Introduction to Coding Theory

Overview

Error control coding is an indispensable part of any digital communication system. In this introductory course, we will discuss theory of linear block codes and convolutional codes, their encoding and decoding techniques as well as their applications in real world scenarios. Starting from simple repetition codes, we will discuss among other codes: Hamming codes, Reed Muller codes, low density parity check codes, and turbo codes. We will also study how from simple codes by concatenation we can build more powerful error correcting codes.

PREREQUISITES: An exposure to linear algebra and probability theory as well as a course in digital communications

Syllabus

Week 1-

Lecture 1: Introduction to error control coding Lecture 2: Introduction to linear block codes, generator matrix and parity check matrix Lecture 3: Properties of linear block codes: Syndrome, error detection

Week 2-

Lecture 4: Decoding of linear block codes Lecture 5: Distance properties of linear block codes

Week 3-

Lecture 6: Some simple linear block codes: Repetition codes, Single parity check codes, Hamming codes, Reed Muller codes Lecture 7: Bounds on size of codes: Hamming bound, Singleton bound, Plotkin bound, Gilbert-Varshamov bound

Week 4-

Lecture 8: Introduction to convolutional codes-I: Encoding, state diagram, trellis diagram Lecture 9: Introduction to convolutional codes-II: Classification, realization, distance properties Lecture 10: Decoding of convolutional codes-I: Viterbi algorithm

Week 5-

Lecture 11: Decoding of convolutional codes-II: BCJR algorithm Lecture 12: Performance bounds for convolutional codes

Week 6-

Lecture 13: Low density parity check codes Lecture 14: Decoding of low density parity check codes: Belief propagation algorithm on BSC and AWGN channels

Week 7-

Lecture 15: Turbo codes Lecture 16: Turbo decoding

Week 8-

Lecture 17: Distance properties of turbo codes Lecture 18: Convergence of turbo codes Lecture 19: Automatic repeat request schemes Lecture 20: Applications of linear codes