Proposed Method Research Articles

Spatial and temporal consistency learning for monocular 6D pose estimation

Monocular 6D pose estimation is a challenging task in the field of computer vision and robotics. Many previous works only input the cropped image of single object during training and inference, aiming to remove the noise from non-object regions. However, most of these methods ignore the viewpoint and spatial relationships of objects in the scene, which are crucial for accurate pose estimation of camera. To address this issue, this paper proposes a novel multi-view and multi-object based learning strategy for monocular 6D pose estimation, which involves the consistency of object coordinate for the same object at different viewpoints and the consistency of world coordinate for different objects in the same space. In the proposed method, the spatial and temporal groups are generated to trained the monocular 6D pose estimation network. Due to the camera motion, scene images taken at different times can be regarded as images captured from different viewpoints. Therefore, a temporal consistency loss is designed to constraint the relationship of the same object at different viewpoints, while a spatial consistency loss is designed to constraint the relationship of different objects at the same space. Finally, the proposed method is verified on the public datasets. Experimental results show that the proposed method is accurate, robust, and outperforms similar state-of-the-art approaches.

Interpretable synthetic signals for explainable one-class time-series classification

This research paper introduces an innovative approach for explainable one-class time-series classification (XOCTSC). The proposed method involves generating pseudounseen synthetic signals by altering the amplitude and cycle of the original signals. Subsequently, a classification process is performed to distinguish between the original and synthetic signals, and the resulting model is applied to testing data. Instances classified as synthetic classes are treated as unseen classes, and the dissimilarity with the training data can be elucidated through an explanation of the synthetic class creation process. This approach aims to enhance the interpretability of one-class time-series classification models by providing insights into the reasoning behind their decisions. The proposed method is demonstrated with a ballistocardiogram (BCG) signal for the breathing dataset and an electroencephalogram (EEG) signal for the epilepsy dataset. The proposed method recognizes BCG amplitude reduction during breath holding. Moreover, EEG cycle changes during epileptic seizures are observed across multiple channels. These observations align with actual epilepsy symptoms and breathing behaviors.

Multiple discrete crack initiation and propagation in Material Point Method

Cracks in MPM (CRAMP) is one of the most prominent discrete crack simulation methods in the Material Point Method (MPM) due to its simplicity and versatility. However, CRAMP is yet to include the capability to simulate concurrent crack initiations and propagations, as well as propagation to the edge of the material domain. The method proposed in this paper enables the simulation of multiple crack paths with CRAMP via the dynamic assignment of particles to separate grids while minimising the number of necessary grids. It also proposes methods of evaluating crack initiation and propagation via the Rankine criterion. The proposed methods are then implemented in an in-house Convected Particle Domain Interpolation (CPDI) MPM developed at Aalto University.To verify the integrity of the CPDI algorithm, our CPDI code with the proposed method implemented simulated a CPDI vortex. Furthermore, six fracture-simulation verification test cases were carried out: (1) through-crack in an infinite plate; (2) mode-I propagation; (3) initiation; (4) initiation with large deformations; (5) merging; (6) multiple initial cracks; and (7) radially-cracked thick ring. All these verification tests show successful initiation, propagation, merging, crack opening, and agreement with the results from the literature, as well as the convergence of various parameters with the expected rates.

An Estimation Method of High-Order LC Circuits in Power Electronic Converters

This paper proposes a parameter estimation method for high-order LC circuits conducted during the power-OFF of power electronic converters. It is based on a battery-powered compact signal injection unit and a circuit modeling method to estimate the LC circuit parameters. The proposed method aims for LC component condition monitoring in applications with routine-based maintenance or frequent ON/OFF operation (e.g., wind power, photovoltaic, traction systems). Compared to LCR meters and impedance analyzers, the proposed method has a significantly lower implementation cost and does not require disassembly of components. Moreover, it is a converter-level method in which the signal injection unit only needs to connect the converter inputs or outputs. Compared to power-ON LC parameter estimation methods, the proposed one is non-invasive without adding any hardware or software algorithm that may introduce new risks to converter operation. This method can handle the converters with high-order LC structures (e.g., <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$2^{nd}$</tex-math></inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$4^{th}$</tex-math></inline-formula> -order) in converter power-OFF state, which has not been achieved by existing power-OFF solutions. Therefore, the proposed method is promising for the specific applications of interest mentioned above. The parameter estimation algorithm and signal injection unit design are presented. The experimental results indicate that the estimation errors of the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$2^{nd}$</tex-math></inline-formula> -order equivalent circuit, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$3^{rd}$</tex-math></inline-formula> -order equivalent circuit, and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$4^{th}$</tex-math></inline-formula> -order equivalent circuit are lower than 1.6%, 2.5%, and 3%, respectively. The estimated LC parameters can be further used as inputs for operation optimization or condition monitoring of the converters.

QbD Assisted Development and Validation of UV Spectroscopic Method in Estimation of Silymarin

Background: Silymarin is a flavonoid utilised in liver dysfunction for years; new studies have emphasised its potential utility as a therapy for nanoparticulate targeting in many other disorders. In order to establish the product quality of such compounds, researchers have been trying to develop a robust method of analysis, but the methods developed till now are too expensive and time-consuming. Here we proposed quality by design-assisted development and validation of the UV spectroscopic method using Design-Expert® software in the estimation of Silymarin for fabrication of nanoparticulate formulations, which is simple, accurate, cost-effective and non-tedious. Aim: The proposed method is a simple, new, robust, accurate, and precise UV visible spectroscopic technique for estimating silymarin produced in nano-formulations utilising an analytical quality by design (AQbD) approach. Methods: A UV spectrophotometric technique was established, as maximum absorption (287.7) was measured using a Shimadzu UV-1800 double-beam UV visible spectrophotometer. The characterization of silymarin was done by melting point, DSC and FTIR techniques. The two critical method variables chosen were scanning speed and sample interval to be analysed by the design of experiment methodology utilizing the central composite design principle, which shows robustness and optimized technique involved in this work. Results: The spectroscopy technique was developed and validated as per International Conference of Harmonization recommendations. The Beer's-Lambert rule was followed in a series of 2-12 μg/ml dilution increments, with a correlation value of R2 = 0.999. The method's linearity was shown to be excellent across the concentration range. The percent recovery of the current method approach was determined to be within the confidential limitations, i.e., less than 2% expressed as % RSD, and the methodology was proved to be precise at inter and intraday variations (% RSD). The LOD and LOQ were found to be 0.264 μg/ml and 0.801 μg/ml which were also determined correctly. During specificity testing, no interfering peaks were found. Conclusion: This UV approach has been used successfully to determine the quantity of silymarin present in the nanoparticulate formulation, which can be used for testing its other pharmaceutical dosage forms.

A Simple Magnetic-Flux-Based Method for Localizing High-Resistance Stator Ground Faults in Synchronous Machines

This paper presents a new offline method based on the analysis of the magnetic flux to determine the location of stator ground faults in synchronous machines. First, the defective phase is identified via the insulation resistance (Megger) test, and then the longitudinal axis of the rotor is aligned with the faulty phase winding. During two measuring modes, a voltage is applied separately to each end of the faulty phase while the other end is floating and the core is grounded. The application of the voltage between each terminal of the faulty phase and ground will lead to the circulation of a current through the stator winding to the fault point, being its path closed by the grounded core. A magnetic flux will then be created, which can be sensed indirectly by measuring the induced voltage in the rotor field winding. The fault point can be identified by measuring the applied current or voltage and the magnetic flux created in both measuring modes. Since stator ground faults usually occur via high resistances, the method is extended for the detection of such faults by processing the data using the stationary wavelet transform. The experimental results show that the proposed method is very sensitive and is able to identify the ground fault location for fault resistances up to a few hundred of k-Ohms.

Active-Control-Based Three-Phase Reclosing Scheme for Single Transmission Line With PMSGs

Electric devices may be damaged due to the strong secondary impact when the circuit breaker recloses on permanent faults, so it is significant to identify the fault nature to reduce the harm of grid faults. To achieve this point, a new control scheme is designed to make permanent magnet synchronous generator (PMSG) based wind turbines behave as a three-phase symmetrical voltage source, so the overcurrent can be detected for permanent faults when the circuit breaker on the PMSG side is reclosed, but only the tiny capacitance current is present for temporary faults. Therefore, the fault nature can be identified by detecting the current amplitude. To accurately extract the power-frequency phasors, a fast matrix pencil extraction method is adopted. This method solves the difficulty in fault nature detection caused by the rapid disappearance of current and voltage after a three-phase trip using the controllability of power electronics. Meanwhile, the severe secondary impact can be reduced since the injected current is small even for permanent faults. Additionally, the proposed method can ensure the fault current at the inverter terminal is within the limiting value and operate properly for remote high-resistance faults. Extensive simulations and experiments are performed to verify this method.

A Series Impedance Reshaping Control Method Considering PLL Dynamics for Grid-Connected Inverters Under Weak Grid Conditions

Grid-connected inverters are the key part in renewable energy power generation systems. Usually, phase-locked loop (PLL) is adopted in grid-connected inverters to achieve synchronization. However, prior-art works have revealed that the negative resistance behavior in the low-frequency band caused by PLLs may lead to system instability under weak grids. In this paper, first, a comprehensive small-signal output impedance modeling of the LCL-filtered single-phase grid-connected inverter is derived for stability analysis, where the digital control delays and PLL dynamics have been taken into account. Then, a simple but effective virtual series based impedance reshaping method is proposed. With that, the inverter system can operate stably even under a very weak grid (e.g., short-circuit ratio (SCR) is as low as 1.2). Compared with the traditional grid impedance compensation by virtually increasing grid resistance or reducing the grid inductance, the proposed method does not require grid impedance estimation and is robust against system parameter variations. Thus, its feasibility is improved. Finally, experimental tests are carried out on a single-phase 5-kW prototype, validating the effectiveness of the proposed scheme.

Loose to compact feature alignment for domain adaptive object detection

Recently, great achievements have been made for deep learning based object detection methods. But their performance drops significantly when domain shifts occur. To address this problem, in this work we propose a loose to compact feature alignment method under an unsupervised domain adaptation framework. The entire feature alignment is performed in a divide and conquer manner, so as to distribute the alignment difficulties at two steps. At the first step, we loosen the goal at both image and instance levels. At the image level, a new Mask Guided Foreground Alignment (MGFA) module is proposed to make the alignment focus more on easier foreground regions, leaving the more diverse and more difficult background regions to the second step; at the instance level, we propose a Class-Wise Instance Alignment (CWIA) module with separated domain classifiers for different categories so as to ease the alignment. At the second step, the alignment is performed per pixel and per instance, achieving a more compact and better aligned feature space. We conduct experiments on three different adaptation scenarios, where we achieve comparable results, demonstrating the effectiveness of our proposed method.

Unsupervised Object-Centric Learning From Multiple Unspecified Viewpoints.

Visual scenes are extremely diverse, not only because there are infinite possible combinations of objects and backgrounds but also because the observations of the same scene may vary greatly with the change of viewpoints. When observing a multi-object visual scene from multiple viewpoints, humans can perceive the scene compositionally from each viewpoint while achieving the so-called "object constancy" across different viewpoints, even though the exact viewpoints are untold. This ability is essential for humans to identify the same object while moving and to learn from vision efficiently. It is intriguing to design models that have a similar ability. In this article, we consider a novel problem of learning compositional scene representations from multiple unspecified (i.e., unknown and unrelated) viewpoints without using any supervision and propose a deep generative model which separates latent representations into a viewpoint-independent part and a viewpoint-dependent part to solve this problem. During the inference, latent representations are randomly initialized and iteratively updated by integrating the information in different viewpoints with neural networks. Experiments on several specifically designed synthetic datasets have shown that the proposed method can effectively learn from multiple unspecified viewpoints.

Implementation of a non-linear SVM classification for seizure EEG signal analysis on FPGA

An automatic detection method for disease diagnosis plays an important role in the medical field. A computer-aided diagnosis (CAD) system for seizure detection using electroencephalogram (EEG) data is significant in protecting patients' lives. This paper presents machine learning algorithm-based seizure detection approach implemented in both software and hardware using EEG recordings. The input EEG signal is subjected to pre-processing, decomposition using wavelet transform and classification by a support vector machine (SVM) classifier. The maximum and variance statistical values of the decomposed EEG sub-bands are computed to generate a feature vector, while classification accuracy and computational complexity are taken into account. Gaussian Radial Basis Function (RBF) kernel based Non-Linear SVM (NLSVM) is used in this proposed approach to solve both two-class and three-class classification problems. In three-class classification, the one-against-one (OAO) technique is used. The competency of the proposed technique is evaluated using two distinct datasets provided by Bonn University, Germany & Neurology and Sleep Centre, New Delhi. The adequate parameters of the NLSVM classifier model are determined by using MATLAB software in the training phase and the best performing NLSVM model is selected for implementation on the Field-Programmable Gate Array (FPGA). The proposed seizure detection method is implemented with Verilog code and examined on a Xilinx Zynq-7000 XC7Z020 FPGA board to evaluate its performance. The experimental results show that the proposed method with a minimal number of features achieved a good performance metric for seizure detection and is comparable with that of the state-of-the-art methods.

Shadow Detection in Remote Sensing Images Based on Spectral Radiance Separability Enhancement.

Shadow detection is a basic task of remote sensing image analysis, but it is often seriously disturbed by vegetation, water bodies, and black objects. It is observed that vegetation and dark objects often show a dark look in visible bands but brighter in the near-infrared (NIR), and is also noticed that the reflection of inland water bodies in the green band is stronger than that in the blue band. Taking advantage of these physical properties and combining them with the bluish and dark appearance of shadows, we propose a simple but effective shadow detection method for multispectral remote sensing images. These physical properties are used to create transformation models that suppress features such as vegetation, water bodies, etc., but at the same time enhance shadows. Then, we transform the shadow representation into a color space to generate candidate shadows using dominant color components. To separate shadows from the others, we propose two indexes, the normalized Color Difference Composite Index (CDCI) and Color Purity Index (CPI), and fuse them to achieve shadows and their confidence. The experimental results indicate that the proposed method can effectively detect the shadows in multispectral images and outperforms the state-of-the-art approaches.

A Common-Mode Voltage Elimination Scheme by Reference Voltage Decomposition for Back-to-Back Two-Level Converters

Common-mode voltage (CMV) can greatly reduce the reliability and performance of back-to-back two-level converter-fed motor drive systems. To address this issue, a novel CMV elimination scheme based on reference voltage decomposition is proposed in this article. Firstly, the reference voltage of each phase is decomposed into two virtual reference voltages, resulting in six virtual reference voltages that are inter-related for back-to-back two-level converters. Then, by comparing each virtual reference voltage with a triangle carrier based on the two-level pulse width modulation (PWM), the final PWM voltage of each phase can be obtained from the differences of the corresponding two virtual PWM outputs. With the proposed method, both the amplitude and RMS value of CMV can be eliminated theoretically with satisfactory output waveform qualities. Simulations and experimental results are presented to verify the proposed method.

Video process detection for space electrostatic suspension material experiment in China’s Space Station

The purpose of video process detection is to identify key processes of interest and associated semantic categories, which is an important computer vision task for localizing key processes in space science experiments. To detect experimental processes in space electrostatic suspension material experiment on China’s Space Station, we propose a Slow-and-Fast Dual-Stream Network. It extracts multi-scale temporal semantic information by designing a Slow-and-Fast Stream structure and effectively detects the experimental processes with various temporal scales. In addition, the Two Path Temporal Pyramid Network (TP2N) enhances the bi-direction flow among multi-scale features on the temporal dimension to alleviate boundary sensitivity. The validation experiments on our established dataset demonstrate the effectiveness of our proposed method with 62.43% mAP, outperforming the related state-of-the-art temporal action localization methods with nearly 7% mAP gain. This paper is the first application of deep learning in space electrostatic suspension material experiments on China’s Space Station and further provides important technical support for the identification of key processes in space science experiments.

Decomposition by Approximation With Pulse Waves Allowing Further Research on Sources of Voltage Fluctuations

Voltage fluctuations are common disturbances in power grids. Initially, it is necessary to selectively identify individual sources of voltage fluctuations to take actions to minimize the effects of voltage fluctuations. Selective identification of disturbing loads is possible using a signal chain consisting of demodulation, decomposition, and assessment of the propagation of component signals. The accuracy of such an approach is closely related to the applied decomposition method. The paper presents a new method for decomposition by approximation with pulse waves. The proposed method allows for a correct identification of selected parameters, that is, the frequency of changes in the operating state of individual sources of voltage fluctuations and the amplitude of voltage changes caused by them. The article presents results from numerical simulation studies and laboratory experimental studies, based on which the estimation errors of the indicated parameters were determined by the proposed decomposition method and other empirical decomposition methods available in the literature. The real states that occur in power grids were recreated in the research. The metrological interpretation of the results obtained from the numerical simulation and experimental research is discussed.

A Novel High-Impedance Fault Detection Technique in Smart Active Distribution Systems

Identification of high impedance fault (HIF) is important in smart active distribution systems (SADSs). The existing protection methods fail to detect HIF, as HIFs have peculiar fault current characteristics. However, if HIFs are left undetected, they result in severe fire hazards and arcing. The paper proposes a novel fault detection indicator that utilizes the differential energy associated with the positive sequence current components during faults. IEEE-13 bus and IEEE-34 bus system is used for simulation in Real Time Digital Simulator (RTDS). The results show that proposed technique accurately identify HIFs and perfectly discriminates the transients due to HIFs, non-linear loads and voltage and reactive power compensation devices. Furthermore, it differentiates HIFs from other switching transients like capacitor or load switching events. Other factors, like transformer inrush currents, effect of noise and sampling rate are also considered to ensure the accuracy and reliability of the proposed method. Moreover, the practical feasibility of the proposed method is also experimentally validated.

Fourier-Based Vibration Model of Electrical Machines Considering Nonideal Orthogonality Between Electromagnetic Forces and Structural Modes

This paper proposes a Fourier-based model of vibration synthesis, improving both precision and speed. For the first time, the nonideal orthogonality between electromagnetic (EM) forces and structural modes is theoretically proposed and modelled by spatial harmonics. Different from the discrete vectors used in previous literature, the proposed method achieves a one-dimension (1-D) continuous model of EM forces and modes. On the basis, the vectors of modes and forces on tooth surfaces are decomposed by fast Fourier transform (FFT), so that the vibration response is synthesized based on Fourier coefficients. Further, the proposed model is applied to a 12-slot/10-pole (12s/10p) interior permanent-magnet (IPM) machine. The spatial harmonics of mode vectors are elaborated to identify the key orders of EM forces before the vibration response is completed. Then, the proposed model is compared with the coupled-field finite element analysis (FEA). The advantages are verified in terms of computational accuracy, speed, and sensitivity to mesh quality. Finally, the feasibility of the proposed model is verified by both coupled-field finite element analysis (FEA) and experiment.

Reconstructing Randomly Masked Spectra Helps DNNs Identify Discriminant Wavenumbers.

Nondestructive detection methods, based on vibrational spectroscopy, are vitally important in a wide range of applications including industrial chemistry, pharmacy and national defense. Recently, deep learning has been introduced into vibrational spectroscopy showing great potential. Different from images, text, etc. that offer large labeled data sets, vibrational spectroscopic data is very limited, which requires novel concepts beyond transfer and meta learning. To tackle this, we propose a task-enhanced augmentation network (TeaNet). The key component of TeaNet is a reconstruction module that inputs randomly masked spectra and outputs reconstructed samples that are similar to the original ones, but include additional variations learned from the domain. These augmented samples are used to train the classification model. The reconstruction and prediction parts are trained simultaneously, end-to-end with back-propagation. Results on both synthetic and real-world datasets verified the superiority of the proposed method. In the most difficult synthetic scenarios TeaNet outperformed CNN by 17%. We visualized and analysed the neuron responses of TeaNet and CNN, and found that TeaNet's ability to identify discriminant wavenumbers was excellent compared to CNN. Our approach is general and can be easily adapted to other domains, offering a solution to more accurate and interpretable few-shot learning.

Self-Calibration Method of Displacement Sensor in AMB-Rotor System Based on Magnetic Bearing Current Control

Displacement sensors are the key components of the active magnetic bearing rotor (AMB-rotor) system to realize the closed-loop control of the rotor displacement. The aging and failure of the displacement sensor components will lead to the changes of sensor sensitivity and zero point, which can directly affect the suspension performance of the rotor and even cause the instability of the AMB-rotor system. Therefore, it is significant to calibrate the sensor regularly. This paper proposes a self calibration method based on the detection and control of the magnetic bearing current without additional calibration equipment and hardware circuits. In addition, a new control structure of the AMB-rotor system is proposed to achieve efficient calibration, which can control the rotor displacement based on current following control method. Finally, experimental results verify the effectiveness of the proposed method on the displacement sensor calibration.

A Noncontact Method for Simultaneous Measurement of di/dt and dv/dt

This letter proposes a non-contact method for simultaneously measuring the current and voltage change rates by Rogowski coils. The dv/dt measurement capability is achieved based on electric field coupling, which is incorporated into the state-of-the-art Rogowski coil based current sensing solutions. The application of the proposed method for power semiconductor devices is demonstrated by a case study of an IGBT switching circuit. The Rogowski coils are integrated with a Printed Circuit Board (PCB) and fit with the IGBT module terminals. The comparable results from high-performance commercial instruments verify the di/dt and dv/dt measurement accuracy of the presented compact and cost-effective solution.