Communication Reliability Study Of Sequence-based Password Attack

Increasing signal-noise ratio is a general method for receivers to improve information transmission performance in noisy channel. And error-correcting code is also an important method to improve information transmission reliability. The correlation attack for stream cipher is a breaking algorithm in common use. This thesis brings a new decoding algorithm for communication error-correcting codes based on correlation attacks which not only enhances communication reliability but also has high communication security.Correlation attack is to acquire the initial states of linear feedback shift register (LFSR) sequences, and fast correlation attacks algorithm is one of the most effective attack methods for stream cipher. In this thesis, we apply fast correlation attack algorithm B to communication information transmission as a new decoding algorithm. We explore the performance of the new decoding algorithm in BSC channel and AWGN channel, and find that the new decoding algorithm shows good performance in low signale to noise ratio. When the length of information bit is 31, the longer the code length the better the performance of fast correlation attack algorithm B. When the length of code is more than 5000, it can decode successfully in BSC channel with error probability of 0.38. When the length of code is more than 4000, it can decode successfullt in AWGN channel with -14dB SNR.This thesis also discusses the performance of the new decoding algorithm particularly in interference conditions. We analyze the anti-jamming capability of the new decoding algorithm and compare it with direct sequence spread spectrum (DSSS). Although it is found that the performance of anti-jamming technology base on FCAA is not better than DSSS in total, but communication based on FCAA is safer than DSSS. Once enemies capture the spread spectrum codes, the spread spectrum information is exposed. While in communication based on FCAA, even if transmitted sequences are acquired, it is difficult to recover transmitted information (the initial state of the LFSR) without knowledge of configuration of LFSR.In the end, this thesis applies the new decoding algorithm to acquisiton the long PN sequence, and brings a new acquisition method based on FCAA. Simulation results show that the performance of FCAA search is better than that of iterative message passing algorithms (iMPAs) and closer to that of parallel search and serial search as the capture sequences are longer. When acquisition length is 256, the performance of new method based on FCAA with 20 iteratives is a little worse than that of parallel search, serial search and iMPAs. When acquisition length is 512, the performance of new method is closer to parallel search and serial search and better than iMPAs.This new method is significant in complexity reduction comparing with parallel search and serial search. Furthermore this approach needs not to extract the cyclic graphical models, so that it can be applied to linear and nonlinear sequence acquisition.