Search

Automatic driving technology has high accuracy and real-time requirements for pedestrian identification and localization. Pedestrian detection is a basic and necessary function in vision-based pedestrian detection systems and collision warning, which can effectively avoid traffic accidents and improve road driving safety to a certain extent. In this paper, a lightweight solution based on the FCOS-DEFPN model is proposed for real-time pedestrian detection. Based on the FCOS model, this paper proposes the FCOS-DEFPN model, which achieves the lightweight of the network by replacing the ResNet50 backbone network with the MobilenetV3 network and using the depth separable convolution instead of the ordinary convolution for parameter compression. While maintaining the detection accuracy, this paper introduces data enhancement methods such as Random Erasing and Morsia to simulate pedestrian occlusion and small target scenarios to improve the robustness of the model. For the pedestrian occlusion scenario, this paper introduces a lightweight attention network ECA, which helps to extract pedestrian features better. For small-target multi-scale pedestrians, the DEFPN feature pyramid network is proposed, which acquires feature information at multiple scales by attentional fusion of feature layers at different scales from top-down, bottom-up, and front-back. The experimental results show that the proposed model is enhanced in terms of detection accuracy for occluded and small-target pedestrians, and satisfies real-time pedestrian detection under the premise of robustness in complex scenes.

Pedestrian trajectory prediction is a key technology in surveillance systems and autonomous driving. However, due to the high uncertainty and dynamic spatial-temporal dependence of pedestrian movement, timely and accurate pedestrian trajectory prediction, especially long-term prediction, is still an open challenge. However, the existing models lack an effective modeling method for temporal dependence and spatial interaction modeling. To solve these problems, in this paper, we propose a novel paradigm of Sparse Transformer Networks with Spatial-Temporal Graph (STGSTN), which captures complex spatial-temporal interactions by stacking spatial-temporal Transformer blocks and improves the accuracy of pedestrian trajectory prediction by combining dynamic spatial dependence and long-range temporal dependence. We propose a new variant of sparse spatial transformer combined with graph neural networks, which uses the self-attention mechanism to dynamically model spatial dependencies to capture the state of pedestrian movement. The multi-head attention mechanism is also used to jointly model various patterns of spatial dependence. In addition, the sparse temporal transformer is used to learn the sparse attention map for time modeling, and then the long-range temporal dependence between pedestrians is modeled. Compared with the existing work, STGSTN can effectively and efficiently train long-range spatial-temporal dependencies. Experimental results show that STGSTN is competitive with state-of-the-art techniques on the ETH-UCY dataset, especially for long-term prediction. The performance of STGSTN may be limited by the quality of the training data, and its generalization to highly dynamic or crowded environments remains a challenge. Future work will focus on addressing these limitations and exploring additional contextual information to further enhance prediction accuracy.

The deep integration of Artificial Intelligence (AI) is gradually becoming a key force in innovating the teaching of English as a Foreign Language (EFL). This study aims to assess the practical effects of AI technology in providing customized instructional support and learning pathways in EFL instruction. The study reveals the benefits of AI in the instruction of English vocabulary, utilizing the Apriori algorithm from association rule mining and empirical analysis from survey data of 110 second-year university students across four different majors using AI-powered language learning platforms and AI-powered mobile language learning applications (such as UNIPUS AIGC platform and iTEST, intelligent assessment mobile application). It also deduces related teaching strategies and learning models. The results indicate that the use of AI-powered language learning platforms positively impacts English vocabulary learning outcomes in EFL instruction, and the combined use of AI-powered mobile language learning applications for self-testing and in-class tests effectively enhances vocabulary learning efficiency. The findings and conclusions of this study provide valuable insights for EFL educational practice and demonstrate the potential of AI in boosting the effectiveness of language learning, offering empirical support and guidance for future educational decision-making.

Dopaminergic treatment has proved effective to Parkinson’s disease (PD), but the conventional treatment assessment is human-administered and prone to intra- and inter-assessor variability. In this paper, we propose a knowledge-driven framework and discover that brain ACtivation-Transition-Spectrum (ACTS) features achieve effective quantified assessments of dopaminergic therapy in PD. Firstly, brain activities of fifty-one PD patients during clinical walking tests under the OFF and ON states (without and with dopaminergic medication) were measured with functional near-infrared spectroscopy (fNIRS). Then, brain ACTS features were formulated based on the medication-induced variations in temporal features of brain regional activation, transition features of brain hemodynamic states, and graph spectrum of brain functional connectivity. Afterwards, a prior selection algorithm was constructed based on recursive feature elimination and graph spectrum analysis for the selection of principal discriminative features. Further, linear discriminant analysis was conducted to predict the treatment-induced improvements. The results demonstrated that the proposed method decreased the misclassification probability from 42% to 16% in the evaluations of dopaminergic treatment and outperformed existing fNIRS-based methods. Therefore, the proposed method promises a quantified and objective approach for dopaminergic treatment assessment, and our brain ACTS features may serve as promising functional biomarkers for treatment evaluation.

Glaucoma is a group of diseases characterized by progressive optic nerve damage, ultimately resulting in irreversible visual impairment. Early diagnosis through color fundus photography, including measurement of the vertical cup-to-disk ratio (CDR), can help prevent vision loss. The normal range of CDR values is usually 0.3-0.5, and if it exceeds 0.6, then there may be some problems. However, asymmetrical thinning at the edges of the bottom-superior temporal-nasal region and large gaps in datasets pose challenges for existing automatic segmentation methods. To address these challenges, this paper proposes a joint segmentation method for the optic disc (OD) and optic cup (OC) based on an adversarial network, incorporating new monitoring functions to guide the network optimization process. The effectiveness and stability of this framework were evaluated using two public performance datasets of retinal fundus images, namely Drishti-GS and REFUGE. On the Drishti-GS dataset, our method achieved a score of 0.850/0.964/0.086, while on the REFUGE dataset, it obtained a score of 0.887/0.975/0.061. These results indicate the effectiveness of our approach.

Gaze estimation significantly enhances user interfaces, road safety, accessibility for the disabled, and consumer behavior analysis. Traditional methods depend on high-resolution images and ideal conditions, neglecting low-resolution scenarios. We propose GAFUSE-Net, designed for low-resolution gaze estimation. GAFUSE-Net includes a novel Facial Attention-Enhanced feature extraction (FAE) module and a Local image Super-resolution based Eye feature extraction (LSE) module for low resolution gaze estimation. The FAE module extracts gaze-relevant features from facial images. The LSE module uses super-resolution to enhance low-resolution eye images, followed by feature extraction with DeepEyeNet. Combined facial and eye features enable accurate gaze estimation. Experiments on the MPIIFaceGaze dataset show GAFUSE-Net’s effectiveness, achieving mean angular errors of 4.37°, 4.62°, 4.84°, and 6.19° at resolutions of $128\times 128$ , $64\times 64$ , $32\times 32$ , and $16\times 16$ , respectively. These errors are 0.2%, 1.3%, 5.7%, and 10.2% lower than state-of-the-art methods. The experimental results highlight the efficiency of GAFUSE-Net across various resolutions, significantly advancing the practical implementation of gaze estimation.

It is well known that distillation learning has the ability to enhance the performance of a light (student) model by transferring knowledge from a heavy (teacher) model, without incurring additional computational and storage costs. This article proposes an improved decoupled knowledge distillation (DKD) strategy for hyperspectral image (HSI) classification. A spatial feature blurring (SFB) module is designed to improve the classification performance of the student network when using DKD strategy. The SFB module utilizes randomly initialized 2-D standard normal distribution tensors to blur the spatial features of HSI, which increases the complexity of the data. This aligns with the characteristics of DKD, which transfers more useful knowledge under the condition of sample complexity. To effectively transfer knowledge, this article proposes a robust teacher network named the dual-branch spatial transformer-spectral transformer (DBSTST) network. This network describes the spatial and spectral long-range dependencies of HSI, addressing the limitations of convolutional neural networks in capturing only local features due to their fixed receptive fields. More specifically, the DBSTST network adopts spatial transformer-spectral transformer, which is composed of a parallel spatial-spectral multihead self-attention (PS2MHSA) module, aiming to describe pixel-level spatial long-range dependencies and spectral correlations in HSI. Simultaneously, the introduction of spatial-spectral positional embedding into PS2MHSA enhances positional awareness. We demonstrated the effectiveness of our proposed method on four publicly available HSI datasets. The student network achieves classification performance improvement and surpasses some other networks. Moreover, when compared with state-of-the-art classification methods, the DBSTST network also exhibits significant improvements in classification performance.

Superpixel segmentation is an essential step of object-oriented remote sensing image classification; the accuracy of the superpixel segmentation boundary will directly affect the classification result. Most of the traditional superpixel segmentation algorithms rely on spectral similarity and spatial connectivity to construct superpixels. They cannot find the accurate boundary in the complex scenes, such as the spatial distribution of ground features being relatively broken, and large differences in the size and shape, especially long-thin shape and circular shape. Aiming at this problem, a superpixel segmentation algorithm based on an anisotropic diffusion model named ADS is proposed and applied to image classification. The anisotropic diffusion model originated in thermodynamics has excellent properties in which the diffusion is continuous and smooth and its diffusion speed depends on the medium, which provides convenience for smoothing homogeneous regions and establishing boundary constraints for different ground objects. With this advantage, the diffusion flux model is established to consider the influence of boundary factors and used to simulate the dissimilarity measure with boundary constraints between pixels and seed points by combining the traditional spectral and spatial distance. Then, the seed points of superpixel are optimized under the K-means framework. The effectiveness of the proposed algorithm is tested and verified with different spatial resolutions, such as Landsat 8 with 30 m, Sentinel-2 with 10 m, and SkySat with 0.5 m. A large number of experiments show that the proposed algorithm can better correct the superpixel boundary-fitting deviation problem in complex scenes and effectively promote the improvement of image classification accuracy.

The development of modern science and technology has had a significant impact on the traditional teaching methods of higher education institutions. The utilization of technology-mediated learning approaches has provided universities and other higher education institutions with an opportunity to promote blended learning. This research aims to empirically analyze the learning effects of blended learning using the machine learning algorithm CART (Classification and Regression Tree) and regression analysis models, with Learning Management Service (LMS, using Rain Classroom as an example in this paper). It explores whether different dimensional factors in blended learning on the Rain Classroom platform are associated rules for English proficiency, while specifically focusing on their correlation with scores in the English proficiency test (CET-4, College English Test Band 4). The study identifies which association rules within these factors contribute to the improvement of students’ scores. Additionally, it classifies students based on their scores and learning effects using a decision tree structure. Based on the classification results, a multiple linear regression model is constructed to quantitatively assess the learning effects of using Rain Classroom for English language learning among students with different English proficiency levels. These analytical results enable educators, during the later stages of using Rain Classroom for blended learning, to adjust teaching decisions based on individual situations of students in different categories. They can design teaching materials and activities tailored to each student, leveraging the tool capabilities of Rain Classroom as an LMS and the advantages of blended learning. This approach aims to create more profound and meaningful learning experiences for students, assisting them in achieving educational goals and unlocking their full potential in academics.

Rectangular orthogonal digital filter banks (ODFBs) based on square-root-raised-cosine (SRRC) functions are widely utilised to realise flexible and elastic multi-channel aggregations for fixed and mobile networks. However, long digital filter lengths are required to minimize digital filtering-associated signal distortions. In this paper, based on the extended Gaussian function (EGF), a novel rectangular ODFB with excellent robustness against the short digital filter length-induced truncation effect is proposed. Optimum digital filter parameters of the EGF-based ODFBs are identified and verified in multi-channel hybrid OFDM-digital filter multiple access (DFMA) PONs based on intensity modulation and direct detection (IMDD). By making use of the identified optimum digital filter parameters, extensive comparisons of digital filter characteristics and corresponding multi-channel upstream PON performances are made between the EGF-based ODFBs and the SRRC-based ODFBs. It is shown that to achieve a similar aggregated upstream signal transmission capacity, the EGF-based ODFB reduces the digital filter DSP complexity by a factor of 4. For a digital filter length as short as 8, in comparison with the SRRC-based ODFB, the EGF-based ODFB introduces >1.5 dB (>0.8 dB) improvements in upstream receiver sensitivity for 5-bits (8-bits) DACs/ADCs, increases the aggregated upstream signal transmission capacity by >5.5%, enlarges the ONU launch power dynamic range by >2.5 dB and improves the frequency offset tolerance by a factor of >1.5. In addition, the EGF-based ODFB also enhances upstream performance robustness against ONU symbol timing offsets. [ABSTRACT FROM AUTHOR]

In PON-based mobile fronthauls, direct inter-ONU communications without passing end-user traffic to the OLT offer a promising solution for fulfilling the stringent latency and bandwidth requirements of 5G and beyond networks. In this paper, with slight modifications to the PON remote node, a concurrent inter-ONU and upstream communication technique is proposed and experimentally demonstrated in 101.6Gbit/s multipoint-to-point hybrid SSB OFDM digital filter multiple access (DFMA) IMDD PONs over 25km SSMFs. Multiple gapless inter-ONU and upstream SSB signals are aggregated by digital orthogonal filtering in each ONU transmitter. A single FFT operation is applied for demultiplex and demodulation in the OLT/ONU receivers. The results show that for both the inter-ONU and upstream transmissions, the optimum length of digital filters is 32, based on which the power penalties due to the fiber transmission and ONU channel interference are <1dB and <2dB, respectively. For the inter-ONU communications, adaptive RF spectral assignments can effectively mitigate the Rayleigh and Brillouin backscattering effects and the upstream channel fading effect, thus giving rise to >30% improvements in aggregated signal transmission capacity. In addition, detailed experimental investigations are also undertaken of the trade-off between differential ONU optical launch power dynamic range and aggregated signal transmission capacity. An approximately 1dB increase in ONU launch power dynamic range is achievable by reducing the aggregated signal transmission capacity by 5Gbit/s. [ABSTRACT FROM AUTHOR]

Vehicle trajectory prediction is a key technology in autonomous driving, one of the aims of which is to ensure the safety of vehicle travelling and lane changing. Especially in complex traffic scenarios, it is critical to accurately predict the vehicle trajectory so that the automated driving system can make appropriate response actions to help the vehicle driver avoid accidents to the greatest extent possible. Some studies on vehicle trajectory prediction have used recurrent networks (LSTMs) to extract temporal correlations and additional convolutional neural networks (CNNs) to capture spatial correlations. This hybrid model can be added more operations, so we propose a new CNN-LSTM hybrid model containing multi-module, interaction regions and spatial-temporal attention mechanisms (IA-CSTM). In this paper, the traffic scene is divided into multiple interaction regions, and vehicles in different regions have different impacts on the predicted vehicles. The model has multiple modules that deal with vehicle information in different interaction regions. The model will cross-fuse vehicle information in important interaction regions to enhance the model’s perception of vehicle information. Multiple modules are linked through a spatial-temporal attention mechanism and capture the relevance of the vehicle in the interaction region to predicted vehicle.

The synthetic aperture radar (SAR) oriented ship detection is a more complex and finer task than the horizontal, and it provides ship orientation for military and civilian applications, such as trajectory prediction. However, many oriented target detection methods discarded high-resolution (low-level) feature map, which provides morphological features for predicting accurate bounding box for small ships, due to computational cost and lack of semantic information. In addition, most methods follow the samples assignment in horizontal detection after padding oriented bounding box. The inaccurate geometric relationship causes that some low-quality positive samples without contributing accuracy improvement are introduced. To improve the precision of small oriented ship detection, we proposed the feature pyramid network contained high-resolution feature to fuse high-level semantics and highlight contours and textures of ships in the low-level feature map. To select high-quality positive rotated samples, dynamic rotated positive samples mining (DRPSM) is designed to mine hard positive samples for oriented ship detection specially. In one word, a two-stage and anchor-free detector trained by DRPSM and utilized high-resolution feature map is proposed for detecting oriented ships in SAR imagery. Subsequently, extensive experiments have been conducted on two oriented SAR datasets: the SSDD and the HRSID. As a result, our method attains recall of 96.2% and 90.1%, and AP50 of 90.7% and 86.5% on SSDD and HRSID, respectively. It shows that our detector achieves state-of-the-art results applied in capturing ships, when compared with other popular remote sensing oriented detectors.

Block adjustment technology can effectively correct the interferometric parameter errors acquired from Interferometric Synthetic Aperture Radar (InSAR) system. It can reduce the number of required ground control points and improve the accuracy of the generated digital elevation model, making it’s an effective means for obtaining three-dimensional information from large-scale terrain information. However, the currently used adjustment models usually only involve surface elevation parameters, which means that the accuracy of the terrain planimetric information entirely relies on the Position and Orientation System (POS) parameters. This may lead to the inaccurate planimetric to be a fundamental weakness which hinder the practical application of this technology. To solve these problems, this paper proposed a DSM (digital surface model) mapping scheme by using very high-resolution synthetic aperture radar (SAR) data that combines the block adjustments of orientation and interferometric parameters. Firstly, the interferometric and orientation parameters are obtained separately by using InSAR and SAR block adjustment techniques. After optimizing the model parameter accuracy, the elevation information obtained from the interferometric parameters is incorporated into the SAR positioning model to acquire the three-dimensional information of the area accurately. Experimental processing was conducted using 23 pairs of airborne InSAR data from two strips in the Weinan area of China. The experimental results demonstrate that the proposed scheme can improve the elevation accuracy by approximately 20%.

In this letter, a high performance quasi-vertical GaN-on-Si Schottky barrier diode (SBD) was fabricated by combing in-situ p-GaN layer with hydrogen plasma treatment and controlled diffusion as edge terminations (ETs), where the main junction region of Schottky contact is emerged by pattern etching the top p-GaN layer. The p-n junction with higher barrier height can screen the Schottky contact from high electric field induced by field crowding effect. The hydrogen plasma treatment and controlled diffusion applied to the p-GaN layer further reduces the electric field crowding effect near the p-n junction edge via a gradient junction termination extension structure. As a result, the breakdown voltage of the edge p-GaN terminated SBD with H treatment and diffusion is significantly boosted from 140 V in conventional SBD to 500 V with optimized fabrication process. The on/off current ratio of the diode is as high as 1011@±3 V. The forward current density is 1.45 kA/cm2@3 V and the specific on-resistance is 1.57 $\text{m}\Omega \cdot $ cm2.

In recent years, large numbers of jellyfish have congregated in marine areas, leading to a decline in other plankton and fisheries. Jellyfish themselves have a certain toxicity and aggression, which have a serious impact on the safety of human life. In order to detect the quantity and distribution of underwater jellyfish, and to be more proactive in the prevention and control of Aurelia outbreaks, this study proposed a method for in-situ detection of underwater jellyfish based on the improved Faster R-CNN network model. Firstly, the real data sets of three species of jellyfish in the Qinhuangdao sea area were established by using underwater high-definition camera. The Multi Scale Retinex with Colour Restoration (MSRCR) algorithm was used to improve the brightness and contrast of the underwater images. Secondly, the residual network Resnet50 was integrated into the backbone network for better feature extraction; then the semi-precision floating-point number FP16 was added to improve the training speed. Finally, comparative experiments were conducted to verify the improved network. The F1 value, the P-R curve, the Loss curve and the AP value of the three detection models were evaluated and compared. The experimental results showed that compared to Vgg16 network and YOLO V3 network, the training speed was improved from 1.85bit/s to 7.35bit/s, and the accuracy was also improved to over 0.98. The experimental results were good, and the research results provided a more accurate and faster method for the in-situ detection of underwater jellyfish.

When ac and dc microgrids get exposed to load power changes or voltage transients, instability could occur owing to constant power loads’ (CPLs) behaviour. CPL has the negative impedance behaviour which is the reason behind such instability. A flexible alternating current transmission system (FACTS) devices are used in ac grid to help improve the power quality under such circumstances. Contrarily, dc networks doesn’t have such devices and suffers under similar situations. In this paper, the T-Type Modular Dc Circuit Breaker (T-Breaker) is proposed to act as a compensator device for dc networks when energy storage is integrated. Its compensation functions in addition to breaking and limiting a system’s current can be achieved using this single device. By implementing the T-Breaker’s shunt (current) and series (voltage) compensations, ride-through of grid transients can be achieved. T-Breaker’s series compensation function and how it can be integrated with the shunt compensation to improve the stability of a dc microgrid is the focus in this paper. In addition to modeling and simulating the system, both small signal and large signal stability are analyzed. Furthermore, a scaled down 270 V 5 kW system is utilized to validate T-Breake’s compensation functions and their influence on a microgrid’s stability.

Parkinson’s disease (PD) is a prevalent brain disorder, and PD diagnosis is crucial for treatment. Existing methods for PD diagnosis are mainly focused on behavior analysis, while the functional neurodegeneration of PD has not been well investigated. This paper proposes a method to signify functional neurodegeneration of PD with dynamic functional connectivity analysis. A functional near-infrared spectroscopy (fNIRS)-based experimental paradigm was designed to capture brain activation from 50 PD patients and 41 age-matched healthy controls in clinical walking tests. Dynamic functional connectivity was constructed with sliding-window correlation analysis, and k-means clustering was applied to generate the key brain connectivity states. Dynamic state features including state occurrence probability, state transition percentage and state statistical features were extracted to quantify the variations of brain functional networks. A support vector machine was trained to classify PD patients and healthy controls. Statistical analysis was conducted to investigate the difference between PD patients and healthy controls as well as the relationship between dynamic state features and the MDS-UPDRS sub-score of gait. The results showed that PD patients had a higher probability of transiting to brain connectivity states with high levels of information transmission compared with healthy controls. The MDS-UPDRS sub-score of gait and the dynamics state features showed a significant correlation. Moreover, the proposed method had better classification performances than the available fNIRS-based methods in terms of accuracy and F1 score. Thus, the proposed method well signified functional neurodegeneration of PD, and the dynamic state features may serve as promising functional biomarkers for PD diagnosis.

Recent advancements in immersive virtual reality head-mounted displays allowed users to better engage with simulated graphical environments. Having the screen egocentrically stabilized in a way such that the users may freely rotate their heads to observe virtual surroundings, head-mounted displays present virtual scenarios with rich immersion. With such an enhanced degree of freedom, immersive virtual reality displays have also been integrated with electroencephalograms, which make it possible to study and utilize brain signals non-invasively, to analyze and apply their capabilities. In this review, we introduce recent progress that utilized immersive head-mounted displays along with electroencephalograms across various fields, focusing on the purposes and experimental designs of their studies. The paper also highlights the effects of using immersive virtual reality discovered through the electroencephalogram analysis and discusses existing limitations, current trends as well as future research opportunities that may hopefully act as a useful source of information for further improvement of electroencephalogram-based immersive virtual reality applications.

Journal editors are putting a lot of effort into selecting appropriate reviewers for fair and reliable peer review of submitted manuscripts. Editors consider whether the reviewers have no affinity with any of the authors of manuscripts and have sufficient expertise in reviewing the manuscripts. The affinity can be evaluated by whether any of the reviewers has been a coauthor and/or a coworker in a common institution with any of the authors of the manuscript. The expertise depends on the similarity of the research topic between the reviewer’s published papers and the submitted manuscripts. In this paper we propose an algorithm to recommend appropriate reviewers to editors, based on the assessment of these scholarly activities and achievements. To implement this algorithm, TextRank and GenSim library are used to extract feature sets from abstract and introduction sections of both submitted manuscripts and the reviewer candidates’ papers. Based on the extracted feature sets, affinity and expertise check are implemented. An experiment has been conducted with a data set of over 1,000 papers in the field of DB research to evaluate the performance of this algorithm. The experiment consists of affinity check by using 2-mode network matrix operations and expertise check based on Max Similarity and/or topic clustering. Experimental results show that the recommendation algorithm is reasonable on the basis of scholarly activity assessment and achievements. Our method is designed to optimize the reviewer pool for each journal, client and this algorithm is designed with an open license and has the advantage of being free to use and operate. In addition, our algorithm has the advantage of journal-specific optimization by designing to adjust the expertise of the reviewer selection and the weights for exclusion of interests to reflect journal-specific policies.

In silicon carbide (SiC) planar insulated-gate bipolar transistor (IGBT), a large distance between neighboring p-bodies is beneficial to enhance the on-state conductivity modulation, but will expose the gate oxide to high electric field in off-state. With p-bodies placed closer, the gate oxide field is reduced, but the conductivity modulation is suppressed. In this work, a new SiC planar IGBT with oxide shield is proposed and studied by TCAD simulations. The proposed SiC IGBT achieves improved trade-off between on-state voltage drop $(V_{\mathrm{ ON}})$ and maximum gate oxide electric field $(E_{\text {ox-m}})$ . When a quite larger distance between neighboring p-bodies is adopted in the proposed SiC IGBT, a low $V_{\mathrm{ ON}}$ is obtained, while the $E_{\text {ox-m}}$ can be kept at a small value with the oxide shielding structures protecting the gate oxide. Switching characteristics are also studied, and the proposed SiC-IGBT delivers much better trade-off between turn-off energy loss $(E_{\mathrm{ OFF}})$ and $V_{\mathrm{ ON}}$ than the conventional SiC planar IGBT.

We propose a circuit topology suitable as a battery charge/discharge tester with a DAB converter and a non-isolated dc-dc converter as a module structure. The module structure can be configured to have a wide input and output voltage range because it is easy to expand. DAB converters are used for bidirectional power transfer and galvanic isolation. By controlling the phase difference of the voltages across the DAB to be almost identical, the generation of circulating current due to the phase difference is minimized, and the ZVS is guaranteed. In addition, by switching at a phase angle that flattens the DAB switch current, the peak current is reduced to diminish the switch’s conduction loss, lowering the switch current rating, and reducing the magnetic core size. The non-isolated dc-dc converter is used to control the battery output voltage. Through the interleaved structure, good dynamic characteristics and current ripple can be reduced, so the effect of improving the battery lifecycle can be expected. We present four series-parallel structures of the proposed converter and compare them with the case of generating an output voltage only with a DAB converter. Analyze the characteristics of automatic voltage balancing due to the difference between the input capacitors’ initial voltage and capacitance connected in series. We analyze the features of the proposed system through simulation and verify the technical feasibility and excellence of the proposed battery charge/discharge tester through charge/discharge experiments using supercapacitors instead of batteries.

This paper presents a robust force/torque-based impedance control with a perturbation observer, particularly for high payload systems. One of the special cases of such systems is an assistive robot for paraplegics (ARP) with a sit-stand mechanism (STSM) and, therefore, its control is investigated for enhanced efficiency and safety. Recently, a nonlinear robust position control known as sliding mode control with sliding perturbation observer (SMCSPO) was implemented on STSM and proven its effectiveness. However, with increasing subject body mass (payload), high switching control gains are required which consequently either result in increased tracking error or chattering in control performance. In this regard, a new framework of integrated control was proposed which combines the feature of impedance control with estimated information from SMCSPO resulting in an enhanced control algorithm. The advantage of this impedance control is that because of SMCSPO, it utilizes the linear system information with position feedback only to estimate the states and perturbation (nonlinearities, uncertainties, and disturbance). This estimated perturbation then helps in compensating the actual perturbation effect from the system response. Moreover, the impedance control excludes the saturation function which helps prevent the chattering with low control gains. The simulation and experiment results validated that the proposed algorithm shows better transient response with minimized estimation error and reduced chattering effect than SMCSPO.

Herein, we introduce “Look and Diagnose” (LAD), a hybrid deep learning-based system that aims to support doctors in the medical field for diagnosing effectively the Benign Paroxysmal Positional Vertigo (BPPV) disorder. Given the body postures of the patient in the Dix-Hallpike and lateral head turns test, the visual information of both eyes is captured and fed into LAD for analyzing and classifying into one of six possible disorders which the patient might be suffering from. The proposed system consists of two streams: (1) an RNN-based stream that takes raw RGB images of both eyes to extract visual features and optical flow of each eye followed by ternary classification to determine left/right posterior canal (PC) or other; and (2) pupil detector stream that detects the pupil when it is classified as Non-PC and classifies the direction and strength of the beating to categorize the Non-PC types into the remaining four classes: Geotropic BPPV (left and right) and Apogeotropic BPPV (left and right). Experimental results show that with the given body postures of the patient, the system is capable of accurately classifying given BPPV disorder into the six types of disorder with an accuracy of 91% on the validation set. The proposed method can successfully classify disorders with an accuracy of 93% for the Posterior Canal disorder and 95% for the Geotropic and Apogeotropic disorder, paving a potential direction for research with the medical data.

Fresh Fruit Bunch (FFB) is the main ingredient in palm oil production. Harvesting FFB from oil palm trees at its peak ripeness stage is crucial to maximise the oil extraction rate (OER) and quality. In current harvesting practices, misclassification of FFB ripeness can occur due to human error, resulting in OER loss. Therefore, a vision-based ripe FFB detection system is proposed as the first step in a robotic FFB harvesting system. In this work, live camera input is fed into a Convolutional Neural Network (CNN) model known as YOLOv4 to detect the presence of ripe FFBs on the oil palm trees in real-time. Once a ripe FFB is detected on the tree, a signal is transmitted via ROS to the robotic harvesting mechanism. To train the YOLOv4 model, a large number of ripe FFB images were collected using an Intel Realsense Camera D435 with a resolution of $1920\times 1080$ . During data acquisition, a subject matter expert assisted in classifying the FFBs in terms of ripe or unripe. During the testing phase, the result of the mean Average Precision (mAP) and recall are 87.9 % and 82 % as the detection fulfilled the Intersect over Union (IoU) with more than 0.5 after 2000 iterations and the system operated at the real-time speed of roughly 21 Frame Per Second (FPS).

In the past few years, aircraft detection in remote sensing (RS) images has been an important research hotspot, and it is very crucial in plenty of military applications. Based on the high computational cost of the model and numerous parameters, deep convolution neural networks-based algorithms have excellent performance in the aircraft detection task. However, it is still difficult to detect aircraft due to the complex background of RS images, various types of aircraft, and so on. In addition, it is difficult and costly to make labels for satellite-based optical RS images. Consequently, we propose an end-to-end lightweight aircraft detection framework called CGC-NET (a network based on circle grayscale characteristics), which can accurately detect aircraft with a few training samples. There are only a small number of trainable parameters in CGC-NET, which greatly reduces the need for large datasets. Extensive evaluations indicate the excellent performance of CGC-NET, in which the F-score can reach 91.06% and the model size is only 0.88 M. Therefore, CGC-NET can be used to accurately detect aircraft targets simply and effectively.

Many applications are equipped to utilize drones to provide various services to users in Internet of Drones (IoD) environments. In such applications, it is necessary to make a session key between a drone and a user to establish an authenticated and secure channel. It is also desirable to provide strong anonymity to increase user(drone) privacy. To provide robust anonymity, a protocol has to provide both pseudonymity and unlinkability. If a protocol provides only pseudonymity without unlinkability, user(drone) privacy could be breached by analyzing communication frequency or user(drone) movement. On the other hand, we consider drone capture attacks in the IoD, because if a drone is captured, the secret information of the drone could be revealed. To minimize the damage against the capturing attacks, a key exchange should provide the forward unlinkability as well as forward secrecy. Forward unlinkability means that even though the secret information of the drone is revealed, the unlinkability is guaranteed. In the paper, we suggest the first key agreement protocols providing both pseudonymity and forward unlinkability, whereas previous key agreement protocols provide only pseudonymity and unlinkability.

Robot manipulators have complex dynamics and are affected by significant uncertainties and external disturbances (perturbations). Consequently, determining the exact mathematical model of a robot manipulator is a tedious task. Accordingly, accurate trajectory tracking is a dominant feature in the design of position control for robot manipulators. The main objective of this research is to design a robust and accurate position control for a robot manipulator despite the absence of an exact model. For this purpose, an extended state observer (ESO)-based sliding mode control (SMC) is proposed. The main concept in the ESO is to define and estimate the assumed perturbation, which includes known and unknown system dynamics, uncertainties, and external disturbances. Additionally, the ESO estimates the system states. This estimated perturbation information, which is combined with the SMC input, is utilized as a feedback term to compensate for the effect of the actual perturbation. The perturbation compensation improves the controller performance, resulting in a slight position error, less sensitivity to perturbation, and a small switching gain required for the SMC. The advantage of the proposed algorithm is that it only requires partial state information and position feedback. Thus, identifying the system parameters for the nominal model before designing the controller is unnecessary. The proposed algorithm is implemented and compared with the conventional SMC and the SMC with a sliding perturbation observer (SMCSPO) in a virtual environment based on MATLAB SimMechanics. The comparison results validate the robustness of the proposed technique in the presence of perturbation and show that the technique has a significantly reduced trajectory tracking error than the conventional SMC and SMCSPO.

Accurate prediction of mesoscale eddy trajectories requires efficient models with large-size available data instances to capture the main eddy characteristics. However, there is a lack of salient attention mechanisms that can recognize the demand of extracting the aggregated features over multidimensional eddy data sources. Additionally, deep learning techniques are very important for eddy trajectory prediction to dynamically capture properties of mesoscale eddies in the South China Sea. In this article, we propose a spatio–temporal attention-based deep learning framework that can orchestrate heterogeneous data integration and propagation trajectory forecast together. It consists of a novel autoencoder equipped with channel and spatial attention mechanisms (CSA-encoder), and a gated recurrent unit (GRU) network with temporal attention layer (TA-GRU). CSA-encoder compresses stereoscopic eddy data with convolutional layers and generates the small-scale and high-quality dataset as the input of TA-GRU. The finer grained TA-GRU method is extended to accurately predict eddy trajectories with more valuable imagery information so that the temporal attention mechanism can automatically select relevant regions within the next 14 days. Our cross-validation results demonstrate that our framework averagely achieves a lower distance error (9 km) and 54% performance improvement over the baseline GRU technique in the next one day, and outperforms two state-of-the-art techniques of long short-term memory and recurrent neural network by 54.9% and 65.6%, respectively.

There are lots of image data in the field of remote sensing, most of which have low-resolution due to the limited image sensor. The super-resolution method can effectively restore the low-resolution image to the high-resolution image. However, the existing super-resolution method has both heavy computing burden and number of parameters. For saving costs, we propose the feedback ghost residual dense network (FGRDN), which considers the feedback mechanism as the framework to attain lower features through high-level refining. Further, for feature extraction, we replace the convolution of the residual dense blocks (RDBs) with ghost modules (GMs), which can remove the redundant channels and avoid the increase of parameters along with the network depth. Finally, the spatial and channel attention module (SCM) is employed in the end of the RDB to learn more useful information from features. Compared to other SOTA lightweight algorithms, our proposed algorithm can reach convergences more rapidly with fewer parameters, and the performance of the network can be markedly enhanced on the image texture and object contour reconstruction with better peak signal-to-noise ratio (PSNR) and structural similarity (SSIM).

A discrete GaN power transistor’s substrate is typically connected to its source electrode. However, on the GaN-on-Si power IC platform, the high-side transistor (HS-transistor) and low-side transistor (LS-transistor) share a common substrate that cannot be simultaneously connected to both source electrodes of the two transistors. Thus, the termination of the common substrate remains an undecided issue. In this work, comprehensive TCAD simulations are exploited to reveal the influences of various substrate termination schemes. It is found the common substrate inevitably leads to severe degradation in the dynamic ${R} _{\mathrm{ ON}}$ due to back-gating effects. The mechanisms for the degradations vary with the substrate termination scheme, and will be discussed in detail. To address these issues, we propose a new GaN power IC platform on an engineered bulk silicon substrate, and study the new platform with TCAD simulations. The proposed platform provides a local electrical substrate (a p+ island) for each GaN power transistor. The source electrode of each GaN transistor is connected to its local electrical substrate, while all devices share a common mechanical substrate. The junctions between the local substrates and the underlying n-layer provide an effective isolation between GaN transistors. The back-gating effects are completely suppressed for the GaN integrated bridge circuit.

Till now, most calibration methods only compensate geometric error caused by inaccurate kinematic parameters, while the desired accuracy may still not be achieved when the robot is performing long-stroke, heavy-duty loading and unloading tasks. In this paper, a generalized semi-analytical beam deformation model is firstly proposed for the Cartesian robot to compensate the non-geometric error due to structural deformation under both distributed and concentrated loads. The adjustment factors are introduced in this model to deal with the over-constrained boundary conditions for both intermediate and side modules of the long-stroke Cartesian robot. This improves the fitness of the coming geometric error model, which assumes the beam to be rigid and straight. In addition, as the major error components which do not conform to the Gaussian distribution have been extracted in earlier steps, the Gaussian process regression model is imported to predict the residual error more accurately. In this way, comprehensive geometric and non-geometric error modeling and compensation procedures are formed for the multi-module, long-stroke Cartesian robot. Simulations and real-time experiments are conducted to validate the effectiveness of the proposed method.

Medium-voltage (e.g., 10 kV rated) silicon carbide (SiC) devices have great potentials in medium-voltage variable speed drives. But their high switching dv/dt can increase the voltage stress on motor windings and cause partial discharges. This paper presents a partial discharge study of a medium-voltage form-wound winding under two-level square-wave voltage pulses. A 10 kV SiC device-based test platform is built to generate voltage pulses with high dv/dt. A three-step test approach is proposed and employed to systematically investigate the effects of various voltage parameters on partial discharges. These include voltage rise/fall time, voltage pulse width, pulse repetitive rate, duty ratio, voltage polarity, fundamental frequency, and modulation index. Partial discharge inception voltages (PDIVs) and repetitive partial discharge inception voltages (RPDIVs) of the sample are measured with varied voltage parameters. Test results show that voltage rise/fall time is a major affecting factor which reduces PDIVs of the winding sample by 6.5% when it decreases from 800 ns to 100 ns. Based on test results, a hypothetical partial discharge mechanism is presented to explain the effects of fast voltage rise/fall edges. An empirical equation is also derived to estimate PDIVs of a winding sample under various voltage rise/fall time and pulse width conditions.

Considering the compensation of thermo-optic coefficients (TOC) between GaN-based cavity and TiO2 photonic crystal (PhC) layer, we have theoretically designed a photonic crystal surface emitting laser (PCSEL) with temperature-insensitive wavelength. The structure includes a freestanding GaN membrane with embedded multiple InGaN/GaN quantum well (QW) layer and the top TiO2 PhC layer. The high refractive index (RI) of PhC and the freestanding membrane cavity structure cooperate to enable a stronger field-coupling in PhC, thus fulfill the athermal coupling condition. Interestingly, the wavelength temperature dependence is only 0.0015 nm/°C for such PCSEL with a 82% PhC confinement factor, and its athermal property is proportional to the field confinement factor in TiO2 PhC layer. To illustrate, a general formula has been developed after the investigations of thermo-optic effect and thermal expansion effect.

GaoFen-4 (GF-4) infrared spectrum (IRS) camera has a special spectral band located in the mid-wave infrared spectrum with the wavelength ranging from 3.50 to 4.10 μm. It is very important for forest fire-detection and other quantitative remote sensing applications at a high spatial resolution of 400 m. Due to the similarity of the spectral characteristics, the thermal emissive band 20 (3.660–3.840 μm) and 22 (3.929–3.989 μm) of medium resolution imaging spectrometer (MODIS) are taken as the reference bands for the radiometric cross-calibration of the GF-4/IRS camera. The time difference between selected GF-4/IRS and MODIS images at the Qinghai Lake and Se lincuo Lake test sites is less than one hour. After spatial matching, single-band and dual-band spectral matching methods are proposed in this article. Then, the DN value of GF-4 image is linearly fitted to the radiance of the corresponding MODIS image to obtain the radiometric calibration coefficients of GF-4/IRS camera. The validation results show that the cross-calibration coefficients calculated by integrating the measurement of two MODIS bands with the mean value spectral matching approach has the highest radiometric calibration consistency with the official calibration coefficients.

As a fundamental component of heatmap-based human pose estimation methods, heatmap decoding is to transform heatmaps into joint coordinates. We found that previous heatmap decoding methods generally ignored the effect of systematic errors introduced by the resolution increaseing operations in the network decoder. This work fills the gap by taking the systematic errors in heatmap decoding into consideration. We proposed a fast method to reduces the systematic and random errors in one shot by error compensation. The proposed method outperforms the previous best method on the COCO and the MPII datasets while being over 2 times faster. Extensive experiments with different networks, resolutions, metrics and datasets have proved the rationality of the proposed idea.

Based on the common existence of the mixed road, a switched controller with consideration of the difference between estimation optimal and current flux (EOCFD) is presented in the lattice hydrodynamic model of traffic flow. Based on the Hurwitz criteria and the $H_{\infty }$ -norm, stability conditions for the curved road and straight road scenarios are obtained with the transfer functions $G_{0}$ and $G_{1}$ respectively. By analyzing the Bode-plot of transfer functions $G_{0}$ and $G_{1}$ , stability analysis is performed with the feedback gain $k$ , the radian $\theta _{j}$ and the curvature radius $R$ . These theoretical results indicate that the switched control scheme prompts the traffic flow to be more stable in both curved and straight roads. Compared with Cheng’s model, numerical simulations with multiple perturbations confirm that the switched control scheme can further suppress the traffic congestion with a lower feedback gain on mixed road.

In view of the fact that the existing behavior recognition algorithms cannot fully extract abstract behavior features, this paper proposes a SE-R3D-LSTM behavior recognition algorithm based on 3D residual convolutional neural network (R3D), which integrates Squeeze-and-excitation network (SENet)and long short-term memory (LSTM). First of all, a residual module is added to the 3D Convolutional Neural Network (3D-CNN) to avoid problems such as gradient dispersion caused by the deepening of the network layer; Secondly, not only the global average pooling layer but also the global maximum pooling layer is used in the SENet network, which can fully extract global information and achieve feature calibration. In the meantime, expand the SENet network to three-dimensional, which can make the connection of the spatiotemporal feature channels closer. Afterwards, the 3D-SE module is introduced into the R3D network, which can enhance the effective spatiotemporal features and suppress the invalid spatiotemporal features; Since, because LSTM can perform timing modeling on high-level features and learn more effective feature information, the LSTM network is introduced into the SE-R3D network. Finally, Softmax is used for classification. Experimental results show that the recognition rate of the SE-R3D-LSTM network on the UCF101 data set reaches 96.5%.

Due to the rapid development of the Intelligent Traffic System (for short, ITS), the shared information among successive vehicles can be acquired accurately and reliably. In this paper, a control scheme utilizing the successive flux difference under honk environment (SFDHE) is investigated in the lattice hydrodynamic model. Based on the Hurwitz criteria and the $H_{\infty }$ -norm, the sufficient condition for the transfer function is derived with the control theory analysis. The Bode-plot of transfer function indicates that the stability of the traffic flow is increased significantly with the novel controller. Moreover, the numerical simulation results of the new model are compared with that of Peng’s model, which demonstrates that traffic jams can be suppressed efficiently when the SFDHE control signal is taken into account.