Search

Efficiently deploying deep neural networks on resource-limited embedded systems is crucial to meet real-time and power consumption requirements. Utilizing metareasoning as a higher-level controller along with tiny machine learning (TinyML) can enhance energy efficiency and reduce latency on such systems by overseeing available resources. This study introduces MetaTinyML, a comprehensive metareasoning framework for self-guided navigation on TinyML platforms. The framework adapts its decision-making process by factoring in environmental changes to select the most suitable algorithms for the current scenario. Implementation of MetaTinyML on an NVIDIA Jetson Nano 4-GB system integrated with a Jetbot ground vehicle demonstrated up to 50% power consumption enhancement. View a video demonstration of the MetaTinyML framework at: Video. [ABSTRACT FROM AUTHOR]

Return-oriented programming (ROP) have emerged as great threats to the modern embedded systems. ROP attacks can be used to either bypass credential verification or modify RAM contents. In this letter, we introduce a simple side-channel technique for the run-time ROP detection. We use processors’ power consumption pattern as an indicator for the potential ROP attacks, which can be deployed across different platforms. We avoid the computational complexities of training machine learning models by using a simple linear comparison algorithm to compare the known and unknown power patterns to discern anomalies. For evaluation, we implement both the ROP attacks in multiple scenarios on the benchmarks with various complexity levels. We demonstrate the robustness of our approach and also outline some potential overheads that the approach incurs for the run-time ROP detection. [ABSTRACT FROM AUTHOR]

This letter introduces ViTSen, optimizing vision transformers (ViTs) for resource-constrained edge devices. It features an in-sensor image compression technique to reduce data conversion and transmission power costs effectively. Further, ViTSen incorporates a ReRAM array, allowing efficient near-sensor analog convolution. This integration, novel pixel reading, and peripheral circuitry decrease the reliance on analog buffers and converters, significantly lowering power consumption. To make ViTSen compatible, several established ViT algorithms have undergone quantization and channel reduction. Circuit-to-application co-simulation results show that ViTSen maintains accuracy comparable to a full-precision baseline across various data precisions, achieving an efficiency of ~3.1 TOp/s/W. [ABSTRACT FROM AUTHOR]

This letter proposes using intermittent chaotic clocks, generated from chaotic maps, to drive cryptographic chips running the advanced encryption standard as a countermeasure against correlation power analysis (CPA) attacks. Five different chaotic maps, namely, the logistic map, the Bernoulli shift map, the Henon map, the tent map, and the Ikeda map, are used in this letter to generate chaotic clocks. The performance of these chaotic clocks is evaluated in terms of timing overhead and the resilience of the driven chip against CPA attacks. All proposed chaotic clocking schemes successfully protect the driven chip against attacks, with the clocks produced by the optimized Ikeda, Henon, and logistic maps achieving the lowest-timing overhead. These optimized maps, due to their intermittent chaotic behavior, exhibit lower-timing overhead compared to previous work. Notably, the chaotic clock generated by the optimized Ikeda map approaches the theoretical limit of timing overhead, i.e., half the execution time of a reference periodic clock. [ABSTRACT FROM AUTHOR]

Javascript object notation (JSON) and extensible markup language (XML) are two data serialization methods that have been compared over many applications, including client-server transmission, Internet communication, and large-scale data storage. Due to the smaller file size, JSON is faster for transmitting data. However, XML is better for sending complex data structures. This letter compares the two data formats in the context of an embedded system, considering factors, such as time, memory, and power to identify efficient characteristics of each method. Programs for each format were written, optimized, and compared for the same dataset. The JSON file was found to be 24.7% smaller than the XML file. This led to a shorter program run-time and less power being consumed when reading and processing the file. However, the program to deserialize the XML file took up 16.7% less flash memory than its JSON counterpart. Overall, JSON was found to be a better choice for systems when collecting large amounts of data, requiring high speed communication, or running for an extended period between battery charges. However, XML is proposed for systems that have limited flash memory. [ABSTRACT FROM AUTHOR]

To achieve high reliability, fault-tolerance techniques are exploited, but they may increase power consumption and temperature beyond safe limits. Therefore, power-aware fault-tolerance techniques should be used to manage power and temperature issues. We tolerate both permanent and transient faults through hybrid fault-tolerance techniques. In this letter, at first, we investigate how much power and temperature are increased when a hybrid fault-tolerance technique is applied to multicore embedded systems. Then, we propose a peak-power-aware hybrid fault-tolerant technique to meet the temperature constraint. Transient-temperature-based safe power (T-TSP) is a new power budgeting technique whose calculation is based on the current temperature of the processing core. Assigning dynamic budgets through T-TSP to processing cores allows us to effectively reach the full performance of processing cores. Experiments show that our proposed method reduces peak power and energy consumption on average by 13.5% (up to 50.7%) and 41.8% (up to 67.4%), respectively and improves the schedulability on average by 6.8% (up to 22.4%) compared to state-of-the-art methods while meeting the system reliability target. [ABSTRACT FROM AUTHOR]

Parsing value change dump (VCD) files through signal turnover behavior is important for power analysis and estimation. In practical applications, the size of VCD files can reach hundreds of GB. Thus, designing an efficient VCD parser for parsing large VCD files is of great significance. Different from the traditional hash search functions applied in many VCD parsers, this letter proposes a specific search algorithm based on the rules of identifiers in VCD files. Then, a high-performance VCD parser is constructed. The parser supports single-core and multicore modes. Based on the regression test, the function of the VCD parser is verified. Experimental results show that the proposed VCD parser is faster and more functional than the vcd2saif. In multicore mode, our VCD parser only takes about 8.139 s to parse 1-GB VCD files, and the time consumption of the search algorithm only accounts for 2% of the total CPU time. [ABSTRACT FROM AUTHOR]

The integration of electric vehicles (EVs) into the power grid poses significant challenges and opportunities for energy management systems. This is especially concerning for parking lots or private building condominiums in which refurbishing is not possible or is costly. This paper presents a real-time monitoring approach to EV charging dynamics with battery storage support over a 24 h period. By simulating EV demand, state of charge (SOC), and charging and discharging events, we provide insights into the operational strategies for energy storage systems to ensure maximum charging simultaneity factor through internal power enhancement. The study uses a time-series analysis of EV demand, contrasting it with the battery's SOC, to dynamically adjust charging and discharging actions within the constraints of the upstream infrastructure capacity. The model incorporates parameters such as maximum power capacity, energy storage capacity, and charging efficiencies, to reflect realistic conditions. Results indicate that real-time SOC monitoring, coupled with adaptive charging strategies, can mitigate peak demands and enhance the system's responsiveness to fluctuating loads. This paper emphasizes the critical role of real-time data analysis in the effective management of energy resources in existing parking lots and lays the groundwork for developing intelligent grid-supportive frameworks in the context of growing EV adoption. [ABSTRACT FROM AUTHOR]

Semiconductor manufacturing requires a substantial amount of high-purity water generated through a complex series of treatment processes. Reverse Osmosis (RO) as the most crucial water treatment process contributes the majority of energy consumption and carbon emission in Ultra-Pure Water (UPW) preparation for semiconductor manufacturing. However, there is an opportunity to drive innovation around the current design of feed flow channel in spiral wound RO elements to promote energy efficiency and cost savings. In this study, a novel design of feed channel with 3D printed spacers was compared with conventional design of feed channel with mesh spacers regarding energy consumption. The average head pressure of 3D printed spacer was found to be 20 psi lower than mesh spacer with same permeate flow rate, which achieved a lower specific power of 0.449 kWh/m3, resulting in a 20% energy saving compared with mesh spacer. This study demonstrated that this novel channel construction with 3D printed spacer significantly improves the overall energy efficiency in RO through reduced pressure loss and increased active area, with a potential merit of decreasing the anti-scalant usage and membrane cleaning. [ABSTRACT FROM AUTHOR]

Featured Application: This article developed an improved InfoGAN wind–solar–load extreme scenario generation approach that can inform power system evaluation in extreme scenarios. In recent years, extreme events have frequently occurred, and the extreme uncertainty of the source-demand side of high-ratio renewable energy systems poses a great challenge to the safe operation of power systems. Accurately generating extreme scenarios related to the source-demand side under a high percentage of new power systems is vital for the safe operation of power systems and the assessment of their reliability. However, at this stage, methods for extreme scenario generation that fully consider the correlation between wind power, solar power, and load are lacking. To address these problems, this paper proposes a method for extreme scenario generation based on information-maximizing generative adversarial networks (InfoGANs) for high-proportion renewable power systems. The example analysis shows that the method for extreme scenario generation proposed in this paper can fully explore the correlation between historical wind–solar–load data, greatly improve the accuracy with which extreme scenarios are generated, and provide effective theories and methodologies for the safe operation of a new type of power system. [ABSTRACT FROM AUTHOR]

In real-world inventory problems, the deterioration rate is usually treated as an uncontrolled factor. However, by adopting certain preservation techniques, such deterioration rate can be controlled up to a desired level. Moreover, in certain business affairs, the supplier also provides a permissible delay in payment to encourage the vendors to have further sales. In this paper, we develop an inventory model that considers the power pattern demand under trade credit, permissible completely backlogged shortages, and suitable investment in preservation technology. Additionally, to make the model more realistic, we apply the learning effect to the holding costs. We develop the mathematical model of the problem and its solving policy in both crisp and fuzzy environments. The main purpose of the model is to obtain the optimum preservation technology-based cost and cycle time that maximize the overall profit. Furthermore, to validate our findings, we consider numerical examples and subsequently demonstrate the concavity of the profit function via Mathematica 11.1.1 software. Finally, we study the sensitivity analysis and accordingly present several managerial insights for the benefit of inventory managers. [ABSTRACT FROM AUTHOR]

A compact MOSFET-C floating/grounded meminductor emulator (MIE) model is presented for high operating frequency and low power operation. The proposed MIE uses only 22 MOSFETs and two capacitors. Its performance is theoretically analyzed and rigorously verified using the Cadence Virtuoso software and hardware prototypes. The proposed MIE operates appropriately for a wide range of frequencies up to 5 MHz with $590 \mu \text{W}$ power consumption at a 180nm CMOS technology node and manifests important signature properties. The MIE layout area in 180 nm CMOS technology is $13107.5 \mu \text{m}^{2}$. To analyze the effects of statistical variations in MIE elements, extensive Monte Carlo simulations have been performed to demonstrate the robustness of the proposed MIE. For experimental validation, hardware prototypes have been developed and tested successfully. An MIE-based adaptive learning neuromorphic circuit is presented to show that it can mimic the behavioral responses of amoeba under varying environments such as temperature. [ABSTRACT FROM AUTHOR]

This study investigates the reliability of commonly utilized digital spiking neurons and the potential side-channel vulnerabilities in neuromorphic systems that employ them. Through our experiments, we have successfully decoded the parametric information of Izhikevich and leaky integrate-and-fire (LIF) neuron-based spiking neural networks (SNNs) using differential power analysis. Furthermore, we have demonstrated the practical application of extracted information from the 92% accurate pretrained standard spiking convolution neural network classifier on the FashionMNIST dataset. These findings highlight the potential dangers of utilizing internal information for side-channel and denial-of-service attacks, even when using the usual input as the attack vector. [ABSTRACT FROM AUTHOR]

In deep learning inference, model parameters are pruned and quantized to reduce the model size. Compression methods and common subexpression (CSE) elimination algorithms are applied on sparse constant matrices to deploy the models on low-cost embedded devices. However, the state-of-the-art CSE elimination methods do not scale well for handling large matrices. They reach hours for extracting CSEs in a $200 \times 200$ matrix while their matrix multiplication algorithms execute longer than the conventional matrix multiplication methods. Besides, there exist no compression methods for matrices utilizing CSEs. As a remedy to this problem, a random search-based algorithm is proposed in this letter to extract CSEs in the column pairs of a constant matrix. It produces an adder tree for a $1000 \times 1000$ matrix in a minute. To compress the adder tree, this letter presents a compression format by extending the compressed sparse row (CSR) to include CSEs. While compression rates of more than 50% can be achieved compared to the original CSR format, simulations for a single-core embedded system show that the matrix multiplication execution time can be reduced by 20%. [ABSTRACT FROM AUTHOR]

The primary objective of this article is to develop a mathematical model and determine the optimal policies of an inventory system involving power demand and controlled deterioration through preservation technology. This model comes in handy in a power demand-oriented inventory system with demand high at the end of the period. The model incorporates backlogged shortages and linear holding cost. The triangular neutrosophic numbers (TNN's) are used for a nuanced representation of uncertain and imprecise inventory-related expenses. An efficient algorithm is constructed to minimize the total cost, and obtain optimal positive inventory time, optimum cycle time and minimum preservation technology investment. Few numerical examples are used to illustrate and validate the model. The comparative study conducted between models with and without preservation technology investment reveals a significant reduction in total inventory costs facilitated by the preservation facility. Also, the numerical results obtained in crisp and neutrosophic environment are compared. Specific previously obtained results are discussed to illustrate the theoretical findings. Sensitivity analysis of the model provides managerial insights replicating reality. [ABSTRACT FROM AUTHOR]

This study focuses on energy saving for a Taiwan high-tech DRAM factory as the primary research subject. Collecting operational parameters related to various facility systems and process equipment is initially performed by using the developed energy conversion factors (ECF) calculator. Moreover, innovative fab energy simulation (FES) software has been designed by Taipei Tech. This software is designed for high-tech fab energy consumption analysis. The annual energy consumption data for fabs can be calculated. This data is then converted into carbon dioxide emissions using the power carbon emission coefficient provided by the Bureau of Energy, Ministry of Economic Affairs Taiwan. In this study, five different energy-saving strategies were proposed. The energy consumption and carbon emissions distribution were evaluated to assess the benefits of those different techniques. The findings show that among the existing operational facilities, the use of an exhaust air conditioning unit with reduced enthalpy value setting, with lowered supply air temperature, demonstrates the highest energy-saving. This technique has the potential to annually reduce carbon emissions by approximately 623,158 kg CO2 and operational costs by NT ${\$}$ 6,005,764 (189,602 U.S. ${\$}$). This can reduce the overall manufacturing cost and is also beneficial for the environment. [ABSTRACT FROM AUTHOR]

Near-fault pulse-like (PL) ground motions generally transmit huge amounts of energy into structures during a relatively short period compared with non-pulse-like (NPL) ground motions. Consequently, power demand has emerged as a direct and distinctive measure for evaluating the risk that PL ground motions pose on structures. This study examines the power demands of structures subjected to ground motions recorded at 10 seismic stations in six city (or town) centers during the Mw 7.8 earthquake that hit Türkiye on 6 February 2023. The six cities (or towns), Golbasi, Kahramanmaras, Nurdagi, Osmaniye, Iskenderun, and Antakya, were the locations where an international team conducted field reconnaissance 2 weeks after the earthquake. This study first evaluates the power histories and other seismic responses of single-degree-of-freedom structures exposed to both PL and NPL ground motions. Subsequently, the power spectra of 20 horizontal ground motions recorded at the 10 stations are constructed and examined. Through these investigations, we hope to gain a better understanding of and raise awareness regarding the threats that PL ground motions pose to structures in the six cities (or towns) during the earthquake. [ABSTRACT FROM AUTHOR]

Radiological source search is a challenging task involving detection and identification of weak sources in a constantly changing radiological background. As of now, many radiological source detection algorithms have been proposed; however, their computational complexity and, hence, reliance on power intensive processing units inhibit low-power applications of radiological source search systems. In this work, we introduce the anomaly filter (AF) algorithm; a computationally light, yet effective time-series source detection algorithm based on exponential weighted moving average (EWMA) and Poisson deviance statistics. Then, we demonstrate that the proposed algorithm can be used in ensemble with other more computationally intensive source detection and identification algorithms to achieve both increased detection performance and reduced power consumption. The proposed AF algorithm and the ensemble algorithms were thoroughly benchmarked against several existing source detection and identification algorithms. The results show that the AF algorithm outperforms existing conventional source detection algorithms, and the ensemble approach improves the overall performance of existing source detection and isotope identification algorithms. Furthermore, the AF algorithm and the non-negative matrix factorization approach-based source identification (NMF-ID) algorithm were combined and implemented on a single-board microcontroller, and the power consumption was measured. This ensemble algorithm reduced the power consumption of the NMF-ID algorithm almost by a factor of 100, while improving the detection performance of the overall system.

The primary objective of this article is to develop a mathematical model and determine the optimal policies of an inventory system involving power demand and controlled deterioration through preservation technology. This model comes in handy in a power demand-oriented inventory system with demand high at the end of the period. The model incorporates backlogged shortages and linear holding cost. The triangular neutrosophic numbers (TNN’s) are used for a nuanced representation of uncertain and imprecise inventory-related expenses. An efficient algorithm is constructed to minimize the total cost, and obtain optimal positive inventory time, optimum cycle time and minimum preservation technology investment. Few numerical examples are used to illustrate and validate the model. The comparative study conducted between models with and without preservation technology investment reveals a significant reduction in total inventory costs facilitated by the preservation facility. Also, the numerical results obtained in crisp and neutrosophic environment are compared. Specific previously obtained results are discussed to illustrate the theoretical findings. Sensitivity analysis of the model provides managerial insights replicating reality.

Electric vehicle (EV) charging is widely considered as a key enabling technology that can support system stability and provide ancillary services to the grid. The present work aims to advance the state of the art in dynamic charging of individual EVs within existing AC-charging facilities. The paper proposes two alternative control solutions for tracking of power setpoints in EVs, based on adaptive feedforward and feedback linear controllers, respectively. The control design, which does not require any ad-hoc hardware adjustment of the standard EV charging infrastructure, nor higher-level communication, is supported by extensive real-world tests that have been performed on the workplace charging facility operated in the JRC Ispra campus. The experimental results validate the effectiveness of the proposed control methods in tracking two different current profiles for flexibility scheme qualification.

The evaluation of power demand is fundamental in the Economic Load Dispatch problem, ensuring that the generated power meets the needs of consumers reliably and efficiently in planning system operations. This paper presented two approaches using an Adaptive Neuro-Fuzzy Inference System (ANFIS) and an Artificial Neural Network (ANN) to evaluate the power demand. The modified IEEE 57-Bus system is considered the thermal units that incorporate renewables. The ANFIS and ANN are implemented using MATLAB online version R2023b. The results show that the ANN and ANFIS techniques are suitable for evaluating power demand. A comparison of both methods indicates that ANFIS is relatively superior to the ANNs techniques, considering the coefficient of determination of the ANNs and ANFIS were equal. The accuracy of its results in terms of prediction RMSE for the ANN and ANFIS of 10.147e-05 and 5.2177e-05 for the training and 14.639e-05 and 5.2177e-05 for the testing, respectively. Finally, the prediction accuracy of the ANFIS can be observed to be higher than that of the ANN, but the ANFIS takes longer to process. ANFIS is the method that can be appropriately applied to evaluate the power demand in this research. However, it could not guarantee for other research topics that ANFIS would be better than ANN for the RMSE. It depends on input and output data complexity and the training function type.

In recent years, extreme events have frequently occurred, and the extreme uncertainty of the source-demand side of high-ratio renewable energy systems poses a great challenge to the safe operation of power systems. Accurately generating extreme scenarios related to the source-demand side under a high percentage of new power systems is vital for the safe operation of power systems and the assessment of their reliability. However, at this stage, methods for extreme scenario generation that fully consider the correlation between wind power, solar power, and load are lacking. To address these problems, this paper proposes a method for extreme scenario generation based on information-maximizing generative adversarial networks (InfoGANs) for high-proportion renewable power systems. The example analysis shows that the method for extreme scenario generation proposed in this paper can fully explore the correlation between historical wind–solar–load data, greatly improve the accuracy with which extreme scenarios are generated, and provide effective theories and methodologies for the safe operation of a new type of power system.

The increasing rates of colorectal cancer and associated mortality have attracted interest in the use of computer-aided diagnosis tools based on artificial intelligence (AI) for the detection of polyps at an early stage. Most AI models are implemented on software platforms; however, due to the demands of embedded devices, hardware implementations have to fulfill the demands of real-time applications with better accuracy and low-power consumption. In this letter, we propose an optimized four-layer network that can be implanted into an embedded device and determine the feasibility of implanting our convolutional neural network (CNN) into a microprocessor. The essential functions of the CNN (i.e., padding, convolution, ReLU, max-pooling, fully connected, and softmax layers) are implemented in the microprocessor. The proposed method achieves efficient classification with high performance and takes only 2.5488 mW at a working frequency of 8 MHz. We conclude this letter with a discussion of the results and future direction of research. [ABSTRACT FROM AUTHOR]

Reconfigurable intelligent surfaces (RISs) are considered a key enabler of highly energy-efficient 6G and beyond networks. This property arises from the absence of power amplifiers in the structure, in contrast to active nodes, such as small cells and relays. However, a certain amount of power is still required for RIS operation. To improve their energy efficiency further, we propose the notion of perpetual RISs, which secure the power needed to supply their functionalities through wireless energy harvesting (EH) of impinging transmitted electromagnetic (EM) signals. Toward this, we initially explain the rationale behind such RIS capability and proceed with a presentation of the main RIS controller architecture that can realize this vision under an in-band EH consideration. Furthermore, we present a typical EH architecture, followed by two harvesting protocols. Subsequently, we study the performance of the two protocols under a typical communications scenario. Finally, we elaborate on the main research challenges governing the realization of large-scale networks with perpetual RISs. [ABSTRACT FROM AUTHOR]

The study utilizes open-access data to generate power demand curves for a hybrid automotive system, testing twelve configurations with three different energy management strategies and four values for the degree of hybridization (DOH), the latter representing the share of the total power of the vehicle powertrain supplied by the battery. The first control logic (Battery Main—BTM) uses mainly batteries to satisfy the power demand and fuel cells as backup, while in the other two controllers, fuel cells operate continuously (Fuel Cell Main—FCM) or within a fixed range (Fuel Cell Fixed—FCF) using batteries as backup. The results are assessed in terms of H2 consumption, overall system efficiency, and fuel cell predicted lifespan. The battery is heavily stressed in the BTM and FCF logics, while the FCM logic uses the battery only occasionally to cover load peaks. This is reflected in the battery's State of Charge (SOC), indicating different battery stress levels between the BTM and FCF modes. The FCF logic has higher stress levels due to load demand, reducing battery lifetime. In the BTM and FCM modes, the fuel cell operates with variable power, while in the FCF mode, the fuel cell operates in a range between 90 and 105% of its rated power to ensure its lifetime. In the BTM and FCM modes, hydrogen consumption decreases at almost the same rate as the DOH increases, due to a decrease in battery capacity and a smaller amount of hydrogen being used to recharge it. In contrast, the FCF control logic results in a larger fuel consumption when the DOH decreases. In terms of FC durability, the FCF control logic performs better, with a predicted lifetime ranging from 1815 h for DOH = 0.5 to 2428 h for DOH = 0.1. The FCM logic has the worst performance, with a predicted lifetime of 800 to 808 h, being almost insensitive to the DOH variation. Simulations were performed on two different driving cycles, and similar trends were observed. Simulations taking into account fuel cell (FC) performance degradation showed an increase in hydrogen consumption of approximately 38% after 12 years. Overall, this study highlights the importance of optimizing control systems to improve the performance of fuel cell hybrid vehicles, also taking into account the component of performance degradation. [ABSTRACT FROM AUTHOR]

In this paper, a hybrid strategy is proposed for an innovative 15-level inverter method for the grid-connected photovoltaic (PV) system. The proposed technique is combined with the Zebra Optimization Algorithm (ZOA) plus the spiking neural network (SNN) and is called the ZOASNN method. The major goals of the ZOASNN approach are to fulfill the power demand of load, lessen harmonics, and improve the PV system power regulation or maximal energy conversion. The multilevel inverter (MLI) is used to operate at symmetrical and asymmetrical configurations aimed at utilizing reduced power components. The proposed ZOASNN controller develops the operating modes of two-generation methods to determine the converter switching states for this purpose. Using this control method, load demands are optimally fulfilled while external disturbances and fluctuations in system parameters are minimized. The proposed strategy is implemented in MATLAB and its performance is estimated with other algorithms, such as the Grasshopper Optimization Algorithm (GOA), the Random Forest Algorithm (RFA), and the Cuckoo Search Algorithm (CSA). The proposed method shows high efficiency, low total harmonics distortion (THD), and lower cost than other existing methods. [ABSTRACT FROM AUTHOR]

The purpose of this study was to enhance electric scooter performance utilizing a novel method consisting of an artificial neural network (ANN) and genetic algorithm (GA) to predict power demand, battery voltage, and identify the optimal performance range. For training, validation, and testing, a dataset comprising 1000 data points for each parameter was extracted from a MATLAB-Simulink model. The ANN application was used to identify the battery voltage and power demand, reflecting the simulated results under varying key input parameters. Additionally, the GA was used to identify the optimal performance after the ANN had been trained. The results showed that the ES can achieve a speed of 28.2 km/h while using an optimal power of 553 W, at a wind velocity of 0 m/s, a slope ratio of 0%, and a wheel diameter of 0.37 m. The achieved results show that the ANN-GA method is appropriate for determining the operating and structural parameters for maximizing the performance of electric scooters. To support the simulated results, an experimental study was carried out with an actual road test along the Taehwa river. [ABSTRACT FROM AUTHOR]

Using a 3-D monolithic stacking memory technology of crystalline oxide semiconductor (OS) transistors, we fabricated a test chip having AI accelerator (ACC) memory for weight data of a neural network (NN), backup memory of flip-flops (FF), and CPU memory storing instructions and data. These memories are composed of two-layer OS transistors on Si CMOS, where memories in each layer correspond to a bank. In this structure, bank switching of the ACC memory and the FF backup memory work together, and thus inference of different NNs is switched with low latency and low power so that the power gating standby time can be extended. Consequently, a 92% reduction in power consumption is achieved in inference at a frame rate of 60 fps as compared with a chip using static random access memory (SRAM) as the ACC memory. [ABSTRACT FROM AUTHOR]

Computing-in-memory (CIM) is a promising candidate for highly energy-efficient neural networks, alleviating the well-known bottleneck in Von Neumann architecture. MRAM has garnered significant attention in the CIM field, providing advantages in terms of non-volatility, high speed, and endurance. However, most existing MRAM-CIM primarily support low-precision operations, which poses a challenge in fulfilling the requirements of complex neural network models for high inference accuracy. To resolve this dilemma, an in-memory Booth Multiplier is proposed with the aim of enhancing the energy efficiency of neural networks performing multi-bit multiply-and-accumulate (MAC) operations. The MRAM array stores the multiplicand, while the multiplier is encoded by a Booth encoder into corresponding control signals, which perform negation and shift operations, reducing half of the partial products and accelerating the overall processing. Simulation results demonstrate at least a 17.3% improvement in energy efficiency compared to the previous in-SRAM counterpart in 8-bit multiplication. [ABSTRACT FROM AUTHOR]

Regulators’ increasingly stringent carbon rules to protect the environment are encouraging practitioners to modify their operational activities that are accountable for releasing emissions into the atmosphere. Thereby, practitioners dealing with product inventory planning are seeking proper management strategies not only to increase profits but also to reduce released carbons from operations. In addition, increasing uncertainty in supply operations has motivated suppliers to impose prepayment mechanisms in recent decades. This study examines the best prepayment installment policy for a practitioner for the first time, where the consumption behavior of consumers changes as a result of the combined effects of unit selling price and storage time. Moreover, to make the present inventory planning more realistic, the unit holding cost function is adopted as a power function of the inventory unit's storage period. The goal of this study is to provide the best combined installment for advance payment, price, and replenishment strategies for a practitioner under cap-and-price, cap-and-trade, and carbon tax environmental guidelines by ensuring maximum profit. For this purpose, an algorithm is created by combining all derived theoretical results from the analytical study, whereas the efficacy of the algorithm is assessed through the examination of five illustrative numerical instances. A plethora of noteworthy management insights for the practitioner are obtained by investigating the dynamic shifts in optimal strategies resulting from fluctuations in system parameters. The results reveal that if the demand is low in the nascent phases of the business cycle, then the prudent approach for the practitioner entails procuring a comparatively smaller lot-size using a modest number of payment frequencies and then setting a relatively small unit selling price to increase profits.

This letter introduces a dedicated processor architecture, called MEGACORE, which leverages vector technology to enhance tracking performance in visual simultaneous localization and mapping (VSLAM) systems. By harnessing the inherent parallelism of vector processing and incorporating a floating point unit (FPU), MEGACORE achieves significant acceleration in the tracking task of VSLAM. Through careful optimizations, we achieved notable improvements compared to the baseline design. Our optimizations resulted in a 14.9% reduction in the area parameter and a 4.4% reduction in power consumption. Furthermore, by conducting application benchmarks, we determined that the average speedup ratio across all stages of the tracking process is 3.25. These findings highlight the effectiveness of MEGACORE in improving the efficiency and performance of VSLAM systems, making it a promising solution for real-world implementations in embedded systems. [ABSTRACT FROM AUTHOR]

Dynamic power management (DPM) techniques on mobile systems are indispensable for deep learning (DL) inference optimization, which is mainly performed on battery-based mobile and/or embedded platforms with constrained resources. To this end, we characterize CNN workloads using object detection applications of YOLOv4/-tiny and YOLOv3/-tiny, and then propose integrated CPU–GPU DVFS governor policies that scale integrated pairs of CPU and GPU frequencies to improve energy–delay product (EDP) with negligible inference execution time degradation. Our results show up to 16.7% EDP improvements with negligible (mostly less than 2%) performance degradation using object detection applications on NVIDIA Jetson TX2. [ABSTRACT FROM AUTHOR]

Stochastic computing (SC) is an emerging paradigm that offers hardware-efficient solutions for developing low-cost and noise-robust architectures. In SC, deterministic logic systems are employed along with bit-stream sources to process scalar values. However, using long bit-streams introduces challenges, such as increased latency and significant energy consumption. To address these issues, we present an optimization-oriented approach for modeling and sizing new logic gates, which results in optimal latency. The optimization process is automated using hardware–software cooperation by integrating Cadence and MATLAB environments. Initially, we optimize the circuit topology by leveraging the design parameters of two-input basic logic gates. This optimization is performed using a multiobjective approach based on a deep neural network. Subsequently, we employ the proposed gates to demonstrate favorable solutions targeting SC-based operations. [ABSTRACT FROM AUTHOR]

Optimal power source sizing and energy management strategies are crucial in the problem of component sizing for hybrid electric vehicles fuel cell. Ensuring cost-effective sizing while meeting power demand necessitates consideration of these factors as well, thereby ensuring a good driving range, reduced energy-loss and consumption, and minimal degradation of fuel cells and batteries for hybrid power sources. The purpose of this work concerns the sizing and the modelling of a power source utilized in a fuel cell hybrid vehicle, the principal source of energy is a Proton Exchange Membrane Fuel Cell, while an Ultra-Capacitor bank serves as an auxiliary source. The sizing algorithm initiates by computing the power demand, which is determined by the mechanical characteristics of the vehicle. This calculation involves considering the instantaneous speed of the chosen drive cycle and the instantaneous road gradient. Subsequently, the algorithm proceeds to determine the mechanical power needed by the motor. In this article, a frequency splitting approach is employed to determine the power distribution between the SC and the fuel-cell for Worldwide Harmonised Light Vehicles Test Procedure (WLTP) driving cycles. The fuel cell operates effectively at low frequencies, whereas the supercapacitor provides power at high frequencies. The efficiencies of every power transformer, including the motor, gearbox, differential, and DC-DC converters, are considered in our work. The data analysis is conducted using the MATLAB software environment. The obtained results demonstrated that the approach outlined in this research article offers a more efficient sizing and energy management between sources in terms of simplicity and adherence to operational conditions of the fuel cell and supercapacitor. [ABSTRACT FROM AUTHOR]

The use of low-resolution Analog-to-Digital Converters (ADCs) is a practical solution for reducing cost and power consumption for massive Multiple-Input-Multiple-Output (MIMO) systems. However, the severe nonlinearity of low-resolution ADCs causes significant distortions in the received signals and makes the channel estimation and data detection tasks much more challenging. In this paper, we show how Support Vector Machine (SVM), a well-known supervised-learning technique in machine learning, can be exploited to provide efficient and robust channel estimation and data detection in massive MIMO systems with one-bit ADCs. First, the problem of channel estimation for uncorrelated channels is formulated as a conventional SVM problem. The objective function of this SVM problem is then modified for estimating spatially correlated channels. Next, a two-stage detection algorithm is proposed where SVM is further exploited in the first stage. The performance of the proposed data detection method is very close to that of Maximum-Likelihood (ML) data detection when the channel is perfectly known. We also propose an SVM-based joint Channel Estimation and Data Detection (CE-DD) method, which makes use of both the to-be-decoded data vectors and the pilot data vectors to improve the estimation and detection performance. Finally, an extension of the proposed methods to OFDM systems with frequency-selective fading channels is presented. Simulation results show that the proposed methods are efficient and robust, and also outperform existing ones.

The use of low-resolution Analog-to-Digital Converters (ADCs) is a practical solution for reducing cost and power consumption for massive Multiple-Input-Multiple-Output (MIMO) systems. However, the severe nonlinearity of low-resolution ADCs causes significant distortions in the received signals and makes the channel estimation and data detection tasks much more challenging. In this paper, we show how Support Vector Machine (SVM), a well-known supervised-learning technique in machine learning, can be exploited to provide efficient and robust channel estimation and data detection in massive MIMO systems with one-bit ADCs. First, the problem of channel estimation for uncorrelated channels is formulated as a conventional SVM problem. The objective function of this SVM problem is then modified for estimating spatially correlated channels. Next, a two-stage detection algorithm is proposed where SVM is further exploited in the first stage. The performance of the proposed data detection method is very close to that of Maximum-Likelihood (ML) data detection when the channel is perfectly known. We also propose an SVM-based joint Channel Estimation and Data Detection (CE-DD) method, which makes use of both the to-be-decoded data vectors and the pilot data vectors to improve the estimation and detection performance. Finally, an extension of the proposed methods to OFDM systems with frequency-selective fading channels is presented. Simulation results show that the proposed methods are efficient and robust, and also outperform existing ones.

With the increasing number of electric vehicles in the transport sector, the relevance of accurate energy and power demand predictions of electromobility is growing. Thereby, different vehicle functions, especially heating and air conditioning and vehicle automation, have a significant influence. In accordance with the upcoming Euro-7 emissions standard, the energy consumption for heating even has to be contained in the manufacturer’s consumption data in the future. To increase the accuracy of energy and power demand predictions of electromobility, the energy consumption of vehicle functions such as heating and air conditioning as well as the energy savings through vehicle automation must be considered.This paper presents approaches for modeling and simulation the energy consumption of heating, air conditioning and vehicle automation which can be used as an extension of electric vehicles WLTP (Worldwide Harmonized Light-Duty Vehicles Test Procedure) consumption simulation on the level of vehicle classes. The Germany-wide results of the electric vehicles energy demand for heating and air conditioning on the level of NUTS3-areas (Nomenclature of territorial units for statistics) and vehicle classes show regionally different results and confirm the relevance of the research approach. Vehicle automation results are described on the level of the five SAE automation levels (Society of Automotive Engineers automation levels) and the vehicle classes. The approaches and results can be used for single vehicles or assumed vehicle fleets.

Characterizing the electric energy curve can improve the energy efficiency of existing buildings without any structural change and is the basis for controlling and optimizing building performance. Artificial Intelligence (AI) techniques show much potential due to their accuracy and malleability in the field of pattern recognition, and using these models it is possible to adjust the building services in real time. Thus, the objective of this paper is to determine the AI technique that best forecasts electrical loads. The suggested techniques are random forest (RF), support vector regression (SVR), extreme gradient boosting (XGBoost), multilayer perceptron (MLP), long short-term memory (LSTM), and temporal convolutional network (Conv-1D). The conducted research applies a methodology that considers the bias and variance of the models, enhancing the robustness of the most suitable AI techniques for modeling and forecasting the electricity consumption in buildings. These techniques are evaluated in a single-family dwelling located in the United States. The performance comparison is obtained by analyzing their bias and variance by using a 10-fold cross-validation technique. By means of the evaluation of the models in different sets, i.e., validation and test sets, their capacity to reproduce the results and the ability to properly forecast on future occasions is also evaluated. The results show that the model with less dispersion, both in the validation set and test set, is LSTM. It presents errors of −0.02% of nMBE and 2.76% of nRMSE in the validation set and −0.54% of nMBE and 4.74% of nRMSE in the test set. [ABSTRACT FROM AUTHOR]

Copyright of Journal of South-Central Minzu University (Natural Science Edition) is the property of Journal of South-Central Minzu University (Natural Science Edition) Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this paper, we present a notion of differential privacy (DP) for data that comes from different classes. Here, the class-membership is private information that needs to be protected. The proposed method is an output perturbation mechanism that adds noise to the release of query response such that the analyst is unable to infer the underlying class-label. The proposed DP method is capable of not only protecting the privacy of class-based data but also meets quality metrics of accuracy and is computationally efficient and practical. We illustrate the efficacy of the proposed method empirically while outperforming the baseline additive Gaussian noise mechanism. We also examine a real-world application and apply the proposed DP method to the autoregression and moving average (ARMA) forecasting method, protecting the privacy of the underlying data source. Case studies on the real-world advanced metering infrastructure (AMI) measurements of household power consumption validate the excellent performance of the proposed DP method while also satisfying the accuracy of forecasted power consumption measurements. [ABSTRACT FROM AUTHOR]

Due to the rugged environmental factors in IoT applications and constrained on-chip resources, PUF, as a critical hardware primitive, is a promising solution for key storage, authentication, and ID generation. The existing CMOS-based Arbiter PUFs mainly suffer from low reliability and vulnerability against modeling attacks due to metastability of flipflops and linearity of additive path delays, respectively. Techniques that have been used to enhance the reliability mainly increased the area occupation. Recently magnetic tunnel junction (MTJ) devices have been widely investigated in various circuits. In this article, the proposed PUF utilizes mCell devices, a class of Magnetoresistive devices employing only Magnetic Tunnel Junction (MTJ) devices, as a building block. In this article a novel nonvolatile latch is proposed to act as an arbiter and generates the responses by comparing the current values instead of delays which leads to increased the reliability by subtracting the constant variation rates of MTJs under environmental variation without adding hardware overhead. The characteristics of MTJ like nonvolatility, stochastic switching, chaotic magnetization, low power consumption, and low occupied area have made the proposed PUF to a low power, highly reliable, high randomness and ultra-compact pure magnetic arbiter PUF. The Monte Carlo HSPICE simulation results reveal that the uniformity, uniqueness, bit-aliasing, power consumption, and area of the proposed PUF are 49.24%, 49.87%, 48.64%, 10.771 $\mu \text{W}$ and 0.106 $\mu \text{m}^{2}$ , respectively. In addition, the average BER across a wide temperature range (−50 $\,^\circ \text{C}$ –150 $\,^\circ \text{C}$) and voltage range (0.05 V–0.1 V) is 0.08% and 0.18%, respectively. [ABSTRACT FROM AUTHOR]

The performance of Network-on-Chip depends a lot on the routers and interconnect circuits used in them. The gradual movement of technology towards nanometer scales has increased the core abilities, leading to an increase in delays and interconnect power consumption. As a result, %60 of the total current chip power consumption is related to the interconnect circuits. This paper has presented a simultaneous Bi-directional current-mode transceiver (SBCT-NoC) with two modes of operation, transmitter, and receiver. Two transceivers connected at both ends of an interconnect can use two-way communication through the same link. The proposed transceiver can significantly reduce the delay and power consumption of Network-on-Chip interconnects and increase the reliability using differential logic. This paper has simulated the circuits using 32-nanometer technology while considering the crosstalk noise effect. The results show the considerable advantage of the proposed method compared to previous interconnected circuit methods. [ABSTRACT FROM AUTHOR]