High-speed Quantum Random Number Generator Based On Laser Phase Fluctuations

The development of the prototype of small-scale integrated high speed quantum random number generator is to meet the requirement of technology and industrial field for random number with high quality and high generating speed including modern secure communication protocol such as quantum key distribution.The real time generating speed of quantum number of the generator reaches to 3.2 Gbps,and can pass NIST randomness test,which conforms to the basic characteristics of true random number.We compare different random number generating projects in the paper,and finally choose the project based on laser phase fluctuations.We theoretically analyze the feasibility of the project by making models,and ascertain the optimal operating point by conducting experiments.In this project,laser phase fluctuations work as quantum signal source,whose randomness originates from spontaneous radiation of laser.Interferometer can change immeasurable phase fluctuations into measurable laser intensity fluctuation,which converting into random fluctuated electric signal by getting through high-speed photoelectric detector,then original random quantum numbers can be collected.In addition,memory modules of test purpose,optical fibers,billibit ethernet as well as data transmission interfaces such as USB 2.0 have also been designed to meet different application requirements,and the effective output rate of random number by getting through optical fibers is 3.2 Gbps.In the process of developing high-speed quantum random number generator based on laser phase fluctuations,the processing speed of information-theoretically provable post-processing for original random number is slow.To deal with the problem,we come up with a post-processing arithmetic applied in FPGA,whose real time postprocessing speed reaches 3.36 Gbps,which breaks the technology bottleneck.Under the adverse impact of factors such as classical noise and undesirable devise,original random numbers are not ideal true random numbers in most cases.Random numbers conforming to true random number characteristics can only be obtained by applying post-processing.However,the existing information-theoretically provable post-processing has large calculated amount which results in low processing speed,so it can only conducts off-line and non-real time and low speed processing,which limits the generating speed of quantum random number in real time.Focusing on this problem,we put forward a parallel algorithm based on matrix decomposition by studying the calculating characteristics of post-processing based on large scale Toeplitz matrix,which remarkably reduces hard-ware logic resource needed of the algorithm,and successfully realizing high-speed post-processing calculation based on large scale Toeplitz matrix in FPGA,and the speed of real-time post-processing reaches 3.36 Gbps.We conduct randomness tests such as statistical homogeneity test,autocorrelation test and NIST test on post-processed quantum random number,and the test results show that the random numbers conform to the basic characteristics of true random number which verifies the effectiveness of the post-processing algorithm.The high speed post-processing parallel algorithm in FPGA solve the mismatching problem between the speed of post-processing and the speed of generating original random number,enabling the quantum number generator based on laser phase fluctuations producing true high speed quantum random numbers.The main innovation point and innovation result in the paper lie in:1)The development of the prototype of small-scale integrated high speed quantum random number generator is to meet the requirement of technology and industrial field for random number with high quality and high generating speed including modern secure communication protocol such as quantum key distribution.The real time generating speed of quantum number of the generator reaches to 3.2 Gbps,and can pass NIST randomness test,which conforms to the basic characteristics of true random number.2)For the problem of the low processing speed of information-theoretically provable post-processing for original random number,we put forward high speed post-processing algorithm based on matrix decomposition,which remarkably reduces hard-ware logic resource needed of the algorithm,and successfully realizing high-speed post-processing calculation based on large scale Toeplitz matrix in FPGA,breaking the technology bottleneck about the speed of post-processing on high speed random number,and developing patent rights.