Communication System Model

The mathematical model of a communication system provides a rigorous framework for analyzing system performance, predicting error rates, and designing optimal systems. The most influential model was proposed by Claude Shannon and Warren Weaver in 1948.

Shannon-Weaver Model

The Shannon-Weaver model is the foundational mathematical model for all modern communication systems:

Mathematical Framework

Source Model

The source generates a message m selected from a set M of possible messages. The information content of a specific message is:

I(mᵢ) = -log₂[P(mᵢ)] bits

The average information (entropy):

H(M) = -Σ P(mᵢ) × log₂[P(mᵢ)] bits/message

Channel Model

The channel is characterized by its transfer function and noise properties:

**y(t) = h(t) * x(t) + n(t)**

Where:

• y(t) = received signal
• h(t) = channel impulse response
• x(t) = transmitted signal
• n(t) = additive noise
• * denotes convolution

AWGN Channel Model

The simplest and most commonly used channel model:

y(t) = x(t) + n(t)

Where n(t) is white Gaussian noise with:

• Mean: E[n(t)] = 0
• Power spectral density: S_n(f) = N₀/2 W/Hz
• Autocorrelation: R_n(τ) = (N₀/2)δ(τ)

Performance Metrics

Signal-to-Noise Ratio (SNR)

SNR = P_signal / P_noise = S/N

In decibels: SNR(dB) = 10 × log₁₀(S/N)

Channel Capacity (Shannon's Theorem)

C = B × log₂(1 + SNR) bits/second

Where:

• C = channel capacity (maximum error-free data rate)
• B = channel bandwidth in Hz
• SNR = signal-to-noise ratio (linear)

Bit Error Rate (BER)

For BPSK in AWGN:

BER = Q(√(2E_b/N₀))

Where:

• E_b = energy per bit
• N₀ = noise power spectral density
• Q(x) = complementary Gaussian distribution function

Types of Communication System Models

Signal Space Representation

Communication signals can be represented as vectors in signal space. For M signals, each can be expressed as:

sᵢ(t) = Σ sᵢⱼ × φⱼ(t) for j = 1 to N

Where φⱼ(t) are orthonormal basis functions satisfying:

∫φᵢ(t)·φⱼ(t)dt = δᵢⱼ (1 if i=j, 0 otherwise)

System Design Trade-offs

The three fundamental resources in communication system design are:

1. Bandwidth — spectral occupancy
2. Power — transmitted energy
3. Complexity — system cost and processing

Solved Example

Problem: A binary communication system operates over an AWGN channel with bandwidth 10 kHz and received SNR = 20 dB. Determine: (a) Channel capacity, (b) Maximum achievable bit rate with error-free transmission.

Solution:

(a) Channel Capacity:

SNR(linear) = 10^(20/10) = 100

C = B × log₂(1 + SNR) C = 10,000 × log₂(1 + 100) C = 10,000 × log₂(101) C = 10,000 × 6.658 C = 66,580 bps ≈ 66.58 kbps

(b) Maximum error-free bit rate = Channel capacity = 66.58 kbps

Any data rate R < C can achieve arbitrarily low error probability with appropriate coding. Any rate R > C will inevitably have errors.

Interview Questions

Q1: State Shannon's channel capacity theorem and its significance.

Shannon's theorem states C = B×log₂(1+SNR). It establishes the theoretical maximum rate for error-free communication. Its significance is that it provides an upper bound for system design — we know the absolute best performance achievable, allowing engineers to benchmark practical systems against the theoretical limit.

Q2: What is the difference between the Shannon model and the OSI model?

The Shannon model is a physical layer abstraction focusing on signal transmission, noise, and information theory. The OSI model is a networking architecture with 7 layers covering application to physical layer protocols. Shannon's model addresses the fundamental limits of point-to-point communication, while OSI defines the entire network protocol stack.

Q3: How does doubling the bandwidth affect channel capacity compared to doubling the SNR?

Doubling bandwidth doubles capacity linearly (C = 2B×log₂(1+SNR)). Doubling SNR increases capacity logarithmically (C = B×log₂(1+2SNR)). Therefore, bandwidth is a more valuable resource — doubling bandwidth always provides greater capacity increase than doubling SNR.

Q4: What assumptions does the AWGN model make, and when does it fail?

AWGN assumes: additive noise (independent of signal), white spectrum (flat PSD), Gaussian distribution, and no channel distortion. It fails in: wireless fading channels, impulsive noise environments (power lines), interference-limited systems, and channels with significant multipath or nonlinear distortion.