Channel coding recognition algorithm based on complementary fusion of explicit and implicit features
The existing methods for channel coding identification mainly fall into two categories:one is the explicit coding feature extraction method based on coding models,and the other is the implicit coding feature ex-traction method based on deep learning.The explicit method suffers from poor adaptability and difficulty in deter-mining classification boundaries,while the implicit method has issues such as high data requirements,weak trans-ferability,and poor interpretability.To address these problems,a channel coding identification algorithm based on the complementary fusion of explicit and implicit features is proposed.Firstly,multidimensional explicit cod-ing features are extracted based on the channel coding model,while implicit coding features are extracted from IQ waveforms using a deep neural network.Then,an attention mechanism is employed to complementarily fuse these two types of features,and the fused features are used for the final channel coding identification.Simulations indicate that this method effectively addresses the shortcomings of both explicit and implicit feature extraction identification methods and demonstrates good performance in channel coding recognition.
万方数据
维普
