基于掺铒光纤激光器真随机序列与神经密码耦合加密方法研究

发布时间：2018-06-29 22:16

本文选题：真随机序列 + 掺铒光纤激光器　；参考：《吉林大学》2016年博士论文

【摘要】：随机序列分为真随机序列和伪随机序列。两种序列都可以作为密码应用在加密领域,但伪随机序列由算法产生,再好的算法也会使伪随机序列具有周期性,往往不能保证密码系统的安全。真随机序列来源于真实的物理世界,能够实现真正的“一次一密”,是密码系统最可靠的来源。但真随机序列的产生需要物理过程的参与,相对较难,后续处理也比较复杂,所以其应用和发展也受到了一定程度的限制。产生真随机序列的方法有很多,大体上可以分为两类,一类是不依赖于数学模型的,如:电器设备的噪声,频率抖动,鼠标轨迹,单光子发射等。另一类是依赖于数学模型的,如光学系统产生的混沌。不依赖于数学模型的方法所产生的随机序列往往特点并不明显,而依赖于数学模型所产生的随机序列往往具备一些特点,例如混沌或超混沌特性。而这些具备混沌或超混沌特性的序列应用在加密系统往往更有优势。当前,光学系统产生的混沌或超混沌序列最适宜应用在加密领域。现有的实验为了得到稳定的系统输出,都采用全光学器件的方法。但全光器件的成本较高,实现较复杂,仅仅为了得到随机序列而采用全光系统,会加大前期实验的难度。所以,如果能够既应用光学系统的模型,同时又不使用光学器件产生真随机序列,就成为亟待解决的课题。本文采用物理方法产生真随机序列,模板为掺铒光纤激光器的动力学系统。针对本文工作,提出以下几个创新点:1、设计了模拟电路实现单环掺铒光纤激光器真随机序列输出。采用电阻、电容、运算放大器等器件按照动力学方程的要求,进行计算,最终实现其运算功能,为带有延迟项的其它不可实现数学模型提供了设计思路。2、通过增加一个延迟项,改进了单环掺铒光纤激光器系统,使系统更复杂,随机性更好,不可预测性更强,并成功输出超混沌序列。改造后的变形系统由于结构复杂,现阶段还只能产生伪随机序列。3、基于双环掺铒光纤激光器模型,设计了模拟电路,为高维超混沌系统的设计提供了思路。4、首次提出将真随机序列同奇偶树型机神经网络耦合产生供加密使用的密钥流。并将密钥流应用于图像加密,并取得较好效果。
[Abstract]:Random sequences are divided into true random sequences and pseudorandom sequences. Both sequences can be used as ciphers in the field of encryption, but pseudorandom sequences are generated by algorithms. However, even good algorithms can make pseudorandom sequences periodic, which often can not guarantee the security of cryptographic systems. The true random sequence comes from the real physical world and can realize the real "one time and one secret", which is the most reliable source of cryptosystem. However, the generation of true random sequences requires the participation of physical processes, which is relatively difficult, and the subsequent processing is more complex, so its application and development are limited to a certain extent. There are many methods for generating true random sequences, which can be divided into two categories, one is independent of mathematical models, such as noise of electrical equipment, frequency jitter, mouse trajectory, single photon emission and so on. The other is dependent on mathematical models, such as chaos generated by optical systems. The characteristics of random sequences produced by methods independent of mathematical models are not obvious, but random sequences that depend on mathematical models often have some characteristics, such as chaos or hyperchaos. These sequences with chaotic or hyperchaotic properties often have more advantages in encryption systems. At present, chaotic or hyperchaotic sequences generated by optical systems are most suitable for encryption. In order to obtain stable system output, all optical devices are used in existing experiments. However, the cost of all-optical devices is high and the implementation is more complex, so it will be more difficult to use all-optical system in order to obtain random sequences. Therefore, if we can not only use the optical system model, but also do not use optical devices to produce true random sequences, it will become an urgent task to be solved. In this paper, the true random sequence is generated by physical method, and the template is the dynamic system of erbium-doped fiber laser. Aiming at the work of this paper, this paper puts forward the following innovative points: 1, and designs an analog circuit to realize the output of true random sequence of single-ring erbium-doped fiber laser. The resistors, capacitors, operational amplifiers and other devices are used to calculate according to the requirements of the dynamic equation. Finally, the calculation function is realized, which provides the design idea for other unrealizable mathematical models with delay term. 2. By adding a delay term, The single-ring erbium-doped fiber laser system is improved to make the system more complex, more random, more unpredictable, and the hyperchaos sequence is successfully output. Because of the complicated structure, the modified deformation system can only produce pseudorandom sequence. 3. Based on the model of dual-ring erbium-doped fiber laser, the analog circuit is designed. For the design of high dimensional hyperchaotic system, a new method is proposed, which is to couple the true random sequence with the odd-even tree type computer neural network to generate the encryption key stream for encryption for the first time. The key stream is applied to image encryption, and good results are obtained.
【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN918.4;TN249

【参考文献】