Cơ bản về hệ thống truyền tin và mã hóa

Fundamentals of Communication Systems and Coding

Welcome to this comprehensive course on the basic concepts of communication systems and coding. Designed for students of computer science and networking, the material covers signal conversion, sampling theory, information measures, classic cryptographic techniques, and linear block codes. Each section expands on the quiz questions you may have encountered, providing deeper explanations, practical examples, and useful mnemonics to aid retention.

1. From Analog to Digital: The Role of Discretization

In any communication system, the first step is to convert the continuous‑time analog signal into a digital representation that can be processed, stored, and transmitted by electronic devices. This conversion is performed by the discretization (or sampling) block.

• What it does: It takes snapshots of the analog waveform at regular intervals (the sampling period) and rounds each snapshot to the nearest quantization level, producing a sequence of numbers.
• Why it matters: Without discretization, digital hardware would never see the signal, and any subsequent digital processing would be impossible.

Remember the mnemonic: "Lấy mẫu = rời rạc hoá" – sampling equals discretization.

2. Shannon’s Sampling Theorem: Minimum Sampling Frequency

Claude Shannon proved that a band‑limited signal can be perfectly reconstructed from its samples if the sampling frequency f_s satisfies:

f_s ≥ 2·B

where B is the highest frequency component present in the signal. This is often called the Nyquist rate.

• Example: Human speech typically contains frequencies up to 4 kHz. Applying the theorem, the minimum sampling rate is 2 × 4 kHz = 8 kHz.
• Practical note: Telephony standards (e.g., PCM) adopt an 8 kHz sampling rate for exactly this reason.

Mnemonic: "Gấp đôi là chìa khóa" – double the highest frequency.

3. Information Quantity: Self‑Information and Entropy

Information theory quantifies the surprise associated with an event. The self‑information of an outcome x with probability p(x) is defined as:

I(x) = -log₂ p(x)

When p = 1 (the event is certain), the logarithm becomes zero, so I(x) = 0. No surprise, no information.

Extending this to a random source X, the average information – called entropy H(X) – is:

H(X) = - Σ p(x) log₂ p(x)

• If the source always emits the same symbol (probability 1), the entropy is 0 bits. There is no uncertainty.
• If the source were perfectly uniform over m symbols, the entropy would be log₂ m bits.

Mnemonic: "Không bất định → entropy 0".

4. Mutual Information and Statistical Independence

Mutual information measures the amount of information shared between two random variables X and Y:

I(X;Y) = H(X) + H(Y) - H(X,Y)

When X and Y are statistically independent, knowing one gives no insight about the other, so the joint entropy equals the sum of the individual entropies and the mutual information collapses to 0 bits.

Mnemonic: "Không chia sẻ, không thông tin chung".

5. Classic Substitution Cipher: Caesar Cipher

The Caesar cipher is a simple shift cipher where each plaintext letter P is moved k positions forward in the alphabet (modulo 26). The encryption formula is:

C = (P + k) mod 26

Decryption reverses the shift:

P = (C - k) mod 26

• Example: If ‘A’ becomes ‘D’, the shift is three positions, so k = 3.
• Key takeaway: The cipher’s security relies solely on the secrecy of k; with only 25 possible keys, it is trivially breakable by brute force.

Mnemonic: "A tới D = ba bước".

6. Linear Block Codes: Structure and Parameters

Linear block codes are defined by two parameters (n, k) where:

• k – number of information (message) bits.
• n – total bits after encoding (information + parity).

The generator matrix G is often written as G = [I_k | P], where I_k is the k × k identity matrix and P is the parity matrix. Consequently, the dimensions of P are:

k × (n‑k)

This layout ensures each of the k message bits contributes to n‑k parity bits, providing redundancy for error detection and correction.

Mnemonic: "Hàng = thông tin, cột = parity".

7. Putting It All Together: A Mini‑Project

To reinforce the concepts, consider building a simple voice‑digitization pipeline:

1. Capture an analog speech signal (max frequency ≈ 4 kHz).
2. Sample at 8 kHz (Nyquist rate).
3. Quantize each sample to 8 bits (256 levels), producing a digital stream.
4. Encode the stream using a (7,4) Hamming code. The parity matrix P will be 4 × 3, adding three parity bits per four data bits.
5. Transmit the encoded bits over a noisy channel.
6. Decode on the receiver side, correct any single‑bit errors, and reconstruct the original samples.

This exercise illustrates how sampling theory, information measures, and coding theory interlock in real‑world communication systems.

8. Frequently Asked Questions (SEO‑Optimized)

• What is the purpose of discretization in a communication system? Discretization converts continuous‑time analog signals into a sequence of digital values, enabling processing by computers and digital hardware.
• How do I calculate the minimum sampling rate for a given signal? Use Shannon’s theorem: f_s ≥ 2·B, where B is the highest frequency component of the signal.
• Why does a certain event have zero information? When the probability of an event is 1, the outcome is certain, producing no surprise; thus self‑information I(x) = 0 bits.
• What is the mutual information between independent sources? It is zero because independent sources share no common information.
• How many possible keys exist for a Caesar cipher? There are 25 possible non‑trivial keys (shifts 1‑25), making the cipher insecure against exhaustive search.
• What are the dimensions of the parity matrix in a (n,k) block code? The parity matrix P has dimensions k × (n‑k).

9. Key Takeaways

• Discretization (sampling) is the bridge from analog to digital.
• Shannon’s theorem dictates that the sampling frequency must be at least twice the highest signal frequency.
• Self‑information is zero for certain events; entropy is zero for deterministic sources.
• Mutual information vanishes for independent random variables.
• The Caesar cipher uses a simple modular addition; its key is the shift amount.
• In a linear block code, the parity matrix size is k × (n‑k), reflecting the relationship between information and parity bits.

By mastering these foundational ideas, you are well‑prepared to explore more advanced topics such as error‑correcting codes, modulation schemes, and modern cryptographic algorithms.