Automatic code and interleaver classification technique for future generation communication systems

Abstract

Error correcting codes (ECCs) and interleavers play a vital role in ensuring reliable and efficient communication over noisy channels. The process of manually selecting appropriate ECCs and interleavers for a given communication system can be time-consuming and error-prone. In this thesis, we propose an automated approach to identify suitable ECCs and interleavers using deep learning techniques. We begin by exploring various ECCs commonly used in digital communication systems, including block codes, convolutional codes, Low-Density Parity-Check (LDPC) codes, BCH codes, and Hamming codes. Each ECC exhibits different error correction capabilities, decoding complexity, and performance characteristics. By training a deep learning model on a diverse dataset of encoded and noisy communication signals, we aim to develop a system that can accurately classify and recommend the most suitable ECC for a given scenario. Additionally, we investigate different types of interleavers, such as block interleavers, convolutional interleavers, and helical interleavers. Interleavers aid in spreading burst errors, reducing error propagation, and improving overall error correction performance. Leveraging the power of deep learning, we seek to build a model that can automatically determine the optimal interleaver based on the specific requirements of a communication system.