Processing Architecture Research Articles

An autocorrelation-less single-bit Weiner filter on FPGA

Filtering, transformations, and convolution in digital signal processing (DSP) are relatively resource-intensive algorithms, typically involving massive computations and complex processing architectures. Effective methods to manage this complexity include reducing the data word length using techniques such as sigma-delta modulation (SDM) and algorithmic optimization, either targeting coding style or via changes to the algorithm itself. In this paper, an SDM-based autocorrelation-less Weiner filter is proposed and compared with traditional multi-bit and SDM-based single-bit Weiner filters. The proposed design is first functionally verified using MATLAB simulations, and statistical parameters such as Signal to Noise Ratio (SNR), Mean Square Error (MSE), and Probability of Error (PE) are derived. The results show that the proposed design offers comparable performance to conventional filters. The proposed autocorrelation-less filter and equivalent conventional filters are then synthesized in FPGA for area-performance analysis. The SDM-based autocorrelation-less Weiner filter demonstrates a 30% improvement in performance over an equivalent traditional filter, which in result provides the shift in operating frequency from approximately 240 MHz to around 340 MHz, with about 63% less resources consumption. This improved performance and reduced resources increase the potential range of applications for this Weiner filter.

Serverless Architecture for Data Processing and Detecting Anomalies with the Mars Express MARSIS Instrument

The Mars Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS) on board Mars Express has been sampling the topside ionosphere of Mars since mid-2005. The analysis of the main reflection (nadir) of the ionosphere through the ionograms provided by the MARSIS instrument is typically performed manually due to the high noise level in the lower frequencies. This task, which involves pattern recognition, turns out to be unfeasible for the >2 million ionograms available at the European Planetary Science Archive. In the present contribution, we propose a modular architecture based on serverless computing (a paradigm that stands on the cloud) for optimal processing of these ionograms. In particular, we apply serverless computing to detect oblique echoes in the ionosphere, which are nonnadir reflections produced when MARSIS is sounding regions above or nearby crustal magnetic fields, where the ionosphere loses the spherical symmetry. Oblique echoes are typically observed at similar frequencies to the nadir reflections but at different times delays, sometimes even overlaying the nadir reflection. Oblique echoes are difficult to analyze with the standard technique due to their nonconstant and highly variable appearance, but they harbor essential information on the state of the ionosphere over magnetized regions. In this work we compare the proposed serverless architecture with two local alternatives while processing a representative data subset and finally provide a study by means of cost and performance.

Performance Improvement of Multi-Robot Data Transmission in Aggregated Robot Processing Architecture with Caches and QoS Balancing Optimization

Robot Operating System 2 (ROS 2) is a robotic software that uses a set of Quality of Service (QoS) policies to manage the quality of robot data transmissions in a network, such as the RELIABLE and KEEP_LAST options. In ROS 2 node communication, the RELIABLE connection guarantees that all message data can be properly sent from the publisher to the subscriber. However, strict reliability is not guaranteed if the RELIABLE connection uses the KEEP_LAST option to transmit the robot data in the publish–subscribe communication. This study aims to analyze the efficiency of local cache, cache control, and QoS balancing optimization to improve ROS 2 node communication when using the RELIABLE and KEEP_LAST options to transmit multi-robot data in Aggregated Robot Processing (ARP) architecture. Our idea in local cache and cache control is to streamline the sensor data output before processing it when the sensor device produces the data with the same value in a row. Furthermore, QoS balancing optimization aims to balance the DEPTH and DEADLINE QoS configuration to determine the rates and buffer size in ROS 2 node communication. This study shows that combining local cache and QoS balancing optimization improves multi-robot data transmission and cooperation in ARP architecture.

Adaptive parallel decision deep neural network for high-speed equalization.

The equalization plays a pivotal role in modern high-speed optical wire-line transmission. Taking advantage of the digital signal processing architecture, the deep neural network (DNN) is introduced to realize the feedback-free signaling, which has no processing speed ceiling due to the timing constraint on the feedback path. To save the hardware resource of a DNN equalizer, a parallel decision DNN is proposed in this paper. By replacing the soft-max decision layer with hard decision layer, multi-symbol can be processed within one neural network. The neuron increment during parallelization is only linear with the layer count, rather than the neuron count in the case of duplication. The simulation results show that the optimized new architecture has competitive performance with the traditional 2-tap decision feedback equalizer architecture with 15-tap feed forward equalizer at a 28GBd, or even 56GBd, four-level pulse amplitude modulation signal with 30dB loss. And the training convergency of the proposed equalizer is much faster than its traditional counterpart. An adaptive mechanism of the network parameter based on forward error correction is also studied.

Fourier transform-based electronic logic information processing architecture design

Abstract To better develop the electronic information industry and make the electronic logic information processing architecture more widely used in production life. This paper uses the Fourier transform model based on the discrete Fourier transform and fast Fourier transform data simulation, for the electronic logic information architecture of single-signal and dual-signal signal spectrum analysis for simulation analysis. The electronic logic signal, whether it contains whole harmonics, non-harmonics, or attenuated DC components, can be accurately decomposed even when the system frequency deviates from 50 Hz, and the error value is guaranteed to be around 0.0016%. It is also pointed out that the presence of the attenuated DC component does not affect the results. The simulation results show that the design of electronic logic information processing architecture based on Fourier transform is better than other models under the sample data conditions. And the Fourier transform model can effectively guarantee the probability strength of logic information transfer and enhance the reliable line of results, which provides a research direction for the design of electronic logic information processing architecture.

Nonlinear System Identification Using Audio-Inspired WaveNet Deep Neural Networks

Nonlinear system identification is a challenging task that requires accurate estimation of the structural model from observations of nonlinear behavior. The WaveNet, which was originally a neural network architecture for audio processing, has been modified and first introduced to the analysis of mechanical signals to capture long-term dependencies in mechanical systems and generate high-quality signals. A novel nonlinear system identification method has been proposed using a modified WaveNet-based approach that constructs a relationship between the vibration response and the nonlinear elements in the inverse model without the need for a definite structural model. This approach uses dilated convolution for feature extraction and a multilayer perceptron for feature transition, with the addition of average pooling along the time dimension for adaptive processing of varying length data, which are more computationally efficient and widely applicable. The 13-layer modified WaveNet models have been designed and applied to the problem. Comparisons with other baseline models were made to demonstrate the method’s superiority in terms of accuracy, effectiveness, and robustness. Additionally, the method has been applied to predict composite models of friction and elastic curves, demonstrating its ability to handle diverse and complex problems.

Development and Evaluation of a Multifrequency Ultrafast Doppler Spectral Analysis (MFUDSA) Algorithm for Wall Shear Stress Measurement: A Simulation and In Vitro Study.

Cardiovascular pathology is the leading cause of death and disability in the Western world, and current diagnostic testing usually evaluates the anatomy of the vessel to determine if the vessel contains blockages and plaques. However, there is a growing school of thought that other measures, such as wall shear stress, provide more useful information for earlier diagnosis and prediction of atherosclerotic related disease compared to pulsed-wave Doppler ultrasound, magnetic resonance angiography, or computed tomography angiography. A novel algorithm for quantifying wall shear stress (WSS) in atherosclerotic plaque using diagnostic ultrasound imaging, called Multifrequency ultrafast Doppler spectral analysis (MFUDSA), is presented. The development of this algorithm is presented, in addition to its optimisation using simulation studies and in-vitro experiments with flow phantoms approximating the early stages of cardiovascular disease. The presented algorithm is compared with commonly used WSS assessment methods, such as standard PW Doppler, Ultrafast Doppler, and Parabolic Doppler, as well as plane-wave Doppler. Compared to an equivalent processing architecture with one-dimensional Fourier analysis, the MFUDSA algorithm provided an increase in signal-to-noise ratio (SNR) by a factor of 4-8 and an increase in velocity resolution by a factor of 1.10-1.35. The results indicated that MFUDSA outperformed the others, with significant differences detected between the typical WSS values of moderate disease progression (p = 0.003) and severe disease progression (p = 0.001). The algorithm demonstrated an improved performance for the assessment of WSS and has potential to provide an earlier diagnosis of cardiovascular disease than current techniques allow.

Adaptation of FPGA architecture for accelerated image preprocessing

The work is devoted to the topical problem at the intersection of communications theory, digital electronics and numerical analysis, namely the study of image processing methods implementation time on different architectures of computational devices, which are used for software and hardware acceleration. The subject of this article is the investigation of reconfigurable FPGA processing systems in the image processing area. The goal of this work is to create a reconfigurable FPGA-based image processing system and compare it with existing processing architectures. Task. To fulfill the requirements of this work, it is necessary to prepare a practical experiment as well as theoretical research of the proposed architecture; to investigate the process of creating a ZYNQ SoC-based image processing system; and to develop and benchmark the speed of execution for the given set of algorithms with the specific range of the picture resolution. Methods used: FPGA simulation, C++ parallel programming with OpenMP, NVIDIA CUDA, performance analysis tools. The result of this work is the development of a resilient SoC Zynq7000–based computing system with programmable logic and the possibility to load images to FPGA RAM using the resources of ARM core for further processing and output via HDMI video interface, which enables the change of PL configuration at any time during the processing process. Conclusions. The efficiency of the FPGA approach was compared with a parallel image processing method implementation with OpenMP and CUDA. An overview of the ZYNQ platform with specific details related to media processing is presented. The analysis of algorithm speed testing findings based on various outputs proved the advantage (of over 60 times) of hardware acceleration of image processing over software analogs. The obtained results may be used in the development of embedded SoC-based solutions that require acceleration of big data processing. Also, the achieved findings can be used during the process of finding a suitable embedded platform for a certain image-processing task, where high data throughput is one of the most desired requirements.

Observation of Narrow Optical Homogeneous Linewidth and Long Nuclear Spin Lifetimes in a Prototypical [Eu(trensal)] Complex

Coherent light–matter interfaces are key elements for many quantum information processing architectures. Rare-earth ion (REI)-containing systems featuring narrow inter-configurational f–f transitions are suitable for creating such interfaces. Recently, narrow homogeneous linewidths (Γh) and long nuclear spin lifetimes (T1spin) associated with two Eu3+ complexes featuring the 5D0 → 7F0 f–f transition have been reported, elucidating that REI-molecule-based coherent light–matter interfaces can be obtained. In this study, we report a homogeneous linewidth of 2.8 MHz at 4.2 K, corresponding to an optical coherence lifetime (T2opt) of 114 ns (0.11 μs), associated with the 5D0 → 7F0 transition of the prototypical charge neutral Eu3+ complex─[Eu(trensal)]. Moreover, we have observed long nuclear spin lifetimes (T1spin) up to 460 ± 80 s at 4.2 K for the complex and efficient optical pumping of hyperfine levels of the 5D0 ground state. The results presented in this study indicate that narrow homogeneous linewidths and long nuclear spin lifetimes could be a generic property of molecular Eu3+ complexes featuring the 5D0 → 7F0 transition.

High-Throughput Single-Molecule Sensors: How Can the Signals Be Analyzed in Real Time for Achieving Real-Time Continuous Biosensing?

Single-molecule sensors collect statistics of single-molecule interactions, and the resulting data can be used to determine concentrations of analyte molecules. The assays are generally end-point assays and are not designed for continuous biosensing. For continuous biosensing, a single-molecule sensor needs to be reversible, and the signals should be analyzed in real time in order to continuously report output signals, with a well-controlled time delay and measurement precision. Here, we describe a signal processing architecture for real-time continuous biosensing based on high-throughput single-molecule sensors. The key aspect of the architecture is the parallel computation of multiple measurement blocks that enables continuous measurements over an endless time span. Continuous biosensing is demonstrated for a single-molecule sensor with 10,000 individual particles that are tracked as a function of time. The continuous analysis includes particle identification, particle tracking, drift correction, and detection of the discrete timepoints where individual particles switch between bound and unbound states, yielding state transition statistics that relate to the analyte concentration in solution. The continuous real-time sensing and computation were studied for a reversible cortisol competitive immunosensor, showing how the precision and time delay of cortisol monitoring are controlled by the number of analyzed particles and the size of the measurement blocks. Finally, we discuss how the presented signal processing architecture can be applied to various single-molecule measurement methods, allowing these to be developed into continuous biosensors.

Application of Multilevel Models in Classification and Regression Problems

There is a constant need to create methods for improving the quality indicators of information processing. In most practical cases, the ranges of target variables and predictors are formed under the influence of external and internal factors. Phenomena such as concept drift cause the model to lose its completeness and accuracy over time. The purpose of the work is to improve the processing data samples quality based on multi-level models for classification and regression problems. A two-level data processing architecture is proposed. At the lower level, the analysis of incoming information flows and sequences takes place, and the classification or regression tasks are solved. At the upper level, the samples are divided into segments, the current data properties in the subsamples are determined, and the most suitable lower-level models are assigned according to the achieved qualitative indicators. A formal description of the two-level architecture is given. In order to improve the quality indicators for classification and regression solving problems, a data sample preliminary processing is carried out, the model’s qualitative indicators are calculated, and classifiers with the best results are determined. The proposed solution makes it possible to implement constantly learning data processing systems. It is aimed at reducing the time spent on retraining models in case of data properties transformation. Experimental studies were carried out on several datasets. Numerical experiments have shown that the proposed solution makes it possible to improve the quality processing indicators. The model can be considered as an improvement of ensemble methods for processing information flows. Training a single classifier, rather than a group of complex classification models, makes it possible to reduce computational costs.

D2OP: A Fair Dual-Objective Weighted Scheduling Scheme in Internet of Everything

In times of the Internet of Everything (IoE), the power of the Internet is growing exponentially, followed by a surge in the number of network requests. The conflict between people’s high requirements for quality of experience (QoE) and limited computing resources are becoming increasingly prominent. Therefore, an appropriate offloading method is required to better ease this conflict. In this paper, a highly efficient scheduling architecture of information processing under the big data flow of the IoE is proposed to enhance the scheduling performance. First, we construct a dual-channel processing model to describe the entire data flow and node devices. Second, we carefully consider the choice of weighting method to better find a balance between dual objectives. Third, a dual-objective deep Q-network (DQN)-based offloading algorithm with principal component analysis weighting method (D2OP) is proposed to collaboratively minimize task response time and machine load in a more reasonable allocation. To verify the performance of the D2OP, a series of experiments are conducted from multiple angles. The experimental results demonstrate its better performance than the three comparison algorithms in reducing response time, load balance, and increasing task success ratio.

Exploring the sustainable development path of green financial system in the context of carbon neutrality and carbon peaking based on quality function deployment and G1-entropy methods: Evidence from China

In the 75th session of the United Nations General Assembly, the Chinese government first proposed the goal of carbon neutrality and carbon peaking. Since then, China’s economy and society have undergone a comprehensive green and sustainable development upgrade and transformation. The development of green finance can provide financial support for achieving dual carbon goals and mitigate the impact of climate change. More importantly, it can contribute to the national economy’s and society’s sustainable development. We innovatively draw on the quality function deployment theory in marketing to logically formulate the research idea of this paper. On this basis, we also apply the G1-entropy method from fuzzy mathematical theory for quantitative research. We innovatively address the actual national conditions in China and fully integrate green elements in constructing the index system from green finance and sustainability perspectives. Finally, we calculate index weights through G1-entropy quantification to assess the development quality of China’s green financial system and qualitatively propose countermeasures for the quality of China’s green financial development with respect to key index factors. Specifically, we sort out this paper in the following three aspects: (1) we innovatively combined the quality function deployment theory and built the quantitative analysis process architecture in this paper, which enhanced the readability of this paper (2) we realized the use of quantitative research for qualitative analysis and proposed the G1-entropy value method, which made up for the defects of the subjective and objective methods in the traditional assessment methods (3) we realized the organic combination of quantitative and qualitative analysis and proposed relevant countermeasure suggestions based on the quantitative index calculation results, which provided relevant countermeasure suggestions for promoting the sustainable and high-quality development of green finance in China. Our study will provide a set of perfect assessment methods for the quality improvement path and sustainable development strategy formulation after the construction of China’s future green financial system. It can also provide a reference assessment idea for the high-quality and sustainable development of China’s green finance, which will further help China’s economic transition to green and low-carbon and the achievement of the double carbon goal.

TRIM: A Design Space Exploration Model for Deep Neural Networks Inference and Training Accelerators

There is increasing demand for specialized hardware for training deep neural networks, both in edge/IoT environments and in high performance computing systems. The design space of such hardware is very large due to the wide range of processing architectures, deep neural network configurations, and dataflow options. This makes developing deep neural network processors quite complex, especially for training. We present TRIM, an infrastructure to help hardware architects explore the design space of deep neural network accelerators for both inference and training in the early design stages. The model evaluates at the whole network level, considering both inter-layer and intra-layer activities. Given applications, essential hardware specifications, and a design goal, TRIM can quickly explore different hardware design options, select the optimal dataflow and guide new hardware architecture design. We validated TRIM with FPGA-based implementation of deep neural network accelerators and ASIC-based architectures. We also show how to use TRIM to explore the design space through several case studies. TRIM is a powerful tool to help architects evaluate different hardware choices to develop efficient inference and training architecture design. Experimental results show that TRIM is a powerful tool for rapidly exploring the design space of DNN architectures for training and inference.

Research on Insulator Defect Detection System Based on Edge Intelligence

The paper aims at the problems of the high occupancy rate of communication resources and heavy server load caused by the image data taken during the inspection of the substation and sent to the remote server for processing. An edge intelligent inspection data processing architecture is constructed, and an insulator defect detection model based on transfer learning and YOLO v5 is proposed. The model can meet the training of a small sample network model in the case of scarcity of power equipment defect data. The proposed algorithm is compared with typical target detection algorithms, and the experimental results show that the proposed algorithm has an excellent performance in model recognition accuracy and recall rate. The trained insulator defect detection model is further deployed on the developed edge intelligent terminal for testing, and its image processing speed can meet the image processing requirements for intelligent inspection of substations. The system has good application value.

HALO: A Hardware–Software Co-Designed Processor for Brain–Computer Interfaces

Brain-computer interfaces (BCIs) enable direct communication with the brain, providing valuable information about brain function and enabling novel treatment of brain disorders. Our group has been building HALO , a flexible and ultra-lowpower processing architecture for BCIs. HALO can process up to 46 Mbps of neural data, a significant increase over the interfacing bandwidth achievable by prior BCIs. HALO can also be programmed to support several applications, unlike most prior BCIs. Key to HALO ’s effectiveness is a hardware accelerator cluster, where each accelerator operates within its own clock domain. A configurable interconnect connects the accelerators to create data flow pipelines that realize neural signal processing algorithms. We have taped out our design in a 12nm CMOS process. The resulting chip runs at 0.88V, peraccelerator frequencies of 3–180 MHz, and consumes at most 5.0mW for each signal processing pipeline. Evaluations using electrophysiological data collected from a non-human primate confirm HALO ’s flexibility and superior performance per watt.

Blockchain-as-a-Service for Business Process Management: Survey and Challenges

Blockchain technology (BCT) has brought a paradigm shift to Business Process Management (BPM). BCT provides a trusted decentralized infrastructure to secure data and process executions using distributed ledgers and smart contract to manage complex business processes. Numerous efforts have been made to exploit BCT in supporting dynamic and trusted collaborations of business processes. This paper aims to understand recent BCT development for its BPM applications and identify the limitations and challenges for further development via a systematic literature review (SLR). It is found that numerous works have reported using BCT as technical solutions to fulfill some traditional BPM functions. This paper is distinguished from existing works, especially several relevant surveys in the sense that (1) the impact of using BCT in BPM is thoroughly explored to identify new constraints and challenges explicitly brought by blockchains; (2) the requirements for Business Process Compliance (BPC) are firstly analyzed in detail. Note that BPC is to assure the adherence of business processes to pre-defined policies, standards, specifications, regulations, and laws when business processes are executed. To fill the gaps of BCT applications in these two aspects, Blockchain-as-a-Service (BCaaS) is adopted in business process architecture, and the trends of BCT developments are identified accordingly.

Examining a generic streaming architecture for smart manufacturing's Big data processing in Anomaly detection: A review and a proposal

The smart manufacturing industry has witnessed a rapid increase in data generation due to the integration of sensors, IoT devices, and other advanced technologies. With this huge amount of data, the need for efficient data processing methods becomes critical for identifying anomalies in real-time. With the rise of Industry 4.0 practices, digitally enabled manufacturing units are shifting their focus towards Smart Manufacturing paradigm for better productivity, throughput and increased business volume. Traditionally digital manufacturing units have considered different AI approaches like Neural Network, Statistical Methods, Deep Learning etc. to detect and predict anomalies in their production lines. But with the Smart Manufacturing ecosystem, a manufacturing unit must integrate manufacturing intelligence in real-time across entire production lines through sensor data of IOT devices. Hence the traditional anomaly detection systems fall short to respond well, under the changed scenario, where large volumes of unstructured and varied types of data are being generated at high velocity, to be processed at (soft) real time. The article reviews the current state-of-the-art in big data processing for anomaly detection in smart manufacturing. The review covers various aspects such as data collection, data processing, anomaly detection, and real-time monitoring. The current paper also proposes a novel stateful data streaming computational model for big data processing in smart manufacturing units which conceptually lays the foundation on top of which any discrete anomaly detection engine would be able to work. The proposed architecture has several benefits, including its ability to handle the large volume, velocity, and variety of data generated in smart manufacturing. The architecture can be applied to various smart manufacturing applications, including predictive maintenance, quality control, and supply chain optimization. It is expected that this proposed architecture will pave the way for the development of more efficient and effective smart manufacturing systems in the future.

Chatbot in Artificial Intelligence

Abstract: This paper discloses a virtual conversational method and system to relieve the psychological stress of adolescents. It also aims at providing some positive information through continuous dialogue answers in order to guide adolescents to think and face difficulties with a positive and optimistic attitude and the agenda of reliving the psychological pressure of the adolescents can be achieved. Conventional face-to-face stress detection and relief methods do not work when confronted with those adolescents who are reluctant to express their negative emotions to the people in real life. In this paper, we would like to present an adolescent- oriented intelligent conversational chatting system called “HappySoul”, which acts as a virtual friend who can assist to encourage, understand, comfort, and guide stressful adolescents to pour out their bad and negative feelings, thereby releasing the stress. Chatbots, or conversational interfaces, present a new way for adolescents to interact with computer systems. This chatbot will allow a user to simply ask questions in the same way that they would address a human. The technology at the core of the proposed chatbot is natural language processing (“NLP”), RNN and client server architecture with the help of Android GUI.

Functional architecture of executive processes: Evidence from verbal fluency and lesion mapping in stroke patients

The functional organization and related anatomy of executive functions are still largely unknown and were examined in the present study using a verbal fluency task. The objective of this study was to determine the cognitive architecture of a fluency task and related voxelwise anatomy in the GRECogVASC cohort and fMRI based meta-analytical data.First, we proposed a model of verbal fluency in which two control processes, lexico-semantic strategic search process and attention process, interact with semantic and lexico-phonological output processes. This model was assessed by testing 404 patients and 775 controls for semantic and letter fluency, naming, and processing speed (Trail Making test part A). Regression (R2 = .276 and .3, P = .0001, both) and structural equation modeling (CFI: .88, RMSEA: .2, SRMR: .1) analyses supported this model.Second, voxelwise lesion-symptom mapping and disconnectome analyses demonstrated fluency to be associated with left lesions of the pars opercularis, lenticular nucleus, insula, temporopolar region, and a large number of tracts. In addition, a single dissociation showed specific association of letter fluency with the pars triangularis of F3. Disconnectome mapping showed the additional role of disconnection of left frontal gyri and thalamus. By contrast, these analyses did not identify voxels specifically associated with lexico-phonological search processes.Third, meta-analytic fMRI data (based on 72 studies) strikingly matched all structures identified by the lesion approach.These results support our modeling of the functional architecture of verbal fluency based on two control processes (strategic search and attention) operating on semantic and lexico-phonologic output processes. Multivariate analysis supports the prominent role of the temporopolar area (BA 38) in semantic fluency and the F3 triangularis area (BA 45) in letter fluency. Finally, the lack of voxels specifically dedicated to strategic search processes could be due to a distributed organization of executive functions warranting further studies.