For future network communication, security is one of the main challenges. In this context, the random number generator, which is responsible for generating public keys, private keys, and other kinds of random numbers, is a critical engine in cryptographic algorithms and data secrecy. Cryptographic algorithms require high computational capabilities and high power, which can be challenging in hardware-constrained devices. Therefore, the hardware-based random number generator, which provides high throughput at low power, is an essential component for future devices. Many approaches have been developed to enhance the randomness and security of random number generators used in constrained devices. However, current hardware-based random number generators have been facing design challenges, such as energy efficiency (which is the energy consumed to generate one single random bit), security, and flexibility. To address these challenges, chaos-based random number generators, in which non-linear chaotic systems are playing a key role, have emerged .
For chaos-based random number generators, the chaotic and dynamical characteristics are crucial. The higher the complexity and dynamics of the chaotic system, the higher the randomness of the output bits; however, the randomness of the output bits also depends on other factors, such as the implementation and data post-processing. Recently, considerable attention has been drawn by the research community to continuous chaotic systems. Therefore, this research focuses on developing novel continuous chaotic systems, which are highly dynamic and highly data dimensional, to improve the security and throughput of the proposed random number generators.
The contributions of this dissertation are fourfold: (i) development of robust chaotic systems and dynamic characteristics analysis to find the best parameter set (ii) hardware implementation in consideration of the target applications and device platforms; (iii) proposition of engineering applications using the proposed random number generators; and (iv) design and fabrication of true random number generators and pseudo-random number generators to provide ready-to-use security devices that can be used as commercial products.
First of all, we develop several chaotic systems which focus on extending dimensions (which can improve the overall throughput of random number generators), improving system sensitivity level by hiding information of equilibrium points and making the system’s characteristics depend on the initial conditions which are difficult to predict. Mathematical analysis shows the advantages of the proposed chaotic systems compared to previous systems. In hardware implementation, the power consumption, device resources, and security level are tradeoffs. In chaotic system, non-linear function is the main component. Therefore, the non-linear functions, which have low complexity hardware design, are prioritized.
Regarding hardware implementation, we present two different strategies for true random number generators and pseudo-random number generators. Analog circuit design is employed to implement true random number generator, in which the circuit noise and circuit imperfection affect the chaotic characteristics. The pseudo-random number generators are implemented in FPGA devices which can be integrated into multiple hardware platforms.
With IoT platforms, developers can build a wide range of applications, such as real-time monitoring, surveillance, in which data secrecy matters. To prove the benefits of using the designated random number generators in these applications, we provide image encryption/decryption using one-time pad data encoding. Therefore, a chaos-based one-time pad cryptosystem is developed, the data (images) is encoded using the proposed chaos-based random number generator. The receiver and transmitter share information on the initial condition of the chaotic system to recover the key and decode the data.
Finally, after nearly four years of research on random number generators with the achievements we have been getting so far, it is possible to start commercializing our research project. Acknowledging the increasing demand for personal security, our product aims to give users the highest control over their data by providing them with the key generator. Therefore, stored data cannot be decoded even if the host (where the data is stored/reserved) is attacked. Moreover, our product aims to provide high-speed data encoding/decoding solutions for real-time applications such as video streaming, especially during the quarantine of the worldwide pandemic COVID-19.
| Date | 3 Feb 2022 |
|---|
| Original language | American English |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Georges Kaddoum (Supervisor) & Pascal Giard (Co-supervisor) |
|---|