| dc.description.abstract | In this thesis, a lower bound between non-orthogonal coherent states and mean photon number in quantum noise randomized stream cipher (Y-00) for a given measurement error probability is proposed and compared against other measurement schemes. In this analysis, recently discovered maximum likelihood positive operator valued measure (ML-POVM) approach in a multiphoton regime is used to provide more accurate and optimum results than greedy scheme, quantum unambiguous measurement (QUM), and random guessing for which we have considered success probability of coherent state detection as a figure of merit. Moreover an analysis about the impact of erroneous output sequence of a pseudo random number generator (PRNG) in predicting the running seed key is studied.
In general, Y-00 scheme utilizes an initial shared secret key between legitimate users for which users experience superior receiver performance than does the intruder who does not know the key. An intruder suffers unavoidable quantum noise while probing the communication between legitimate users, owing to the user’s ignorance of the secret key. In particular, an indefinite bound was proposed earlier between the number of non-orthogonal coherent states and the mean photon number in Y-00 scheme. In this research work, a lower bound is proposed using ML-POVM, where ML-POVM provides better probability of detection of a given number of coherent states and mean photon number than other measurement techniques can detect.
Finally, a simulation of linear feedback shift register (LFSR) is carried out as an example of PRNG for various number of bit-flip errors in the output sequence of LFSR to analyze the impact of erroneous output sequence in predicting the running seed key of LFSR, which demonstrates that a significant number of bit-flip errors is required to make the seed key indistinguishable from the observation of the output sequence of LFSR. | en_US |
