The Research On The Blind Recognition For Turbo Coding Parameters

The channel coding blind recognition technology can analyze the coding type and coding parameters through various technical means based on the bit stream of intercepted data without any or only partial prior knowledge,thereby providing decoding and obtaining subsequent information.Based on this,the technology has important military significance and practical value.This paper mainly focuses on the blind recognition of coding parameters of Turbo codes.Turbo codes are a type of code that is approaching the Shannon limit.Since its introduction,it has received a lot attention and has been applied to various communication systems.Due to the special structure of turbo coder,its parameter identification is very difficult.At present,there are few researches related to the blind recognition of Turbo codes.The related results mainly studied the coding parameters of non-punctured Turbo codes under low bit error rate condition.However,for the recognition of parameters under high bit error rate condition and recognition of punctured Turbo codes,the recognition effect is not ideal,therefore this paper make some improvements for these deficiencies.Firstly,aiming at solving the problem of high complexity when recognize the recursive systematic convolutional(RSC)under high bit error rate,we propose a new iterative algorithm.Based on the linear relationship between the codewords in the analysis matrix,the iterative algorithm extracts features of rows which contain error code in the matrix,and then remove these rows.After several iterations,we obtain the analysis matrix with low bit error rate,furthermore,the parity check vector is solved and is converted into the form of generator matrix,then the parameter recognition of the RSC encoder is completed.This algorithm avoids the problem that the existing method needs to solve the satisfaction of all the parity check vectors,which greatly reduces the complexity.Second,for the recognition of the length of interleaving and the frame synchronization,the traditional methods mostly use the rank criterion method or Gauss transform analysis method,the rank-criterion method has poor error-resistance performance,and the Gauss transform analysis method requires a large number of codewords and make statical analysis through the feature of weight.The Gauss transform analysis method still can not solve the recognition problem under high bit error rate condition,and requires the reception of long enough sequences.Unlike the methodmentioned above,this paper decomposes the parity check vector of the RSC encoder and uses a linear combination search method to solve the length of interleaving and the frame synchronization,and optimizes it by using the Chose-Joux-Mitton algorithm,which make a good error-resistance performance.Finally,for the recognition of interleaving mappings,the existing algorithms mainly recognize interleaving mappings by recovering interleaved sequences.The method is poor in bit error tolerance.This paper directly recovers the interleaving mappings by solving and comparing the constraint relationship between the preinterleaved sequence and the parity check sequence.And this method is applied to the recognition of the punctured Turbo codes,which also achieved good performance,overcoming the problem that the punctured Turbo codes cannot recover the interleaved sequence and then fails to recognize the interleaving mapping relationship.The simulation results show that the proposed algorithm can realize the blind recognition of coding parameters of punctured and non-punctured Turbo codes under high bit error rate condition.