Hardware Platform Research Articles

A Modified Control Strategy for Three-Phase Four-Switch Active Power Filters Based on Fundamental Positive Sequence Extraction

Three-phase four-switch active power filters (APFs) have attracted attention due to their low amount of semiconductors and low cost. The traditional control strategy of three-phase four-switch APFs usually includes phase-locked loops (PLLs) and rotating coordinate transformation for harmonic detection, resulting in complicated calculations and increased computation. In this paper, a modified control strategy for three-phase four-switch APFs based on fundamental positive sequence extraction is proposed, eliminating PLLs and rotating coordinate transformation with trigonometric calculations. Harmonic extraction is based on the fundamental positive sequence extraction method, while non-locked phase loop coordinate transformation is proposed to eliminate trigonometric calculations. Quasi-PR control is adopted for current tracking, and DC voltage control is designed to suppress voltage imbalance between the two split capacitors on the DC side. The space vector pulse width modulation (SVPWM) method is modified for a reduced-switch APF topology. The proposed control strategy guarantees excellent harmonic compensation: harmonic currents are significantly suppressed when the APFs are working, the THD of the source current decreases to 3.86%, the bus voltage fluctuation on the DC side is small, the voltage remains stable, and the computational complexity is reduced. Finally, a simulation and an experimental hardware platform are established to validate the feasibility and performance of the proposed control strategy. The experimental results show that it has good performance in suppressing harmonics and improving power quality.

A Survey on Design Space Exploration Approaches for Approximate Computing Systems

Approximate Computing (AxC) has emerged as a promising paradigm to enhance performance and energy efficiency by allowing a controlled trade-off between accuracy and resource consumption. It is extensively adopted across various abstraction levels, from software to architecture and circuit levels, employing diverse methodologies. The primary objective of AxC is to reduce energy consumption for executing error-resilient applications, accepting controlled and inherently acceptable output quality degradation. However, harnessing AxC poses several challenges, including identifying segments within a design amenable to approximation and selecting suitable AxC techniques to fulfill accuracy and performance criteria. This survey provides a comprehensive review of recent methodologies proposed for performing Design Space Exploration (DSE) to find the most suitable AxC techniques, focusing on both hardware and software implementations. DSE is a crucial design process where system designs are modeled, evaluated, and optimized for various extra-functional system behaviors such as performance, power consumption, energy efficiency, and accuracy. A systematic literature review was conducted to identify papers that ascribe their DSE algorithms, excluding those relying on exhaustive search methods. This survey aims to detail the state-of-the-art DSE methodologies that efficiently select AxC techniques, offering insights into their applicability across different hardware platforms and use-case domains. For this purpose, papers were categorized based on the type of search algorithm used, with Machine Learning (ML) and Evolutionary Algorithms (EAs) being the predominant approaches. Further categorization is based on the target hardware, including Field-Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), general-purpose Central Processing Units (CPUs), and Graphics Processing Units (GPUs). A notable observation was that most studies targeted image processing applications due to their tolerance for accuracy loss. By providing an overview of techniques and methods outlined in existing literature pertaining to the DSE of AxC designs, this survey elucidates the current trends and challenges in optimizing approximate designs.

Determinants of Financial Information Disclosure Online: Empirical Evidence in Portugal

The rapid advancement of information technology, particularly the Internet, has significantly transformed communication and corporate reporting practices. The evolution of the Internet and Information and Communication Technologies (ICT) has enabled companies to disseminate extensive financial and non-financial information quickly and cost-effectively. Typically published on company websites, this information is accessible to users worldwide. Digital financial reporting has evolved to meet the needs of both information users and companies. The Internet is increasingly the primary channel for companies to share financial information, driven by technological advancements and the growing demand for online company data. However, the rise of technology in financial reporting raises concerns about information overload, potentially diminishing the quality and relevance of economic data. Additionally, managing the costs associated with technology, such as website development, customer relationship management systems, hardware, software, and other digital platforms, is crucial. This study aims to determine the extent of online financial information disclosure among the 150 largest Portuguese companies and identify the factors influencing this disclosure. The research methodology involves content analysis of the websites of 52 of the largest and best-performing companies operating in Portugal, as ranked by Revista Exame, using a predefined disclosure index to evaluate the degree of financial information disclosure on the Internet. The objective is to enhance understanding of the financial reporting practices adopted by major Portuguese companies and identify factors determining the level of disclosure. The results confirm that the disclosure index has an average value of 0.52147 and that there is a relationship between company size and sector of activity and the level of internet financial reporting.

Dynamical analysis and hardware verification of a new multistable memristive hyperchaotic map

In this paper, a three-dimensional (3D) memristive hyperchaotic map is designed by coupling discrete memristor (DM). Structurally, the map features the existence of infinite fixed points. Based on the differential setting of parameters for various initial values, it can produce an infinite number of heterogeneous coexisting attractors or homogeneous coexisting attractors. Simultaneously, it can realize the attractors symmetric control. The numerical analysis and simulation results highlight its rich dynamical behaviors and excellent performance. A digital hardware platform based on a microcontroller is designed, clearly demonstrating the efficacy of the design through distinct visualization of attractors on an oscilloscope, providing tangible evidence of the successful hardware platform implementation.

Real-time fully digital control scheme for pulse coherent beam combining.

A fully digital control scheme for non-polarization-maintaining (non-PM) nanosecond pulse coherent beam combining (CBC) is proposed, where digital locking of optical coherence by single-detector electronic-frequency tagging (LOCSET) for active phase control and stochastic parallel gradient descent (SPGD) for active polarization control is proposed. The fully digital control scheme is integrated on a real-time field-programmable gate array (FPGA) empowered hardware platform and then experimentally validated in a four-channel all-fiber non-fully polarization-maintaining nanosecond pulse CBC system. Consequently, the system can be fully locked in 9.5 ms, and the polarization extinction ratio (PER) of the combined beam is 21.5 dB with a CBC efficiency of 95.3%. The fully digital control scheme integrates the advantages of digital LOCSET and multi-channel active polarization control, enhancing the channel scalability and the potential output power of the non-PM pulse CBC system.

MS-YOLO: A Lightweight and High-Precision YOLO Model for Drowning Detection.

A novel detection model, MS-YOLO, is developed in this paper to improve the efficiency of drowning rescue operations. The model is lightweight, high in precision, and applicable for intelligent hardware platforms. Firstly, the MD-C2F structure is built to capture the subtle movements and posture changes in various aquatic environments, with a light weight achieved by introducing dynamic convolution (DcConv). To make the model perform better in small object detection, the EMA mechanism is incorporated into the MD-C2F. Secondly, the MSI-SPPF module is constructed to improve the performance in identifying the features of different scales and the understanding of complex backgrounds. Finally, the ConCat single-channel fusion is replaced by BiFPN weighted channel fusion to retain more feature information and remove the irrelevant information in drowning features. Relative to the Faster R-CNN, SSD, YOLOv6, YOLOv9, and YOLOv10, the MS-YOLO achieves an average accuracy of 86.4% in detection on a self-built dataset at an ultra-low computational cost of 7.3 GFLOPs.

Joint Design of Transmit Waveform and Altitude for Unmanned Aerial Vehicle-Enabled Integrated Sensing and Wireless Power Transfer Systems

Recently, unmanned aerial vehicle (UAV)-enabled wireless power transfer (WPT) has received great attention as a promising technology for providing stable power to energy-constrained devices by navigating three-dimensional (3D) space, particularly in challenging environments such as maritime networks and smart cities. Additionally, UAV-enabled radar sensing has gained significant attention as a key technology for future 6G networks, as it enables high-accuracy sensing for various applications, such as target detection and tracking, surveillance, and environmental monitoring, as well as autonomous UAV operation. In this regard, we investigated UAV-enabled integrated sensing and wireless power transfer (ISWPT) systems that combine radar sensing and WPT operations on a unified hardware platform, sharing the same spectrum of resources. In order to accurately sense multiple targets and efficiently transfer power to multiple devices at the same time, we propose a method for jointly designing the transmit waveform and UAV altitude, taking into account the fundamental trade-off between radar sensing performance with the desired beam pattern and WPT performance with the desired harvested power of the devices. We first developed an effective method to obtain the optimal waveform and altitude by solving a challenging non-convex optimization problem. Based on this, we developed another efficient, low-complexity method by exploring a novel transmit waveform and optimizing its parameters to reduce computational complexity and thereby lower power consumption in UAVs. The numerical results verify that the proposed method significantly improves both radar sensing and WPT performance, as well as substantially reduces computational complexity.

Towards high-performance deep learning architecture and hardware accelerator design for robust analysis in diffuse correlation spectroscopy

This study proposes a compact deep learning (DL) architecture and a highly parallelized computing hardware platform to reconstruct the blood flow index (BFi) in diffuse correlation spectroscopy (DCS). We leveraged a rigorous analytical model to generate autocorrelation functions (ACFs) to train the DL network. We assessed the accuracy of the proposed DL using simulated and milk phantom data. Compared to convolutional neural networks (CNN), our lightweight DL architecture achieves 66.7% and 18.5% improvement in MSE for BFi and the coherence factor β, using synthetic data evaluation. The accuracy of rBFi over different algorithms was also investigated. We further simplified the DL computing primitives using subtraction for feature extraction, considering further hardware implementation. We extensively explored computing parallelism and fixed-point quantization within the DL architecture. With the DL model's compact size, we employed unrolling and pipelining optimizations for computation-intensive for-loops in the DL model while storing all learned parameters in on-chip BRAMs. We also achieved pixel-wise parallelism, enabling simultaneous, real-time processing of 10 and 15 autocorrelation functions on Zynq-7000 and Zynq-UltraScale+ field programmable gate array (FPGA), respectively. Unlike existing FPGA accelerators that produce BFi and the β from autocorrelation functions on standalone hardware, our approach is an encapsulated, end-to-end on-chip conversion process from intensity photon data to the temporal intensity ACF and subsequently reconstructing BFi and β. This hardware platform achieves an on-chip solution to replace post-processing and miniaturize modern DCS systems that use single-photon cameras. We also comprehensively compared the computational efficiency of our FPGA accelerator to CPU and GPU solutions.

On Optimizing Resources for Real‐Time End‐to‐End Machine Learning in Heterogeneous Edges

ABSTRACTDeploying end‐to‐end ML applications on edge resources becomes a viable solution to achieve performance and data regulations. With the microservice architecture, these applications can scale dynamically, improving service availability under dynamic workloads. However, orchestrating multiple end‐to‐end ML applications within heterogeneous edge environments must deal with numerous challenges while sharing computing resources. Prevalent orchestration tools/frameworks supporting edge ML serving are inefficient in provisioning methods due to constrained resources, diverse resource demands and utilization patterns. In this work, we present a provisioning method to optimize resource utilization for end‐to‐end ML applications on a heterogeneous edge. By profiling all microservices within the application, we estimate scales and allocate them on desired hardware platforms with sufficient resources when considering their runtime utilization patterns. We also provide several practical analyses on runtime monitoring metrics to detect and mitigate resource contentions, guaranteeing performance. The experiments with three real‐world ML applications demonstrate the practicality of our method on a heterogeneous edge cluster of Raspberry Pis and Jetson Developer Kits.

Molli: A General Purpose Python Toolkit for Combinatorial Small Molecule Library Generation, Manipulation, and Feature Extraction.

The construction, management, and analysis of large in silico molecular libraries is critical in many areas of modern chemistry. Herein, we introduce the MOLecular LIibrary toolkit, "molli", which is a Python 3 cheminformatics module that provides a streamlined interface for manipulating large in silico libraries. Three-dimensional, combinatorial molecule libraries can be expanded directly from two-dimensional chemical structure fragments stored in CDXML files with high stereochemical fidelity. Geometry optimization, property calculation, and conformer generation are executed by interfacing with widely used computational chemistry programs such as OpenBabel, RDKit, ORCA, NWChem, and xTB/CREST. Conformer-dependent grid-based feature calculators provide numerical representation and interface to robust three-dimensional visualization tools that provide comprehensive images to enhance human understanding of libraries with thousands of members. The package includes a command-line interface in addition to Python classes to streamline frequently used workflows. Parallel performance is benchmarked on various hardware platforms, and common workflows are demonstrated for different tasks ranging from optimized grid-based descriptor calculation on catalyst libraries to an NMR chemical shift prediction workflow from CDXML files.

Software-Defined Platform for Global Navigation Satellite System Antenna Array Development and Testing

With the increasing demand for accurate and robust positioning solutions, the use of GNSS antenna arrays has gained significant attention. However, their development and testing are frequently constrained by the inflexibility of traditional hardware platforms, often requiring extensive reconfiguration throughout the development cycle. This paper presents a platform based on a system on chip to develop a highly flexible software-controlled system that is capable of directly sampling up to 16 antenna elements. Multibeam digital beamforming is implemented using the available FPGA resources and the resulting signal is reproduced by the integrated DAC and can be connected to any conventional single antenna GNSS receiver. This paper presents the architecture of the platform, detailing its components and capabilities. Our experimental results demonstrate that the system can phase shift every channel with errors of less than 0.5° and can reconfigure 4 simultaneous beams of a 16-antenna array at speeds of 1.2 kHz, and 20 beams at around 400 Hz. The average delay introduced by each channel of the system is around 381 ns with a maximum deviation of 1.05 ns. The delay was also measured for the implementation using 4 beams, which achieves a slightly bigger average delay of 384.6 ns while keeping the variation to 5 to 6 ns. This system is intended to be used as the backbone for the development of antenna arrays and beamforming algorithms. Given its flexibility, it is not necessary to develop new hardware between development iterations or even for different systems, as only the software layer needs to be modified. Consequently, it is possible to expedite the development stage before producing dedicated solutions for industrial applications.

Introducing edge intelligence to smart meters via federated split learning

The ubiquitous smart meters are expected to be a central feature of future smart grids because they enable the collection of massive amounts of fine-grained consumption data to support demand-side flexibility. However, current smart meters are not smart enough. They can only perform basic data collection and communication functions and cannot carry out on-device intelligent data analytics due to hardware constraints in terms of memory, computation, and communication capacity. Moreover, privacy concerns have hindered the utilization of data from distributed smart meters. Here, we present an end-edge-cloud federated split learning framework to enable collaborative model training on resource-constrained smart meters with the assistance of edge and cloud servers in a resource-efficient and privacy-enhancing manner. The proposed method is validated on a hardware platform to conduct building and household load forecasting on smart meters that only have 192 KB of static random-access memory (SRAM). We show that the proposed method can reduce the memory footprint by 95.5%, the training time by 94.8%, and the communication burden by 50% under the distributed learning framework and can achieve comparable or superior forecasting accuracy to that of conventional methods trained on high-capacity servers.

Spaceborne particle identification platform and its application based on convolutional neural network

The identification of particle types in space radiation particle detection is a problem of great concern in scientific research and engineering applications. At present, the particle types are mainly identified by the difference in energy deposition of particles in the sensor, the difference in waveform in the sensor, the flight time and energy of particles, and the difference in the deflection path of particles in the electric field, such as traditional telescope detectors, particle pulse waveform analysis, flight time TOF system and electrostatic analyzer. However, the current on-orbit particle identification logic is relatively simple and the identification accuracy is limited. Convolutional neural networks have powerful target classification capabilities and are good at capturing and extracting target feature details, which can improve the accuracy of particle energy measurement and identification. Based on the environment commonly used in space environment detection payloads, this paper proposes a method for building an on-orbit convolutional neural network particle identification platform to achieve particle type identification. The platform first constructs a multi-dimensional input data set, completes the model training and weight derivation with the help of the software platform, and completes the waveform inference and data set expansion through the hardware platform. The established identification platform is used to train and test the neutron and gamma waveform data obtained in actual tests, and the accuracy of the identification of the software and hardware platforms is analyzed, completing the verification of the platform. The establishment and application of this platform provides a new idea and method for particle measurement and identification in future space environment detection, and has strong engineering practice significance.

An Online Data-Driven Method for Accurate Detection of Thermal Updrafts Using SINDy

Utilizing thermal updrafts shows potential for enabling long-endurance cruising of fixed-wing unmanned aerial vehicles without energy consumption. This article presents a novel online method based on sparse identification of nonlinear dynamics (SINDy) approach to achievement identification of thermal sources in the atmosphere. Initially, the algorithm is incorporated into the upper-level planning system, interacting with the lower-level controller. Then, experiments are conducted through software-in-the-loop simulations (SITL) to validate the implementation of the proposed algorithm. It is found that direct observation of thermal sources through measurements using SINDy is unfeasible during straight and circular flight modes. Nevertheless, simulation analysis of the proposed approach indicates that under unobservable conditions, a portion of the parameters can still be identified. By comparing results obtained using the particle filter algorithm, this method is shown to accurately estimate the parameters with negligible errors under observability conditions. The novelty of this approach lies in its significant improvement of the localization accuracy of the thermal source, without the need for parameter adjustments in the algorithm. Finally, the proposed methods are integrated into commonly used hardware platforms, and their online feasibility is verified through hardware-in-the-loop simulations.

Motion Control System for USV Target Point Convergence.

The goal of this paper is to establish a motion control system for unmanned surface vehicles (USVs) that enables point-to-point tracking and dynamic positioning. This includes the heading control and path following control of USVs. A hardware and software platform for USVs using microcontrollers is designed. This paper presents the construction of a kinematics and dynamics model for an unmanned catamaran. The motion process is divided into two segments. In the target point tracking segment, the heading coordinate system and the ship coordinate system are established. Based on these, a control method using differential steering to track the desired yaw angle is designed to improve the tracking efficiency. And the accuracy of heading keeping and path following is improved by combining the cascade PID controller. In the dynamic positioning segment, a self-adjusting mechanism is designed, thereby enhancing the flexibility of thrust distribution and improving the accuracy of the USV's positioning retention in wind and wave environments. Finally, experimental validation is carried out to verify the effectiveness of the design proposed in this paper by issuing control commands and saving the return data through the upper computer, and then analyzing the return data with MATLAB (R2022b, MathWorks, Natick, MA, USA).

ECG signal analysis and detection system based on deep learning

In view of the problems of large workload and easy error in traditional manual recognition of ECG signals, and the existing ECG monitoring equipment still has few ECG signal recognition types, low diagnostic accuracy, and excessive reliance on network services, in order to improve the performance of ECG monitoring equipment, an ECG (Electrocardiogram Signals) analysis and detection system is designed based on deep learning technology. The SENet-LSTM (Squeeze-and-Excitation Networks Long Short Term Memory) network model is built to realize automatic diagnosis of 7 categories of ECG signals. The model is built on an intelligent hardware platform that uses ADS1292R as ECG acquisition module, STM32F103 as data processing module, and Raspberry Pi as central processing module. The system uses the integrated high- performance microcomputer Raspberry Pi for calculation and analysis, providing users with offline artificial intelligence (AI: Artificial Intelligence) services. At the same time, the accuracy and precision of the model are and respectively 98.44%, 90.00%thereby realizing real-time detection and accurate classification of ECG, providing patients with accurate disease diagnosis.

A new two-parameter controllable multi-scroll 4D Hamiltonian conservative hyperchaotic system with improved nested COS-PWL function

A new two-parameter controllable multi-scroll Hamiltonian conservative hyperchaotic system is proposed in this paper. By utilizing an improved nested cosine-piecewise linear (COS-PWL) saturation function in combination with a 4D Hamiltonian conservative system, a 4D Hamiltonian conservative hyperchaotic system that can regulate multi-scroll according to two parameters is proposed. Compared with the nested Sine-PWL function, the system can produce hyperchaos. And the control parameter is more than just the saturation value E, the gain p is more sensitive to the wider control range of the system’s multi-scroll. The effect of the variation of the saturation value E and gain p of the control parameters on the phase trajectories of the multi-scroll is consistent with the behavior of the Lyapunov exponents (LEs) and bifurcations. Finally, we successfully constructed the system via the DSP hardware platform, demonstrating its practicality.

Extreme multi-stability and circuit implementation for a two-ReLU-memristor-based jerk oscillator

Memristor-based oscillation circuits are prone to produce coexisting infinite attractors depending on the initial conditions of memristors, leading to the appearance of extreme multi-stability. In this paper, we propose a novel memristive jerk oscillator by bringing two ReLU-type memristors in a simple jerk oscillator and investigate its dynamical behaviors associated with the coupling parameters using bifurcation plots and Lyapunov exponent plots. Further, we discuss the planar equilibrium state and its stability, and then numerically explore the coexisting infinite attractors driven by the initial conditions of two ReLU-type memristors. Because of the intervention of the two ReLU-type memristors, the memristive jerk oscillator has a planar equilibrium state whose stability closely relies on the initial conditions of two ReLU-type memristors, and different initial conditions cause different attractors to coexist, resulting in bidirectional extreme multi-stability. Finally, the memristive jerk oscillator is implemented by analog circuit and digital hardware platform, and the numerical results are confirmed by circuit simulations and hardware experiments.

PolyJuice: Detecting Mis-compilation Bugs in Tensor Compilers with Equality Saturation Based Rewriting

Tensor compilers are essential for deploying deep learning applications across various hardware platforms. While powerful, they are inherently complex and present significant challenges in ensuring correctness. This paper introduces PolyJuice, an automatic detection tool for identifying mis-compilation bugs in tensor compilers. Its basic idea is to construct semantically-equivalent computation graphs to validate the correctness of tensor compilers. The main challenge is to construct equivalent graphs capable of efficiently exploring the diverse optimization logic during compilation. We approach it from two dimensions. First, we propose arithmetic and structural equivalent rewrite rules to modify the dataflow of a tensor program. Second, we design an efficient equality saturation based rewriting framework to identify the most simplified and the most complex equivalent computation graphs for an input graph. After that, the outcome computation graphs have different dataflow and will likely experience different optimization processes during compilation. We applied it to five well-tested industrial tensor compilers, namely PyTorch Inductor, OnnxRuntime, TVM, TensorRT, and XLA, as well as two well-maintained academic tensor compilers, EinNet and Hidet. In total, PolyJuice detected 84 non-crash mis-compilation bugs, out of which 49 were confirmed with 20 fixed.

Denoising by neural network for muzzle blast detection

Acoem develops gunshot detection systems, consisting of a microphone array and software that detects and locates shooters on the battlefield. The performance of such systems is obviously affected by the acoustic environment in which they are operating: in particular, when mounted on a moving military vehicle, the presence of noise reduces the detection performance of the software. To limit the influence of the acoustic environment, a neural network has been developed. Instead of using a heavy convolutional neural network, a lightweight neural network architecture was chosen to limit the computational resources required to embed the algorithm on as many hardware platforms as possible. Thanks to the combination of a two hidden layer perceptron and appropriate signal processing techniques, the detection rate of impulsive muzzle blast waveforms (the wave coming from the detonation and indicating the position of the shooter) is significantly increased. With a rms value of noise of the same order as the muzzle blast peak amplitude, the detect rate is more than doubled with this denoising processing.