Research On Reconstruction Method Of Linear Scrambler

Linear scrambler is the key technique of modern digital communication, it is mainly used in transmission data’s randomizing processing. The reconstruction of linear scrambler has vital importance on correct obtainment of the transmission message, and so this probelm has been deeply studied by this thesis.There are two types of linear scrambler:synchronous scrambler and self synchronous scrambler. First this thesis raise that the reconstruction of synchronous scrambler contains the recognition of generating polynomial and the initial states of LFSR, and the reconstruction of self synchronous scrambler is the recognition of generating polynomial. This thesis also come up that the premise of the reconstruction of linear scrambler is the source sequence is biased.Secondly, the reconstruction of synchronous scrambler is studied by this thesis. For the recognitiong of generating polynomial, three algorithms are studied: Walsh-Hadamard algorithm based on algebra, Cluzeau algorithm and Probability Distribution Distance algorithm which are based on statistics. For the reconstruction of LFSR’s initial states, a fast correlation attacks theory on stream ciphers is studied by this thesis, Walsh-Hadamard transform is used to resolve the error equations of LFSR’s initial states, then the initial states of LFSR is reconstructed. Simulation and contrastive analysis verifies the validity of all the algorithms above on certain conditions.Finally, the recognition of self synchronous scrambler’s generating polynomial is studied by this thesis. In order to solve the deficiencies of existing algorithms, including the additional priori knowledge, not ideal false-alarm probability and recognition performance, the autocorrelation function of self synchronous scrambler is deeply studied by this thesis. The difference of values of autocorrelation function whose generating polynomials are binomial or trinomial is founded. Then a new recogniton algorithm of self synchronous scrambler’s generating polynomial is raised by the difference. Simulation verifies the validity of the algorithm and contrastive analysis indicates that this algorithm is more effective than any other algorithms.