Parameter Identification for Recursive Systematic Convolutional Codes Using Hybrid Polak-Ribière-Polyak and Hestenes-Stiefel Conjugate Gradient

In non-cooperative communications, receivers lack prior channel coding knowledge, making error correction code recognition critical. Recursive systematic convolutional (RSC) codes, core to Turbo codes, suffer from low-SNR generator matrix identification due to existing algorithms’ inadequate noise resilience. This work enhances RSC recognition accuracy by formulating an exponential cost function via RSC constraint relationships. To overcome local optima and slow convergence in optimization, an improved the Polak-Ribière-Polyak and Hestenes-Stiefel (PRP-HS) conjugate gradient algorithm is proposed. Simulations confirm faster convergence and markedly higher recognition accuracy under low SNR versus benchmark methods, significantly boosting noise robustness.