∀i
▶ Representation: Signal as a vector in N-dimensional space

N
X
s(t) = ai Ψi (t) s(t) ≡ [a1 , a2 , . . . , aN ]T
i=1
Orthonormal Basis

▶ Given a set of M signals: {s1 (t), s2 (t), . . . , sM (t)}, they can be


represented by means of an orthonormal basis
{Ψ1 (t), Ψ2 (t), . . . , ΨN (t)} with dimension N ≤ M
▶ We can now analyze, study and manipulate the signals in a
N-dimensional vector space
N
X
sk (t) = ak,i Ψi (t) ≡ [ak,1 , ak,2 , . . . ak,N ]T , k = 1, . . . , M
i=1
Optimal receiver for M symbols

▶ M equiprobable signals (symbols) in AWGN: {s1 (t), s2 (t), . . . , sM (t)}


▶ Orthonormal basis {Ψ1 (t), Ψ2 (t), . . . , ΨN (t)} with N ≤ M
▶ Receiver based on N correlators (matched filters for Ψi (t))
▶ Decisor: Minimum Distance Criterion: iˆ = arg mink (∥sk − z∥)
▶ Symbol representation: si = [ai,1 , ai,2 , . . . , ai,N ]T
▶ Noise representation n = [n1 , n2 , . . . , nN ]T
▶ Received Signal representation: z = [z1 , z2 , . . . , zN ]T = si + n
Receiver Analysis

▶ Received signal: z = si + n
▶ n is a Gaussian random vector N (0, σ 2 I)
▶ Independent components with zero mean and variance σ 2 = N0 /2
▶ If the symbol si is transmitted: z|si ∼ N (si , σ 2 I)
Optimal Decoding

▶ Given the received signal (z = si + n), we need to determine which


symbol was transmitted
▶ This is a classification problem:
▶ Minimum error criterion (MAP): P(si |z) = f (z|si )P(si )
f (z)
▶ For equiprobable symbols reduces to maximizing f (z|si ) (ML)
▶ In the AWGN case, it further reduces to the minimum distance
criterion: iˆ = arg mink (∥sk − z∥)
▶ The probability of error will depend on:
▶ Distance among symbols: di,j = ∥si − sj ∥
▶ Noise variance: σ 2 = N0 /2
▶ Mean Bit Energy: Eb = Es / log2 (M)
▶ Symbol Energy: Esi = ∥si ∥2
▶ Finally: Pe = g (d, σ) = h(Eb , N0 )
Analysis and Simulation of Communication Signals

▶ Complete System:
▶ Includes Modulation, Amplifiers, Filters, Channel, Demodulation . . .
▶ Sampling Frequency: fs ≥ 2fmax : Impractical in most of the cases
▶ Low-Pass Equivalent:
▶ Analysis of the baseband signals
▶ The effect of channel and TX/RX parts is modeled by their low-pass
equivalent
▶ Sampling Frequency moderate: In practice fs = Rs L with L > 2
▶ Equivalent Discrete System:
▶ Analysis in Signal Space
▶ Sampling Frequency: fs = Rs One sample per symbol

Consider WiFi: fc = 2.4GHz, W = 20MHz

You might also like