Study On Blind Identification Of Convolutional Encoders

Blind identification technology of channel coding refers to the technology of identifying and analyzing channel coding types and parameters in a communication link,which only uses the received sequences and the redundancy of the coded sequences.The blind identification technology of channel coding can be mainly used in two different scenarios:cooperative communication and non-cooperative communication.In cooperative communication scenario,the blind identification technology of channel coding is one of the key technologies of adaptive modulation coding technology.By utilizing the blind identification technology of channel coding,the transmitters may choose the appropriate channel coding according to the current channel conditions,without frequent transmission of control signals to receivers.Thus,the spectrum efficiency can be improved.In non-cooperative communication scenario,the non-cooperative receiver has to utilize blind identification technology to identify the channel encoder from the intercepted signal sequences without any prior information,which lays a foundation for the subsequent channel decoding.Therefore,the blind identification technology of channel coding is significant in practical application.Aiming at the convolutional codes in channel coding,this thesis studies the identifiability and robustness identification methods of convolutional encoders,and the main work is as follows.1.Identifiability and identification methods of convolutional encodersFor the noiseless case,the ambiguity of the existing blind identification methods of convolutional generator matrix are studied to analyze the identifiability of the optimal convolutional encoders at different code rates.Then an identification method for eliminating the the ambiguity of convolutional generator matrix is given by using the auxiliary information.The basic idea is to transform the identification equation of generator matrix into a polynomial congruence equation,and to use a brute-force method to solve the equation.The simulations show that the presented method can blindly identify convolutional generator matrix effectively.2.Robust method for blind identification of convolutional encodersFor the noisy case,a robust method for blind identification of k/n convolutional encoders is presented,and the key behind the presented method lies on bind estimation of parameters.Firstly,the parameter estimation for convolutional encoders is formulated as the problem of solving linear equations over GF(2).Secondly,the Walsh-Hadamard transformation is used to solve these linear equations,and an algorithm for the estimation of the coding parameters of the convolutional encoder is proposed.Moreover,the analysis for the threshold setting in the proposed algorithm is provided.The simulations show that the proposed method outperforms the existing method.Especially,the correct rate of parameter estimation for the proposed method is more than 95%for different encoders when the bit error rate of the channel transmission is less than or equal to 0.07,which demonstrates the practical value of the proposed method.