Hasil untuk "Computer engineering. Computer hardware"

YU Yongtao, SUN Ao, LI Ang, ZHU Linlin

In industrial surface Quality Control (QC) scenarios, deep classification neural networks are widely used to classify product images for qualified judgment or quality grading. However, surface QC equipment equipped with deep classification neural networks must meet Attribute Reproducibility and Repeatability (AR&R) assessment requirements. Perturbations in product images, caused by assembly tolerance, equipment vibrations, and other factors, lead to variations in position, angle, brightness, and blurring. These perturbations result in inconsistent classification outputs, causing the surface QC equipment to fail the AR&R assessment, a problem referred to as the network output reproducibility issue. To address this issue, this study proposes a training method for classification neural networks based on Siamese networks. The Siamese primary network is trained using original samples for supervised learning to learn correct classification categories. The Siamese secondary network copies the weights of the primary network via exponential smoothing and generates feature embeddings of perturbed samples corresponding to the original ones. These embeddings are used for comparative learning training of the primary network, enabling it to output consistent classification probabilities for both original and perturbed sample inputs. During inference, only the primary network is retained for product defect classification. The results show that the classification accuracy reaches 99.346 2%, with a classification probability variance of 0.001 016. The described method effectively improves the output reproducibility of deep classification neural networks for industrial product image classification by reducing classification probability variance and enhancing accuracy.

Hayato Ishida, Amal Elsokary, Maria Aslam et al.

Realisation of significant advances in capabilities of sensors, computing, timing, and communication enabled by quantum technologies is dependent on engineering highly complex systems that integrate quantum devices into existing classical infrastructure. A systems engineering approach is considered to address the growing need for quantum-secure telecommunications that overcome the threat to encryption caused by maturing quantum computation. This work explores a range of existing and future quantum communication networks, specifically quantum key distribution network proposals, to model and demonstrate the evolution of quantum key distribution network architectures. Leveraging Orthogonal Variability Modelling and Systems Modelling Language as candidate modelling languages, the study creates traceable artefacts to promote modular architectures that are reusable for future studies. We propose a variability-driven framework for managing fast-evolving network architectures with respect to increasing stakeholder expectations. The result contributes to the systematic development of viable quantum key distribution networks and supports the investigation of similar integration challenges relevant to the broader context of quantum systems engineering.

Zhicheng Xu, Jiawei Liu, Sitao Huang et al.

The growing demand for edge computing and AI drives research into analog in-memory computing using memristors, which overcome data movement bottlenecks by computing directly within memory. However, device failures and variations critically limit analog systems' precision and reliability. Existing fault-tolerance techniques, such as redundancy and retraining, are often inadequate for high-precision applications or scenarios requiring fixed matrices and privacy preservation. Here, we introduce and experimentally demonstrate a fault-free matrix representation where target matrices are decomposed into products of two adjustable sub-matrices programmed onto analog hardware. This indirect, adaptive representation enables mathematical optimization to bypass faulty devices and eliminate differential pairs, significantly enhancing computational density. Our memristor-based system achieved >99.999% cosine similarity for a Discrete Fourier Transform matrix despite 39% device fault rate, a fidelity unattainable with conventional direct representation, which fails with single device faults (0.01% rate). We demonstrated 56-fold bit-error-rate reduction in wireless communication and >196% density with 179% energy efficiency improvements compared to state-of-the-art techniques. This method, validated on memristors, applies broadly to emerging memories and non-electrical computing substrates, showing that device yield is no longer the primary bottleneck in analog computing hardware.

Jung Hee Cheon, Hyeongmin Choe, Julien Devevey et al.

We present HAETAE (Hyperball bimodAl modulE rejecTion signAture schemE), a new lattice-based signature scheme. Like the NIST-selected Dilithium signature scheme, HAETAE is based on the Fiat-Shamir with Aborts paradigm, but our design choices target an improved complexity/compactness compromise that is highly relevant for many space-limited application scenarios. We primarily focus on reducing signature and verification key sizes so that signatures fit into one TCP or UDP datagram while preserving a high level of security against a variety of attacks. As a result, our scheme has signature and verification key sizes up to 39% and 25% smaller, respectively, compared than Dilithium. We provide a portable, constanttime reference implementation together with an optimized implementation using AVX2 instructions and an implementation with reduced stack size for the Cortex-M4. Moreover, we describe how to efficiently protect HAETAE against implementation attacks such as side-channel analysis, making it an attractive candidate for use in IoT and other embedded systems.

Frances Fengyi Yang, Michele Sasdelli, Tat-Jun Chin

Gate quantum computers generate significant interest due to their potential to solve certain difficult problems such as prime factorization in polynomial time. Computer vision researchers have long been attracted to the power of quantum computers. Robust fitting, which is fundamentally important to many computer vision pipelines, has recently been shown to be amenable to gate quantum computing. The previous proposed solution was to compute Boolean influence as a measure of outlyingness using the Bernstein-Vazirani quantum circuit. However, the method assumed a quantum implementation of an $\ell_\infty$ feasibility test, which has not been demonstrated. In this paper, we take a big stride towards quantum robust fitting: we propose a quantum circuit to solve the $\ell_\infty$ feasibility test in the 1D case, which allows to demonstrate for the first time quantum robust fitting on a real gate quantum computer, the IonQ Aria. We also show how 1D Boolean influences can be accumulated to compute Boolean influences for higher-dimensional non-linear models, which we experimentally validate on real benchmark datasets.

Péter Horváth, Attila Nagy, Flóra Hajdu

An important aspect of a sustainable society is the care of elderly people. For a sustainable society, it is necessary to develop cost-effective, culture-appropriate, and sustainable eldercare services, which should guarantee both safety and quality. It is also important to prolong the time in which elderly people can live independently to contribute to sustainable healthcare. For elderly people, a daily activity, climbing stairs, might be difficult. This paper presents an overview of stair-climbing aid devices. The paper introduces the sustainable healthcare of elderly people. Then, different stair-climbing aid devices are explained and analyzed, ranging from passive ones to high-tech stair-climbing exoskeletons. A prototype of a new solution, which is an Intelligent Crutch, is presented. After elaborating on the design criteria, the detailed conceptual design and the working principle of the device are presented. The presented concept fills a gap between expensive and complicated stair climbing aid devices and simple and cheaper solutions. It is cheaper and simpler than the active, complicated stair-climbing devices but performs almost the same function. The device helps elderly and disabled people to climb stairs alone, thus improving the quality of life of the target group. The study concludes with further development tasks in order to help elderly people and contribute to a sustainable society.

Fábio Eid Morooka, Adalberto Manoel Junior, Tiago F. A. C. Sigahi et al.

Applications of deep learning (DL) in autonomous vehicle (AV) projects have gained increasing interest from both researchers and companies. This has caused a rapid expansion of scientific production on DL-AV in recent years, encouraging researchers to conduct systematic literature reviews (SLRs) to organize knowledge on the topic. However, a critical analysis of the existing SLRs on DL-AV reveals some methodological gaps, particularly regarding the use of bibliometric software, which are powerful tools for analyzing large amounts of data and for providing a holistic understanding on the structure of knowledge of a particular field. This study aims to identify the strategic themes and trends in DL-AV research using the Science Mapping Analysis Tool (SciMAT) and content analysis. Strategic diagrams and cluster networks were developed using SciMAT, allowing the identification of motor themes and research opportunities. The content analysis allowed categorization of the contribution of the academic literature on DL applications in AV project design; neural networks and AI models used in AVs; and transdisciplinary themes in DL-AV research, including energy, legislation, ethics, and cybersecurity. Potential research avenues are discussed for each of these categories. The findings presented in this study can benefit both experienced scholars who can gain access to condensed information about the literature on DL-AV and new researchers who may be attracted to topics related to technological development and other issues with social and environmental impacts.

Hien T.N. Nguyen, Thinh H. Pham, Thien D. H. Nguyen

Despite the growing demand for more sustainable food production and consumption like plant-based products, there has not been much research on the sensory properties of plant-based sausages. This research aims to compare the sensory characteristics of plant-based sausages and meat sausages, as well as determine the drivers of likings for sausage products. 8 samples (6 meat products and 2 plant-based products) were evaluated by 103 consumers using the check-all-that-apply (CATA) method with 33 attributes. The consumers also rated their overall liking of each sample. The results showed a large difference in appearance and texture between plant-based sausages and meat sausages, with plant-based sausages more associated with “coarse surface”, “tooth packing” and “crumbliness”. The liking scores of plant-based sausage samples were significantly lower than that of meat sausage samples. The results of the penalty-lift analysis showed that “smooth surface”, “firmness”, “springiness” and “dense surface” were must-have attributes for sausages, while “coarse surface”, “crumbliness”, “fishy odor” and “bitterness” should not be presented in sausage products. The results of this study could potentially be applied to improve the formulation of plant-based sausages.

Roberto Morabito, Mallik Tatipamula, Sasu Tarkoma et al.

The network edge's role in Artificial Intelligence (AI) inference processing is rapidly expanding, driven by a plethora of applications seeking computational advantages. These applications strive for data-driven efficiency, leveraging robust AI capabilities and prioritizing real-time responsiveness. However, as demand grows, so does system complexity. The proliferation of AI inference accelerators showcases innovation but also underscores challenges, particularly the varied software and hardware configurations of these devices. This diversity, while advantageous for certain tasks, introduces hurdles in device integration and coordination. In this paper, our objectives are three-fold. Firstly, we outline the requirements and components of a framework that accommodates hardware diversity. Next, we assess the impact of device heterogeneity on AI inference performance, identifying strategies to optimize outcomes without compromising service quality. Lastly, we shed light on the prevailing challenges and opportunities in this domain, offering insights for both the research community and industry stakeholders.

Casey Duckering

The field of quantum computing is at an exciting time where we are constructing novel hardware, evaluating algorithms, and finding out what works best. As qubit technology grows and matures, we need to be ready to design and program larger quantum computer systems. An important aspect of systems design is layered abstractions to reduce complexity and guide intuition. Classical computer systems have built up many abstractions over their history including the layers of the hardware stack and programming abstractions like loops. Researchers initially ported these abstractions with little modification when designing quantum computer systems and only in recent years have some of those abstractions been broken in the name of optimization and efficiency. We argue that new or quantum-tailored abstractions are needed to get the most benefit out of quantum computer systems. We keep the benefits gained through breaking old abstraction by finding abstractions aligned with quantum physics and the technology. This dissertation is supported by three examples of abstractions that could become a core part of how we design and program quantum computers: third-level logical state as scratch space, memory as a third spacial dimension for quantum data, and hierarchical program structure.

YU Zhichao, LI Yangzhong, LIU Lei, FENG Shengzhong

Through superposition and entanglement, a quantum computing displays significant advantages over classical computers in dealing with problems that require large-scale parallel processing capabilities.At present, a physical quantum computer is limited in scalability, coherence time, and precision of quantum gate operations, so it is feasible to simulate quantum computing on a classical computer for studying quantum advantage and quantum algorithms.However, the computer resources required for quantum computing simulation grow exponentially with the number of qubits.Therefore, it is of great importance to study how to reduce the resources required for large-scale simulation with ensured computational accuracy, precision and efficiency.This paper describes the basic principles and background knowledge of quantum computing, including qubits, quantum gates, quantum circuits and quantum operating systems.Meanwhile, this paper summarizes the classical computer-based methods for simulating quantum computing, and analyzes their design ideas, advantages and disadvantages.Some commonly used simulators are also listed.On this basis, this paper discusses the communication overhead problem of quantum computing simulation, and presents some supercomputer-based methods for optimizing quantum computing simulation from the two aspects of node analysis and communication optimization.

Jesús Roberto del Ángel Ruíz, Xochitl Maya Rosales, Juan Gerardo Avalos et al.

Beamforming is a wireless communication technique used in telecommunications applications, which is used to separate a desired signal from interfering signals. This technique increases the coverage range and reduces the interference problem, improving the performance of the systems. To achieve this operation, adaptive algorithms are required. In this work, an alternating structure for beamforming systems is presented, which is composed of two algorithms, the Sign Regressor Least Mean Square (SR-LMS) and the Affine Projection Like (APL) algorithm. The results show that the proposed structure has the best characteristics of the combined algorithms, obtaining an algorithm with a high convergence speed and lower computational cost compared to other algorithms based on conventional convex combinations.

Homa Esfahanizadeh, Alejandro Cohen, Muriel Medard

To improve the utility of learning applications and render machine learning solutions feasible for complex applications, a substantial amount of heavy computations is needed. Thus, it is essential to delegate the computations among several workers, which brings up the major challenge of coping with delays and failures caused by the system's heterogeneity and uncertainties. In particular, minimizing the end-to-end job in-order execution delay, from arrival to delivery, is of great importance for real-world delay-sensitive applications. In this paper, for computation of each job iteration in a stochastic heterogeneous distributed system where the workers vary in their computing and communicating powers, we present a novel joint scheduling-coding framework that optimally split the coded computational load among the workers. This closes the gap between the workers' response time, and is critical to maximize the resource utilization. To further reduce the in-order execution delay, we also incorporate redundant computations in each iteration of a distributed computational job. Our simulation results demonstrate that the delay obtained using the proposed solution is dramatically lower than the uniform split which is oblivious to the system's heterogeneity and, in fact, is very close to an ideal lower bound just by introducing a small percentage of redundant computations.

Mohamed Mohamed, Abd El-Nabi El-Sayed, Khairy Megalla et al.

Thermal storage tank is a standout amongst the most encouraging methods in solar thermal power stations operation. Accurate selection of appropriate storage system is a significant parameter to ensure the continuous working of thermal solar station during the absence of the sun. This work describes financial analysis of different locations of a 500MW Solar Plant in Egypt and also thermal tank design. The selected three locations which are investigated in this study are Aswan, EL-Arish and Hurghada to build this challenged size solar station. These locations cover the tree levels of the solar intensity in Egypt. This study is achieved by System Advisor Model (SAM) as financial analysis simulation tool. All the solar thermal power plants are working twenty-four hours per day and with sixteen full load hours of thermal energy storage (TES). Parametric design and cost analysis for each location, comparison between these locations are introduced to select the optimum location for 500MW solar power plant. The results of this study is considered a good orientation for feasibility study for CPS (concentrators parabolic system) projects, and it is needed in all over the world in particular, in Egypt for future to produce clean energy.

Min Yee Chin, Zhen Xin Phuang, Marlia Hanafiah et al.

Drastic global population and economic growth have escalated energy and water crises across the world. The overconsumption of energy and water resources without proper planning has entailed adverse environmental impacts. The emergence of microbial fuel cell (MFC) as a bio-electrochemical system (BES) for wastewater treatment and electricity generation shows potential as a prospective solution for the crises. To date, there is no study focused on the environmental impact comparison of different MFC configurations from a life cycle perspective. In this study, the environmental performance of five common MFC configurations (MFC 1: air-cathode MFC; MFC 2: H-type MFC; MFC 3: U-type MFC; MFC 4: flat MFC; and MFC 5: modularized MFC) from the construction stage to the operational stage are being investigated and compared via life cycle assessment methodology. MFC 1, MFC 3, and MFC 4 are single chamber reactors, while MFC 2 and MFC 5 are double chamber reactors. Data collected for this study are mainly sourced from peer-reviewed journal articles and evaluated using the ReCiPe 2016 impact assessment method in SimaPro 9.0 software. The results reveal that the MFC 4 induces the highest overall environmental burdens due to the high hydraulic retention time for wastewater treatment. The other options share significantly low and similar overall environmental burdens. It is also found that the energy consumption from MFC options accounts for 60-90 % of environmental loads in wastewater treatment. The COD level of the treated effluent in all options meets the discharge standard, but the nitrogen and phosphorus content level have to be further reduced to minimise the eutrophication risk to the aquatic ecosystems. This study provides data-driven insights to the renewable energy policymakers and wastewater treatment stakeholders on the environmental potential of different MFC configurations in relieving energy and water crises.

Adarsha Balaji, Shihao Song, Anup Das et al.

Neuromorphic hardware with non-volatile memory (NVM) can implement machine learning workload in an energy-efficient manner. Unfortunately, certain NVMs such as phase change memory (PCM) require high voltages for correct operation. These voltages are supplied from an on-chip charge pump. If the charge pump is activated too frequently, its internal CMOS devices do not recover from stress, accelerating their aging and leading to negative bias temperature instability (NBTI) generated defects. Forcefully discharging the stressed charge pump can lower the aging rate of its CMOS devices, but makes the neuromorphic hardware unavailable to perform computations while its charge pump is being discharged. This negatively impacts performance such as latency and accuracy of the machine learning workload being executed. In this paper, we propose a novel framework to exploit workload-specific performance and lifetime trade-offs in neuromorphic computing. Our framework first extracts the precise times at which a charge pump in the hardware is activated to support neural computations within a workload. This timing information is then used with a characterized NBTI reliability model to estimate the charge pump's aging during the workload execution. We use our framework to evaluate workload-specific performance and reliability impacts of using 1) different SNN mapping strategies and 2) different charge pump discharge strategies. We show that our framework can be used by system designers to explore performance and reliability trade-offs early in the design of neuromorphic hardware such that appropriate reliability-oriented design margins can be set.

Laura-Selin Harding, Georg Fieg

The reactive dividing wall column (RDWC) is a highly integrated apparatus, which combines a reactive distillation and a dividing wall column in one single shell. Thus, high savings in investment as well as operational costs can be achieved compared to conventional process alternatives. Due to the high grade of integration the process behavior of the RDWC is strongly nonlinear and extremely complex. Hence, it is not trivial to understand how the process performs in detail and to predict the advantageousness of the RDWC for a given task during conceptual design. Still, for a safe and energy efficient design as well as a steady operation in the industrial praxis it is essential to know, how the process performances and for which tasks the RDWC is more advantageous than less integrated process alternatives. Therefore, the aim of this research is to generate a profound process understanding and identify process applications for the RDWC. The investigations carried out focus on the influence of non-ideal reaction system properties, such as azeotropic phase equilibria. To determine the best process integration level, the energy optimal designs of the RWDC and less integrated process alternatives are determined and compared. By applying this procedure suitable reaction system characteristics can be identified and so, process applications for the RDWC can be derived. Moreover, heuristics regarding the optimal level of process integration can be deduced. This leads to an easier determination of the optimal process integration level and an acceleration of the conceptual design phase.

Florian Kirchbacher, Martin Miltner, Walter Wukovits et al.

Increasing amounts of renewable energy produced by volatile sources like photovoltaic and wind turbines demand for higher energy storage capacities to achieve a sustainable energy generation and supply. Power-to-methane – storing excess energy via chemical conversion as methane – is one of the most interesting technologies to reach this goal, as it bundles the advantages of large storage capacities, fast response time and use of existing infrastructure. First commercial plants have been realised, but alternative concepts are still heavily researched and in demonstration stage. As most are using complex process setups or achieving low methane product concentrations, process optimisation and simplification is needed, but information on this topic is scarce. To close this gap, a power-to-methane process consisting of catalytic methanation and membrane gas upgrading using biogas as carbon dioxide source was simulated in ASPEN Plus®. Four different process setups were modelled to assess influences of fermentation setup, recycling of membrane off-gas and multistage membrane gas separation as well as pressure and GHSV. Models were parameterised with experimental results obtained from a demonstration plant. It was shown that a process without off-gas recycling requires less energy but leads to hydrogen losses of up to 25 %. Preventing this loss by recycling the off-gas leads to an increase of specific energy demand for hydrogen storage by 17 % and relative membrane area by 11 % for the base case.

Raúl Nozal, Jose Luis Bosque, Ramón Beivide

Heterogeneous systems have become one of the most common architectures today, thanks to their excellent performance and energy consumption. However, due to their heterogeneity they are very complex to program and even more to achieve performance portability on different devices. This paper presents EngineCL, a new OpenCL-based runtime system that outstandingly simplifies the co-execution of a single massive data-parallel kernel on all the devices of a heterogeneous system. It performs a set of low level tasks regarding the management of devices, their disjoint memory spaces and scheduling the workload between the system devices while providing a layered API. EngineCL has been validated in two compute nodes (HPC and commodity system), that combine six devices with different architectures. Experimental results show that it has excellent usability compared with OpenCL; a maximum 2.8% of overhead compared to the native version under loads of less than a second of execution and a tendency towards zero for longer execution times; and it can reach an average efficiency of 0.89 when balancing the load.

Kitti Gelle, Szabolcs Ivan

A prefix grammar is a context-free grammar whose nonterminals generate prefix-free languages. A prefix grammar $G$ is an ordinal grammar if the language $L(G)$ is well-ordered with respect to the lexicographic ordering. It is known that from a finite system of parametric fixed point equations one can construct an ordinal grammar $G$ such that the lexicographic order of $G$ is isomorphic with the least solution of the system, if this solution is well-ordered. In this paper we show that given an ordinal grammar, one can compute (the Cantor normal form of) the order type of the lexicographic order of its language, yielding that least solutions of fixed point equation systems defining algebraic ordinals are effectively computable (and thus, their isomorphism problem is also decidable).