Blind Classification of Linear Block Codes Based on Multi-Dimensional Features

Linear Block Codes (LBCs) classification is a critical component of non-cooperative communication. Existing researches mainly rely on semi-blind classification techniques based on predefined candidate sets, while few research directed towards blind classification of new scheme LBCs widely used in modern communication systems. Therefore, this paper concentrates on classifying new scheme LBCs, including Polar code, Low Density Parity Check (LDPC) code, Bose-Chaudhuri-Hocquenghem (BCH) code, and Reed-Solomon (RS) code. Addressing the existing issues, this paper proposes an innovative method to classify key LBCs types based on multi-dimensional codeword features. To achieve better classification of new scheme LBCs types in error scenarios, a classification method based on multi-dimensional codeword features is proposed. This method utilizes four key features, including code length, frozen bits feature of Polar code, sparse parity check matrix feature of LDPC, and Galois Field Fourier Transform (GFFT) spectrum of RS and BCH codes. And the proposed method is not limited to the specific candidate sets. The experimental results have verified the effectiveness and practicality of the proposed method.