On the use of stochastic logic for nonlinear circuits and systems
[eng] The increasing pervasion of devices related to wireless networks, data process and transport and in general the Internet of Things (IoT) has led to a resultant rapid increase of edge devices. The numbers are enormous and the predictions are wondrous [1]; in a few years (by 2025) 180 ZBytes wil...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2023 |
| País: | España |
| Institución: | CBUC, CESCA |
| Repositorio: | TDR. Tesis Doctorales en Red |
| OAI Identifier: | oai:www.tdx.cat:10803/688339 |
| Acceso en línea: | http://hdl.handle.net/10803/688339 |
| Access Level: | acceso abierto |
| Palabra clave: | Computación estocástica Memristores Sistemas no lineales Computación Estocástica 621.3 |
| Sumario: | [eng] The increasing pervasion of devices related to wireless networks, data process and transport and in general the Internet of Things (IoT) has led to a resultant rapid increase of edge devices. The numbers are enormous and the predictions are wondrous [1]; in a few years (by 2025) 180 ZBytes will be the amount of data to be handled (International Data Corporation - IDC). In addition, IoT devices will exceed 150 billions and it is estimated that the data produced by them will be about 70% of the data produced worldwide (IDC) [1]. It is apparent that all the types of centralized processing, even in the form of cloud, cannot properly and efficiently support this new computing landscape; taking also into account the fact that IoT has to go hand by hand with other technologies regarding artificial intelligence, big data, mobile computing etc., which are being referred to as ubiquitous computing platforms. Therefore, edge computing rises in the horizon, calling for data processing at the edge of the network. All these technologies call for innovative approaches [2, 3] and some of those approaches are approximate computing [4–6], deep learning [7], new post-CMOS devices and architectures [8], or advanced processing techniques [9,10]. One of the fields where more computational power is required, is communications, especially when data privacy protection and security are included. Thus, data encryption emerges as a mandatory element. Nowadays, many techniques and approaches are proposed in the context of the IoT, in all hierarchical levels of data communications, others for ensuring privacy [11] and others for practically ensuring security [12]. As a result, there exist numerous proposals in this sense, usually ubiquitous ones; one such promising option seems to lean towards using chaotic-based encoder-decoder schemes to secure and/or authenticate data transmission in general [13]. There are examples of such secure-communication systems, analog [14, 15], and digital [16,17] ones demonstrating merits like low-cost, circuit simplicity, low-power operation etc. [9, 10, 18, 19]. On the other hand, it seems that approximate computing enables high power savings [6], making this technique a viable candidate for IoT edge devices. This framework offers energy savings by trading accuracy for energy. There are a handful of methods that can successfully implement approximate computing: programming methods and algorithms, hardware impleix mentations and other ubiquitous solutions. As a curious note, this strongly reminds of the way chaos was encountered by Lorenz, finding different solutions of the homonym set of equations because of truncated number storage [20]. An interesting approach had been already introduced by Von Neumann in 1956 [21], based on a series of lectures given by R.S. Pierce in 1952 at California Institute of Technology. This approach, namely Stochastic Computing (SC) or Stochastic Logic, makes a trade-off between calculation time and accuracy. This approach has been successfully applied in fields as diverse as neural network implementation [22, 23], data mining [24], data compression [25], or mathematical calculations (FFT) [26], control [27], or even A/D conversion [28], among others. An important advantage is that it allows for a high reduction in the number of components, thus reducing the power required to run the circuit. However, this comes to a price, since the time required to perform the operation also increases exponentially with the number of bits. There are cases where this trade-off plays a crucial role, like in the case of chaotic systems. It is known that key features of all nonlinear, chaotic systems are: long-term bounded aperiodic behavior, enhanced sensitivity to parameters and initial conditions, and fast de-correlation between past and present [29]. These features, especially sensitivity to initial conditions and parameter values, make proper implementation of chaotic systems within approximate computing frameworks, difficult, not impossible though. A successful example is the case of implementing a chaotic oscillator in a purely digital environment [17], but with the cost of creating a much more complicated (higher-dimensional) implementation. In this thesis, we have focused on the use of Stochastic Computing (SC) applied to the solution of several issues. As a first step, we’ve evaluated the performance of SC when implementing nonlinear circuits with chaotic behavior. Specifically, in chapter 2 we have implemented the so-called Shimizu- Morioka system in SC. The results, partially published in [30]) have shown that this implementation can be useful, if it’s done with a limited number of bits with parallel implementations. In chapter 3, we present three different implementations of memristors and memristor-based systems based on SC. The first one is a purely digital memristor based on a flux-charge model. This part of the chapter was partially published in [31]. The second proposal, partially published in [32], includes the implementation of a switched capacitor memristive emulator. Finally, at the last part of the chapter regarding the third presented implementation, we propose an improvement to the previous emulator by adding SC. This last part of the chapter was partially published in [33], [34] and [35]. In the next chapter, we propose a few applications developed with memristor emulators and SC. At the first part of the chapter 4, we design and implement a system for solving mazes, using the memristive emulator as a x delay element. We initially checked the system’s operation in Matlab and then we exported it to two different FPGAs. This part of the chapter was partially published in [36]. The end of chapter 4 includes a proposal for the use of SC in designing and implementing Cellular Nonlinear Networks (CNN). By combining Matlab and a FPGA, we develop a CNN and apply it to three real-time processes (Store, edge detection and image sharpening) for both grey and color images. This part of the chapter was partially published in [37] and [38]. Finally, the thesis ends with a concluding chapter, where next to briefly describing the presented work, we discuss about the value and The efficiency of using SC in several cases, especially for edge computing applications. |
|---|