The heart of modern data transmission.

Key Concepts:

• Modulation: ASK, FSK, PSK, QPSK, QAM (16-QAM, 64-QAM, 256-QAM), OFDM (used in LTE, 5G, Wi-Fi)
• Coding: Source coding, Channel coding, Error detection & correction (Parity, CRC, Hamming, Convolutional codes, Turbo & LDPC)
• Performance Metrics: Bit Error Rate (BER), Throughput, Latency, Spectral efficiency

Why It Matters: Essential for high-speed, reliable digital networks like 4G/5G.

Labs/Practice: Simulated QAM modulation and error correction; analyzed BER curves.

Tools Used: MATLAB, GNU Radio, NS3.

Lesson 3: Digital Communications

This is the heart of modern telecommunications. Almost everything in 4G, 5G, Wi-Fi, satellite links, fiber high-speed links, and even deep-space comms boils down to digital communications principles.

Lesson 1 gave you math and signals; Lesson 2 gave you sampling/filtering/LTI systems. Now we take bits → turn them into waveforms (modulation), protect them against noise/errors (coding), send them over channels, and measure how well they survive.

Why Digital Communications Is “The Heart”

• All modern telecom is digital at baseband: bits rule.
• The channel is noisy, fading, dispersive → we must fight entropy with smart modulation + coding.
• Goal: Maximize throughput (bits/s) while keeping BER low, latency acceptable, and spectral efficiency high (bits/Hz).
• This directly determines why 5G can do 10 Gbps, why Starlink works in LEO, why Wi-Fi 6/7 pushes higher rates.

1. Modulation – Turning Bits into Waveforms

Modulation maps bit groups to changes in carrier signal (amplitude, phase, frequency).

Basic schemes (historical but still used in low-rate links)

• ASK (Amplitude Shift Keying): bits change amplitude → simple but noise-sensitive.
• FSK (Frequency Shift Keying): bits change frequency → robust to amplitude noise (used in Bluetooth classic).
• PSK (Phase Shift Keying): bits change phase → constant envelope, good for power amplifiers.

Advanced / Modern schemes

QPSK (Quadrature PSK)

• 4 symbols (2 bits/symbol).
• Phases: 45°, 135°, 225°, 315° (or ±45° offsets).
• Each symbol = one complex number in I-Q plane.
• Constant amplitude → efficient for non-linear amplifiers.
• Spectral efficiency: 2 bits/Hz (ideal).

QAM (Quadrature Amplitude Modulation)

Combines amplitude + phase.

• 16-QAM: 16 points (4 bits/symbol) → 4×4 grid in I-Q plane.
• 64-QAM: 64 points (6 bits/symbol) → used in good-channel conditions (LTE/5G downlink).
• 256-QAM: 8 bits/symbol → Wi-Fi 6/7, 5G in excellent SNR.

Higher order = more bits/Hz but needs higher SNR (points closer → more errors from noise).

OFDM (Orthogonal Frequency Division Multiplexing) – The king of wideband digital comms

Used in: Wi-Fi (802.11a/g/n/ac/ax), LTE, 5G NR, DAB, DVB-T, power-line comms.

Idea: Split wideband carrier into hundreds/thousands of narrow orthogonal subcarriers.
Each subcarrier carries low-rate data (usually QAM-modulated).
Orthogonality → no interference between subcarriers (even though spectra overlap).

Key features:

• Cyclic Prefix (CP): Copies end of symbol to beginning → absorbs multipath delay spread → turns ISI into simple phase rotation (equalized per subcarrier).
• FFT/IFFT: Efficient implementation.
• Pilot subcarriers: Known symbols for channel estimation.

Why OFDM wins in mobile/wireless: Multipath fading is flat per subcarrier → simple one-tap equalizer per subcarrier.

2. Coding – Adding Intelligence to Fight Errors

Raw modulation is not enough — noise flips bits. Coding adds redundancy.

Error Detection (simple)

• Parity bit: Even/odd count → detects odd number of errors.
• CRC (Cyclic Redundancy Check): Polynomial division → very good detection (used in Ethernet, Wi-Fi, LTE/5G PDCP/RLC).

Error Correction

• Hamming code: Corrects single errors.
• Convolutional codes + Viterbi decoding → good for continuous streams (classic GSM, CDMA, early Wi-Fi).

• Turbo codes: Parallel concatenated convolutional codes + iterative decoding → approach Shannon limit (3G UMTS).
• LDPC (Low-Density Parity-Check): Sparse parity-check matrix → iterative belief propagation. Used in: 5G data channel, Wi-Fi 6/7, DVB-S2, 10GBASE-T. Very close to Shannon limit.

Modern trend: 5G uses LDPC for data, Polar codes for control channels.

3. Performance Metrics – How We Judge Success

• Bit Error Rate (BER): Probability a bit is wrong after decoding. Target: 10⁻⁵ to 10⁻¹² for data.
• Throughput: Effective bits/s after overhead/coding/retransmissions.
• Latency: End-to-end delay (important for URLLC in 5G).
• Spectral efficiency: bits/s/Hz — higher = better use of bandwidth (Shannon limit: log₂(1+SNR)).

Observe:

• QPSK robust (works at low SNR) but low efficiency.
• 64-QAM/256-QAM need high SNR but pack many bits/Hz.
• Coding shifts curves left (coding gain).

How to Study & Practice This Lesson

Resources:

• “Digital Communications” by Proakis
• “Fundamentals of Wireless Communication” by Tse & Viswanath (free PDF online)
• YouTube: “IIT Bombay Digital Communication” or “Wireless Future Blog” series

Hands-on (crucial):

• Python/MATLAB: Generate QPSK/16-QAM symbols, add AWGN, demodulate → plot constellation scatter.
• Compute BER vs SNR curve (Monte Carlo simulation).
• Simple OFDM: IFFT → add CP → FFT → recover.

Milestone Questions (Portfolio Showcase)

• Why does higher-order QAM need higher SNR?
• How does cyclic prefix fix multipath in OFDM?
• Sketch QPSK constellation and explain decision regions.
• Why LDPC/Turbo beat convolutional in modern systems?
• Interpret a BER waterfall: what does “coding gain” mean?

Master this → 4G/5G architectures, RF impairments, and security attacks will make sense.