Blind recognition of BCH and RS codes with interleaver parameter estimation: A deep learning and algebraic framework with hardware validation

Abstract

Blind estimation of code and interleaver parameters plays a vital role in various applications such as adaptive modulation and encoding, non-cooperative systems, reconfigurable radio systems, signal intelligence, etc. In this paper, we have proposed blind recognition algorithms for Bose-Chaudhuri-Hocquenghem (BCH) and Reed-Solomon (RS) codes, along with block and convolutional interleavers. Initially, deep learning frameworks are employed on the received data to classify the channel coding scheme and interleaver type. Further, we propose novel algorithms based on algebraic framework for recognition of code and interleaver parameters. Simulation results are provided for various test cases considering noisy channel conditions. Finally, the accuracy of classification and probability of correct recognition of code and interleaver parameters are given with detailed inferences. We have validated our findings by comparing the simulation results with those obtained using a USRP-based software-defined radio (SDR) setup, demonstrating that the real-time results closely match the simulation outcomes. Finally, the proposed algorithms outperform the existing methods for the joint recognition of code and interleaver parameters reported in the literature. © 2026 Elsevier B.V.