Error detection and correction using Viterbi decoding and path-metric

Task undertaken as this project is to write a MATLAB code for convolution (2, 1, 9) decoding with g1 = 110111 and g2 = 111011 and perform analysis of percentage error detection and correction using Viterbi Algorithm and path-metric. [Q. B-48]

Input:

• A 4-bit dataword (1 1 0 1) is chosen as an input to the Convolutional Encoder and corresponding codeword is obtained. The length of the codeword obtained is 8.

Algorithm:

• All the possible 1-bit errors (8 possibilities), 2-bit errors (28 possibilities) and 3-bit errors (56 possibilities) are introduced.
• Each resulting codeword is then decoded using the Viterbi Algorithm (dataword with minimum path-metric is chosen).
• The output dataword is then compared to the correct dataword.
  • If (path-metric != 0 and obtained dataword == correct dataword), error is detected and corrected.
  • If (path-metric != 0 and obtained dataword != correct dataword), error is detected but not corrected.
  • If (path-metric == 0 and obtained dataword == correct dataword), no error occurred.

Output:

• The program calculates the Percentage Error Correction and Detection for all possible 1, 2 and 3-bit errors.
• A graph is plotted to analyse the Percentage Error Correction and Detection visually.

How to run the program?

• The project contains two files:
  • main.m
  • decoding.m (A function called for decoding the codeword)
• While the decoding.m is open, run main.m file on MATLAB to obtain the results.

Contributors:

---

• Namrata Ladda (16CO121)
• Aditi Gupta (16CO202)