Computer Hardware Research Articles

Cybercrimes in the Darknet and Their Detections: A Comprehensive Analysis and Future Directions

Although the Dark web was originally used for maintaining privacy-sensitive communication for business or intelligence services for defence, government and business organizations, fighting against censorship and blocked content, later, the advantage of technologies behind the Dark web were abused by criminals to conduct crimes which involve drug dealing to the contract of assassinations in a widespread manner. Since the communication remains secure and untraceable, criminals can easily use dark web service via The Onion Router (TOR), can hide their illegal motives and can conceal their criminal activities. This makes it very difficult to monitor and detect cybercrimes over the dark web. With the evolution of machine learning, natural language processing techniques, computational big data applications and hardware, there is a growing interest in exploiting dark web data to monitor and detect criminal activities. Due to the anonymity provided by the Dark Web, the rapid disappearance and the change of the uniform resource locators (URLs) of the resources, it is not as easy to crawl the Drak web and get the data as the usual surface web which limits the researchers and law enforcement agencies to analyse the data. Therefore, there is an urgent need to study the technology behind the Dark web, its widespread abuse, its impact on society and the existing systems, to identify the sources of drug deal or terrorism activities. In this research, we analysed the predominant darker sides of the world wide web (WWW), their volumes, their contents and their ratios. We have performed the analysis of the larger malicious or hidden activities that occupy the major portions of the Dark net; tools and techniques used to identify cybercrimes which happen inside the dark web. We applied a systematic literature review (SLR) approach on the resources where the actual dark net data have been used for research purposes in several areas. From this SLR, we identified the approaches (tools and algorithms) which have been applied to analyse the Dark net data, the key gaps as well as the key contributions of the existing works in the literature. In our study, we find the main challenges to crawl the dark web and collect forum data are: scalability of crawler, content selection trade off, and social obligation for TOR crawler and the limitations of techniques used in automatic sentiment analysis to understand criminals’ forums and thereby monitor the forums. From the comprehensive analysis of existing tools, our study summarizes the most tools. However the forum topics rapidly change as their sources changes; criminals inject noises to obfuscate the forum’s main topic and thus remain undetectable. Therefore supervised techniques fail to address the above challenges. Semi-supervised techniques would be an interesting research direction.

End-to-End Ultrasonic Hand Gesture Recognition.

As the number of electronic gadgets in our daily lives is increasing and most of them require some kind of human interaction, this demands innovative, convenient input methods. There are limitations to state-of-the-art (SotA) ultrasound-based hand gesture recognition (HGR) systems in terms of robustness and accuracy. This research presents a novel machine learning (ML)-based end-to-end solution for hand gesture recognition with low-cost micro-electromechanical (MEMS) system ultrasonic transducers. In contrast to prior methods, our ML model processes the raw echo samples directly instead of using pre-processed data. Consequently, the processing flow presented in this work leaves it to the ML model to extract the important information from the echo data. The success of this approach is demonstrated as follows. Four MEMS ultrasonic transducers are placed in three different geometrical arrangements. For each arrangement, different types of ML models are optimized and benchmarked on datasets acquired with the presented custom hardware (HW): convolutional neural networks (CNNs), gated recurrent units (GRUs), long short-term memory (LSTM), vision transformer (ViT), and cross-attention multi-scale vision transformer (CrossViT). The three last-mentioned ML models reached more than 88% accuracy. The most important innovation described in this research paper is that we were able to demonstrate that little pre-processing is necessary to obtain high accuracy in ultrasonic HGR for several arrangements of cost-effective and low-power MEMS ultrasonic transducer arrays. Even the computationally intensive Fourier transform can be omitted. The presented approach is further compared to HGR systems using other sensor types such as vision, WiFi, radar, and state-of-the-art ultrasound-based HGR systems. Direct processing of the sensor signals by a compact model makes ultrasonic hand gesture recognition a true low-cost and power-efficient input method.

Enhancing Regular Expression Processing through Field-Programmable Gate Array-Based Multi-Character Non-Deterministic Finite Automata

This work investigates the advantages of FPGA-based Multi-Character Non-Deterministic Finite Automata (MC-NFA) for enhancing regular expression processing over traditional software-based methods. By integrating Field-Programmable Gate Arrays (FPGAs) within a data processing framework, our study showcases significant improvements in processing efficiency, accuracy, and resource utilization for complex pattern matching tasks. We present a novel approach that not only accelerates database and network security applications, but also contributes to the evolving landscape of computational efficiency and hardware acceleration. The findings illustrate that FPGA’s coherent access to main memory and the efficient use of resources lead to considerable gains in processing times and throughput for handling regular expressions, unaffected by expression complexity and driven primarily by dataset size and match location. Our research further introduces a phase shift compensation technique that elevates match accuracy to optimal levels, highlighting FPGA’s potential for real-time, accurate data processing. The study confirms that the benefits of using FPGA for these tasks do not linearly correlate with an increase in resource consumption, underscoring the technology’s efficiency. This paper not only solidifies the case for adopting FPGA technology in complex data processing tasks, but also lays the groundwork for future explorations into optimizing hardware accelerators for broader applications.

A New Axial Compressor Test Facility for the Investigation of Aerodynamic Damping Featuring an Electromagnetic Excitation System

Abstract Aerodynamic damping simulations involving detailed computational fluid dynamics (CFD) models are nowadays an integral part of turbomachinery design processes, thanks to advancements made in simulation tools and in computer hardware over the last decades. However, the test cases available in the open literature suitable for the validation and continued development of simulation tools is rather limited. On this background, a new test facility is introduced that allows acquiring relevant and high-quality aerodynamic damping test data on axial compressor rotor test objects in a controlled manner. The test facility is built as a closed loop and thus enables operation at variable pressure levels ranging from near-vacuum to overpressure and with different operating media. Overall damping test data are acquired at controlled synchronous and nonsynchronous vibration conditions. While vibrations of the blades at controlled nonsynchronous vibration conditions can be introduced in the test facility by means of a proprietary electromagnetic excitation system that is acting below the hub line, vibrations at controlled synchronous conditions can be introduced from a set of magnets integrated into the casing. Great care has been spent to create a clean test case without any intruding parts or openings in the test section. In addition, an axial compressor blisk is presented, which acts as test object to show the capabilities of the test facility as well as to allow for detailed investigations of aerodynamic damping. While the first part of the article discusses the setup and the instrumentation of the test facility to excite and measure blade vibration in a nonintrusive way, flow-field measurements and the validation accompanying steady-state Reynolds-averaged Navier-Stokes (RANS) simulations are shown in the second part. The third part of the article focuses on aerodynamic damping measurements for seven different modes at three different pressure levels in total. It can be shown that aerodynamic damping is by far the largest contributor to the overall damping.

Fire Detection Systems Using Feature Entropy Guided Neural Network

Fire detection from video has become possible and more feasible in prevention of fire disaster due to deep convolutional neural networks (CNNs) and embedded processing hardware. Artificial intelligence (AI) methods generally require more computational time and hardware with powerful graphical processing unit (GPU). In this paper, we propose cost-effective deep CNN architecture for fire detection from video with respect to computational performance of Jetson Nano from NVIDIA. In our paper we compare CNN networks (AlexNet and SqueezeNet) with our proposed CNN architecture. The proposed CNN architecture finds equilibrium between efficiency and accuracy for target system (Jetson Nano). We used CNNs which show high accuracy and low loss.

Effect of compressive strain on electronic and optical properties of Cr-doped monolayer WS2.

The electronic properties and optical properties of Cr-doped monolayer WS2 under uniaxial compressive deformation have been investigated based on density functional theory. In terms of electronic structure properties, both intrinsic and doped system bandgaps decrease with the increase of compression deformation, and the values of the bandgap under the same compression deformation after Cr doping are reduced compared with the corresponding intrinsic states. When the compressive deformation reaches 10%, both the intrinsic and doped system band gaps are close to zero. New electronic states and impurity energy levels appear in the WS2 system when doped with Cr atoms. For the optical properties, the calculation and analysis of the dielectric function under each deformation regime of monolayer WS2 show that the compression deformation affects the dielectric function, and when the compression deformation is 10%, the un-doped and Cr-doped regimes show a decrease in ε1(ω) compared to the compression deformation of 8%. For each deformation system, the peak reflections occur in the ultraviolet region. Near the position where the second peak of the absorption spectrum appears, it can be seen that the ability of each system to absorb light gradually decreases with the increase of the amount of deformation and appears to be red-shifted to varying degrees. This study follows the initial principles of the density functional theory framework and is based on the CASTEP module of Materials-Studio software GGA and PBE generalizations are used to perform computations such as geometry optimization of the model. We have calculated the energy band structure of monolayer WS2 with intrinsic and compressive deformations of 2% and 4% using PBE and HSE06, respectively. The band gap values calculated using PBE are 1.802 eV, 1.663 eV, and 1.353 eV, respectively, and the band gap values calculated with HSE06 are 2.267 eV, 2.034 eV, 1.751 eV. The results show that the bandgap values calculated by HSE06 are significantly higher than those calculated by PBE, but the bandgap variations calculated by the two methods have the same trend, and the shape characteristics of the energy band structure are also the same. However, it is worth noting that the computation time required for the HSE06 calculation is much longer than that of the PBE, which is far beyond the capability of our computer hardware, and the purpose of this paper is to investigate the change rule of the effect of deformation on the bandgap value, so to save the computational resources, the next calculations are all calculated using the PBE. The Monkhorst-Pack special K-point sampling method is used in the calculations. The cutoff energy for the plane wave expansion is 400 eV, and the K-point grid is assumed to be 5 × 5 × 1. Following geometric optimization, the iterative precision converges to a value of less than 0.03 eV/Å for all atomic forces and at least 1 × 10-5 eV/atom for the total energy of each atom. The vacuum layer's thickness was selected at 20 Å to mitigate the impact of the interlayer contact force.

Laser writing of memristive logic gates and crossbar arrays

Memristor-based logic circuits are gaining a lot of attention due to the potential for high logic density hardware and novel in-memory computing applications. Readily available methods for fabricating of memristor logic structures that are suitable for integration with conventional computer hardware are a growing need. This work presents a direct laser writing process capable of rapidly fabricating memristor logic circuits by laser irradiation of metal salt precursor solutions. Planar memristor patterns are fabricated, and their I–V response is characterized. Boolean logic gates are fabricated from planar memristor pairs that exhibit low programming voltages and rapid switching. Cu/Cu2O/Cu and Ag/Cu2O/Cu memristors are also fabricated in crossbar arrays, showing the ability to be programmed to multiple resistance states through ultrashort voltage pulses. The devices also show the potential to have high endurance and nonvolatile resistance state retention.

Feed Forward Neural Network With Column-Wise Matrix–Vector Multiplication on FPGAs

This article presents a reconfigurable accelerator for recurrent neural networks with fine grained column wise matrix vector multiplication (RENOWN).We propose a novel latency hiding architecture for recurrent neural network accelerator using column wise matrix vector multiplication instead of the state of the art row wise operation. This hardware (HW) architecture can eliminate data dependencies to improve the throughput of RNN inference systems. Besides, we introduce a configurable checkerboard tiling strategy which allows large weight matrices, while incorporating various configurations of element-based parallelism (EP) and vector-based parallelism (VP). These optimizations improve the exploitation of parallelism to increase HW utilization and enhance system throughput. Evaluation results show that our design can achieve over 29.6 tera operations per second (TOPS) which would be among the highest for field-programmable gate array (FPGA)-based RNN designs. Compared to state-of-the-art accelerators on FPGAs, our design achieves 3.7–14.8 times better performance and has the highest HW.

A Hardware-Based Orientation Detection System Using Dendritic Computation

Studying how objects are positioned is vital for improving technologies like robots, cameras, and virtual reality. In our earlier papers, we introduced a bio-inspired artificial visual system for orientation detection, demonstrating its superiority over traditional systems with higher recognition rates, greater biological resemblance, and increased resistance to noise. In this paper, we propose a hardware-based orientation detection system (ODS). The ODS is implemented by a multiple dendritic neuron model (DNM), and a neuronal pruning scheme for the DNM is proposed. After performing the neuronal pruning, only the synapses in the direct and inverse connections states are retained. The former can be realized by a comparator, and the latter can be replaced by a combination of a comparator and a logic NOT gate. For the dendritic function, the connection of synapses on dendrites can be realized with logic AND gates. Then, the output of the neuron is equivalent to a logic OR gate. Compared with other machine learning methods, this logic circuit circumvents floating-point arithmetic and therefore requires very little computing resources to perform complex classification. Furthermore, the ODS can be designed based on experience, so no learning process is required. The superiority of ODS is verified by experiments on binary, grayscale, and color image datasets. The ability to process data rapidly owing to advantages such as parallel computation and simple hardware implementation allows the ODS to be desirable in the era of big data. It is worth mentioning that the experimental results are corroborated with anatomical, physiological, and neuroscientific studies, which may provide us with a new insight for understanding the complex functions in the human brain.

Implementation of Lightweight Convolutional Neural Networks with an Early Exit Mechanism Utilizing 40 nm CMOS Process for Fire Detection in Unmanned Aerial Vehicles.

The advancement of unmanned aerial vehicles (UAVs) enables early detection of numerous disasters. Efforts have been made to automate the monitoring of data from UAVs, with machine learning methods recently attracting significant interest. These solutions often face challenges with high computational costs and energy usage. Conventionally, data from UAVs are processed using cloud computing, where they are sent to the cloud for analysis. However, this method might not meet the real-time needs of disaster relief scenarios. In contrast, edge computing provides real-time processing at the site but still struggles with computational and energy efficiency issues. To overcome these obstacles and enhance resource utilization, this paper presents a convolutional neural network (CNN) model with an early exit mechanism designed for fire detection in UAVs. This model is implemented using TSMC 40 nm CMOS technology, which aids in hardware acceleration. Notably, the neural network has a modest parameter count of 11.2 k. In the hardware computation part, the CNN circuit completes fire detection in approximately 230,000 cycles. Power-gating techniques are also used to turn off inactive memory, contributing to reduced power consumption. The experimental results show that this neural network reaches a maximum accuracy of 81.49% in the hardware implementation stage. After automatic layout and routing, the CNN hardware accelerator can operate at 300 MHz, consuming 117 mW of power.

Special Section on Emerging Topics in Hardware Computing Systems Security

Special Section on Emerging Topics in Hardware Computing Systems Security

Comparative Study: Preemptive Shortest Job First and Round Robin Algorithms

Abstract: Operating system is a software acting as an interface between computer hardware and user. Operating system is known as a resource manager. The main responsibility of operating system is to handle resources of computer system. Scheduling is a key concept in computer multitasking and multiprocessing operating system design by switching the CPU among process. Shortest job first (SJF) and round robin are two wellknown algorithms in CPU processing. For shortest job first, this algorithm can be preemptived. In preemptive shortest job first, when a new process coming in, the process can be interupted. Where with round robin algorithm there will be time slices, context switching, or also called quantum, between process. In this journal we wil discuss comparative study between preemptive shortest job first and round robin algorithms. Three comparative studies will be discussed to understand these two algorithms more deeply. For all comparative study, the average waiting time and average turnaround time is more for round robin algorithm. In the first comparative study, we get average waiting time 52% more. For average turnaround time, 30% more. In second comparative analysis, we get 52 % average waiting time more and we get 35 % average turnaround time more. For third comparative analysis, average waiting time we get 50% more and for average turnaround time, we get 28% more. Thus it is concluded in our comparative study for these kind of data the preemptive shortest job first is more efficient then the round robin algorithm.
 
 Keywords: comparative study, premptive shortest job first algorithm, round robin algorithm, turn around time, average waiting time, time slice

Identification of Mineral Zircon (ZrSiO4) Rare Earth Metal (REM) As Smart Material For Electric Tactical Motor Bike Battery

Rare earth metals are associated minerals that are magnetic used as the main component in providing power to most electronic devices. Rare earth metal (REM) as a smart material that has 14 chemical elements. These elements are grouped according to their molecular weight, namely light rare earth metals (Ce, Pr, Nd, Pm, Sm, Eu) and heavy rare earth metals (Gd, Tb, Dy, Ho, Tr, Tm, Yb, Lu) . The use of this REM includes electric cars, health identification devices, LCDs and computer hardware. From some of the benefits and contents of REM, its development is not only in the industrial world but also in the world of defense, one of which is as a battery for motorized vehicles developed by the ministry of defense (Electric Tactical Motor Bike) that uses lithium batteries as a source of electrical energy. However, the use of lithium batteries which have shortcomings in terms of charging makes them look for other alternatives so that they can be charged optimally, one of which is using REM. The strategic position of rare earth minerals in the future must be developed so that they can be used sustainably, considering that Indonesia has potential resources that can be utilized.

PyDAP: Automated dental OPG beam area measurement using python and raspberry Pi camera

IntroductionThis study investigates if inexpensive computer hardware, open-source computer vision and a phosphor screen from disused CR (computed radiography) Cassette can be used to quantitatively assess beam shape and area. Materials and MethodsThe phosphor screen was affixed to a Carestream CS 8100 dental OPG system and the camera was mounted above the X-ray tube. Videos were acquired of the green light emissions during the tomographic irradiation. Images of a chessboard pattern, attached to the detector were used to correct for camera angulation and provide image pixel size calibration. K-Means colour clustering was used to define beam area. The effect of light conditions on beam dimension measurement was also investigated. The beam width measurement from optimised methodology was compared with that determined from dose calibrated GAFChromicTM XR-SP2film. ResultsVideos in dark conditions provided the most reproducible results. FW20M gained from initial sampling matched that obtained using the GAFChromicTM film within the errors of the measurements,6.41 ± 0.09 mm FW20M from this methodology,compared with FW20M (full width at 20 % of maximum) 6.4 ± 0.1 mm from film. The height and area were 126 ± 0.22 mm and 807 ± 11 mm2 respectively. The chess pattern imaging provided a robust means of perspective correction and pixel calibration. There is potential for this methodology to be employed using any digital camera, provided the camera acquisition settings remain constant, the sensor pixels are square, and the camera position is fixed. ConclusionThe potential of this low-cost open-source method of beam area measurement using computer vision is thus proven.

The application of reinforcement learning algorithms in intelligent games

The ongoing advancement of computer hardware has spurred the need for high-quality gaming experiences. The pressing need for intelligent games, the proven efficacy of reinforcement learning in various fields, the advancement of deep reinforcement learning technology, and the desire for customized gaming experiences collectively propel the ongoing exploration and implementation of reinforcement learning in intelligent games.This study use algorithms sourced from the literature review to examine multiple facets of reinforcement learning, including the modeling of game settings, reward design, the utilization of deep learning approaches, and adversarial learning. The utilization of reinforcement learning in intelligent games holds immense importance. Initially, it enhances the gaming encounter and renders intelligent games more demanding and tailored. Furthermore, the game gains complexity and intrigue by the implementation of intelligent adversaries capable of assimilating and adjusting to the player's tactics. This not only fosters the advancement of AI technology through research and development, but also broadens the utilization of reinforcement learning in other domains, like autonomous driving and smart manufacturing. Furthermore, tailored encounters, cutting-edge game design, and utilization in schooling are likewise significant contributions, rendering reinforcement learning a pivotal element in propelling technological advancement and augmenting user experience.

Pengaruh Motivasi Dan Semangat Kerja Karyawan Dalam Upaya Meningkatkan Produktivitas Pada PT. X Jakarta

PT. X is an official distributor company which operates in the field of providing laptop products and providing various kinds of computer hardware and software. The increasing mobility of people and the importance of devices to support their productivity has made the company want to increase PT. X's sales turnover, which is one of the company's goals. However, efforts to increase sales turnover can cause higher levels of employee pressure and decreased work morale. The increasing level of employee pressure and decreasing work morale at PT. X could be a negative impact of the company's efforts to increase sales turnover. Increased stress and decreased morale can become a dangerous cycle for a company. Increasing employee motivation and morale is an important responsibility for the leadership of an organization. A visionary leader needs to create a conducive organizational climate to maintain the morale and work dedication of his team members. The main goal of effective leadership is to inspire and motivate employees. This research uses quantitative methods, the samples in this research are all employees in the IT and Service division of PT. X, totaling 30 people. The research results show that based on the presentation of the results that have been presented, it can be seen that the Fcount value is more than Ftable (3.354) with a significance level of <0.05, so it can be concluded that the motivation and work enthusiasm variables together have an effect on productivity (Fcount 14.101 > Ftable 3.354). Bearti H3 is accepted, motivation and work enthusiasm have a simultaneously positive and significant influence on productivity.

Efficient Look-Up Table from Expanded Convolutional Network for Accelerating Image Super-resolution

The look-up table (LUT) has recently shown its practicability and effectiveness in super-resolution (SR) tasks due to its low computational cost and hardware independence. However, most existing methods focus on improving the performance of SR, neglecting the demand for high-speed SR on low-computational edge devices. In this paper, we propose an efficient expanded convolution (EC) layer, which expands the output size of regular convolution to enlarge the receptive field (RF) indirectly. It can increase the size of the LUT corresponding to the network linearly with the increase of RF. Additionally, after introducing the EC, multiple LUTs are merged into one LUT, achieving faster running speed while maintaining SR performance. More specifically, we expand the coverage of the convolutional output so that the output at the current position covers the target position and its surroundings, forming an overlapping sliding window at the output end. We sum up the overlapping parts of the sliding window as the output, thereby achieving the effect of enlarging the RF size. Moreover, by expanding the numerical range of the accumulated results and rescaling them to [0,255], the method can mitigate the error caused by quantization output. Experiments indicate that the proposed method performs better than the baseline method and is faster than other LUT-based SR methods.

Time transfer via single‐record TDoA measurements of GNSS satellites using direct cross‐correlation and relative pilot code phases

AbstractRadar systems are evolving towards distributed receiver networks. As individual stations might be separated too far to install a cable link, novel methods are required to synchronise individual data records in the time domain. State‐of‐the‐art GNSS receivers disciplining a highly stable oscillator are able to output a timing signal with several nanosecond accuracy solely using non‐proprietary signals. However, they typically require a stable environment and become a major cost factor for receiver networks with a high number of nodes. A method is presented to passively synchronise data records via GNSS raw signals in a single record requiring only a GNSS antenna, an analogue‐to‐digital converter and computation hardware. The clock bias is estimated via the common view method with either full raw signal correlation or software‐based code correlation of individual signals from the GPS, Galileo and BeiDou constellation with sub‐nanosecond precision.

An Overview of Procedurally Generating Virtual Environments

Procedural Generation is a method of using computers to generate content, allowing users to manually create content for virtual worlds is a very time consuming and laborious task, with the development of computer hardware, Procedural Generation is a good alternative, Procedural Generation content can be used as models or sketches to provide inspiration for the artist or to improve efficiency, this paper focuses on the representative algorithms used in Procedural Generation, which are commonly used to produce architecture, plants and terrain.

A universal parallel simulation framework for energy pipeline networks on high-performance computers

Energy distribution networks represent crucial infrastructures for modern society, and various simulation tools have been widely used by energy suppliers to manage these intricate networks. However, simulation calculations include a large number of fluid control equations, and computational overhead limits the performance of simulation software. This paper proposes a universal parallel simulation framework for energy pipeline networks that takes advantages of data parallelism and computational independence between network elements. A non-pipe model of an energy supply network is optimized, and the input and output of the network model in the proposed framework are modified, which can reduce the development burden during the numerical computations of the pipeline network and weaken the computational correlation between different simulated components. In addition, independent computations can be performed concurrently through periodic data exchange procedures between component instances, improving the parallelism and efficiency of simulation computations. Further, a parallel water pipelines network simulation computing paradigm based on a heterogeneous computer hardware architecture is used to evaluate the proposed framework’s performance. A series of tests are conducted to verify the accuracy of the proposed framework, and simulation errors of less than 5% are achieved. The results of multi-threaded simulation experiments have demonstrated the feasibility of the proposed framework in a parallel computing approach. Moreover, an Advanced Micro Devices (AMD) Deep Computing Unit (DCU)-parallel program is implemented into a water supply network simulation system; the computational efficiency of this system is compared with that of its serial counterpart. The experimental results show that the proposed framework is appropriate for high-performance computer architectures, and the 18x speed-up ratio demonstrates that the parallel program based on the proposed universal framework outperforms the serial program. That provides the basis for the application of pipe network simulation on high-performance computers.