Design,Implementation And Application Of Pseudo Random Number Generator Based On Hopfield Memristive Neural Network

With the rapid development of digital communication technology,especially in the increasingly popular smart phones and network communication,more and more people require the security of private information.Chaos is a common phenomenon in nature,which has complex dynamic characteristics.Memristor is a nonlinear device with the characteristics of nanometer characteristic size,very low energy consumption and fast processing and storage of information.Combined with chaotic system,memristive chaotic system is formed,which can improve the complexity of the dynamic characteristics of chaotic system.Hopfield neural network has the characteristics of real-time,continuity,boundedness and operation synchronization,which provides a basic model for studying the chaotic behavior of neural system,but the fixed synaptic weight between neurons hinders the development of Hopfield neural network.Hopfield neural network,as a kind of variable synaptic memory,can improve the characteristics of Hopfield neural network.As the core part of encryption algorithm,Random Number Generator plays an important role in the field of information security,such as physical model simulation or image encryption.Pseudo Random Number Generator(PRNG)based on Hopfield memristive neural network has higher randomness and output rate,which is more suitable for engineering applications.Therefore,based on the special activation gradient Hopfield memristive neural network and the Hopfield memristive neural network under electromagnetic radiation,this paper proposes two PRNG design methods,which are implemented on the Field Programmable Gate Array(FPGA)development platform respectively.The security analysis shows the randomness and feasibility of the two PRNG design methods.Taking the PRNG based on Hopfield memristive neural network under electromagnetic radiation as the encryption sequence source,an image encryption system is designed and implemented on the hardware development platform.The research work of this paper is as follows:(1)By adding a suitable memristor to the Hopfield neural network with special activation gradient,a Hopfield memristive neural network with special activation gradient is proposed.It is simulated and dynamically analyzed,and implemented on FPGA development platform.Then,based on Hopfield memristive neural network with special activation gradient,a PRNG is proposed.Its post-processing unit is composed of nonlinear post processor and XOR calculator,which effectively ensures its randomness.It is implemented on Vivado 2018.3 design tool using Xilinx XC7Z020CLG400-2 FPGA chip and Verilog HDL.Relevant experiments comply with IEEE 754-1985 high-precision 32-bit floating-point number standard.Finally,the proposed PRNG passed the NIST SP800-22 randomness test and safety analysis.(2)A PRNG based on Hopfield memristive neural network under electromagnetic radiation is proposed,which has a negative feedback controller and is implemented on FPGA development platform.The feedback controller takes the magnetic flux passing through the neuron cell membrane as the feedback condition to interfere with the output of other neurons and avoid periodic and chaotic degradation.The post processor of PRNG consists of 32 registers and 15 XOR comparators.The post-processing unit composed of 4 post-processing determines the performance of PRNG.Finally,the random sequence output by the PRNG is tested and analyzed by using NIST SP800-22 statistical test suite,which shows its safety and randomness.(3)PRNG based on Hopfield memristive neural network under electromagnetic radiation can output high-quality random sequences.Based on it,an image encryption system is designed and implemented on FPGA development platform.Finally,the simulation and security analysis are carried out on the MATLAB software platform,and the results prove the security and feasibility of the image encryption system.