This paper presents a novel, real time, high speed and robust chaos-based pseudo random number generator (PRNG) design using the structures of artificial neural network (ANN)-based 2D chaotic oscillator and ring oscillator. In this study, four different robust PRNGs have been implemented using four different approaches (TS-55, Elliott-93, Elliott-2, Cordic-LUT) of TanSig activation functions (TSAF) that have been used in the design of ANN-based 2D chaotic oscillators. The designs have been coded in VHDL using IEEE-754–1985 number standard. The PRNGs have been synthesized for Virtex-6 FPGA chip using Xilinx ISE Design Tools. After Place&Route operation, FPGA chip statistics and maximum operating frequencies have been presented. The maximum operating frequencies of the proposed PRNGs range between 184 and 241 MHz. The 1 Mbit of bit streams generated by PRNGs have been subjected to NIST-800–22 randomness tests. Among 4 different proposed PRNGs, the proposed PRNGs that designed using the Elliott-93 and Cordic-LUT approaches have successfully passed all NIST-800–22 tests and have a bit production rate of 241 Mbps. The proposed secure hybrid chaos-based PRNG structures were compared with similar studies conducted in the literature in recent years. According to the results, the proposed FPGA-based secure new chaotic PRNG structures are useful in cryptographic applications.

本文利用基于人工神经网络（ANN）的二维混沌振荡器和环形振荡器的结构，提出了一种新颖，实时，高速，鲁棒的基于混沌的伪随机数发生器（PRNG）设计。在这项研究中，已使用在基于ANN的设计中使用的TanSig激活函数（TSAF）的四种不同方法（TS-55，Elliott-93，Elliott-2，Cordic-LUT）实现了四种不同的鲁棒PRNG。 2D混沌振荡器。设计已使用IEEE-754–1985数字标准在VHDL中进行了编码。PRNG已使用Xilinx ISE设计工具为Virtex-6 FPGA芯片合成。布局布线后，将显示FPGA芯片统计信息和最大工作频率。建议的PRNG的最大工作频率范围为184至241 MHz。PRNG生成的1 Mbit比特流已经过NIST-800-22随机性测试。在4种不同的建议PRNG中，使用Elliott-93和Cordic-LUT方法设计的建议PRNG已成功通过所有NIST-800-22测试，并且比特率为241 Mbps。提出的安全的基于混合混沌的PRNG结构与近年来文献中进行的类似研究进行了比较。根据结果​​，所提出的基于FPGA的安全的新型混沌PRNG结构可用于密码学应用。提出的安全的基于混合混沌的PRNG结构与近年来文献中进行的类似研究进行了比较。根据结果​​，所提出的基于FPGA的安全新混沌PRNG结构可用于密码学应用。所提出的基于混合混沌的安全PRNG结构与近年来文献中进行的类似研究进行了比较。根据结果​​，所提出的基于FPGA的安全新混沌PRNG结构可用于密码学应用。