International Core Journal of Engineering 2020-26 | Page 185

2019 International Conference on Artificial Intelligence and Advanced Manufacturing (AIAM) Research on pseudo-random characteristics of new random components Hua Jiang* Changchun Li Jinpo Fan Department of electronic and communication engineering Beijing Electronic Science and Technology Institute Beijing, China Department of cyberspace security Beijing Electronic Science and Technology Institute Beijing, China [email protected] Department of electronic and communication engineering Beijing Electronic Science and Technology Institute Beijing, China [email protected] with good randomness are widely used in many fields such as multiple access communication, spread spectrum communication, and cryptography because they can be determined in advance, and copying and generation are relatively easy. Abstract—With the maturity of modern cryptography theory and research, the security of traditional stream cipher systems is increasingly vulnerable to various attacks. In this regard, people began to work on various new pseudo-random sequence generation algorithms. The Linear Feedback Shift Register (LFSR) is the basic component of the most commonly used keystream generators, but because of its constant coefficients, it is weak against algebraic attacks, rational approximation algorithms, and Berlekamp-Massey (B-M) algorithm attacks. Thus we propose a method of automatically changing the coefficients to improve the linear feedback shift register. It is proved by theory that the sequence generated by the new variable coefficient linear feedback shift register has better pseudo-randomness and the security is greatly improved. This paper analyzes the current attack algorithm based on the characteristics of pseudo-random sequences generated by linear feedback shift registers, such as algebraic attack [2], rational approximation algorithm, Berlekamp-Massey (BM) [3] algorithm, etc. An improvement is proposed: a linear feedback shift register (LFSR) with a fixed coefficient is changed into a linear feedback shift register whose coefficients change continuously with the sequence output. This method has been proven to increase the randomness and complexity of the sequence and improve security. Keywords—LFSR, Variable coefficient, Tap changeable, m- sequence II. S EQUENCE T HEORY A ND R ANDOMNESS I NDEX I. I NTRODUCTION A. Random Sequence Concept Random sequences can be generally classified into three types according to their nature: true random sequences, quasirandom sequences, and pseudo-random sequences. With the rapid development of computers and communication networks, the degree of informatization in society has rapidly increased. However, this has also brought about a series of problems such as information security and confidentiality, which has aroused widespread concern. As a key technology in information security, cryptography not only guarantees information confidentiality, but also ensures information integrity and verification to prevent information from being stolen or forged. Truly random sequences, such sequences are unpredictable, and it is not possible to have two identical true random sequences. The generation of true random sequences can only be achieved by some random physical processes, such as electronic noise and thermal noise present in electronic devices, decay of radioactive elements, and the like. Manufacturing costs and difficulty are enormous. A stream cipher, also known as a sequence cipher, is based on the principle that a pseudo-random sequence with excellent performance is generated by means of certain components such as a shift register, and is encrypted as a key pair to obtain a ciphertext sequence. Moreover, compared with block ciphers, stream ciphers have no or only limited error propagation, which is difficult to implement, fast in encryption, and dependent on the randomness and unpredictability of key stream sequences. The synchronization and security are very high. How to design a sequence with good randomness has always been the focus of research on sequence ciphers [1]. Qusai-random sequences, also called low-difference sequences. As the name implies, for all lengths n, the , , …, has a lower difference. In subsequence general, quasi-random sequences are generated by some mathematical formulas. Of course, in a strict sense, they are not true random numbers, but they are considered to have true random number properties as long as they satisfy the conditions of use. Pseudo-random number, which has similar properties to true random sequences, including some statistical properties, but differs in that it can be generated and processed repeatedly. Pseudo-random sequences are typically generated by circuits with periodic characteristics [4]. The inability to predetermine and not repeat implementation is a feature of random sequences. Another feature is that most sequences have some random characteristics. There is currently no conclusion as to whether a true random sequence can be generated, but we can be sure that generating and controlling random sequences is currently difficult to achieve. Therefore, pseudo-random sequences 978-1-7281-4691-1/19/$31.00 ©2019 IEEE DOI 10.1109/AIAM48774.2019.00040 B. Pseudo-Random Sequence Evaluation Index From the perspective of Shannon's unconditional security, 163