Sequential-Decoder

Sequential Decoder for Convolutional Codes (Fano's Algorithm) in Matlab for Data Communication and Networking

Problem Statement

Q42. MATLAB code for convolution code (2, 1, 11) decoding with g1 = 110111001 and g2 = 111011001 and the analysis of % of error detection and correction using sequential decoding with threshold upto 5.

Contributors

• Praveen Kumar Gupta - 16CO235
• Shashank P - 16CO247

Usage

• In main.m , update generating functions, number of memory units, bit length to be passed to encoder, and maximum error bits for error analysis.

• Run all the scripts here once to load the functions to the matlab console.

• Run the main module by entering main in the matlab console.

• Note that Error Analysis may take some time to generate all cases and compile the errors based on the compute capabilities required for the given case.

Example Usage

Parameters

g1 = [1 1 0 1 1 1 0 0 1];
g2 = [1 1 1 0 1 1 0 0 1];
threshold = 5;
memory_bits = 11;
input_bits = 4;
max_errors = 6;

Modules

1. main.m : Module to enter the required parameters for error analysis.
2. error_percentage.m : Returns the percentage of error detected and corrected with the given parameters.
3. encode.m : Returns the encoded codeword for a given input data stream.
4. analyze_kerror.m : Analyze errors by introducing errors and decoding for k-bits in the encoded codeword.
5. decode.m : Recursively decodes the received codeword (which may contain errors) using sequential decoding.
6. getNextState.m : Returns the next state and 2-bit convolutional output during state transition.
7. conv_2bit.m : Calculates the 2 bit convolutional output during state transition.

Limitations

For a n-bit input to be encoded by the encoder, the modules generate all possible codewords by the convolutional encoder with the given generating functions.

Then error are introduced in all possible places in all the encoded codewords from 1,2....max_errors count, and every codeword with error is decoded and compared with the original codeword without errors for analysis of error detection and error correction.

This requires a lot of computation to be done. So error analysis for reasonably sized input and max_errors might take some amount of time for simulation.