Search

The main objective of this article is to investigate the stability and synthesis of controllers for discrete linear repetitive processes that exhibit polyhedral uncertainty. A condition is proposed, which utilizes a parameter-dependent Lyapunov function to alleviate conservatism resulting from uncertainty. The solution involves the use of a Lyapunov function that depends on uncertain parameters represented by a polyhedron, obtained through Linear Matrix Inequality (LMI) conditions. In practical scenarios, it is often the case that the frequency range of reference signals, noise, and interference is predefined, rendering full frequency range controller synthesis inadequate. To tackle this issue, we introduce a finite frequency controller using the generalized Kalman-Yakubovich-Popov (KYP) lemma. This approach enables designers to specify the desired control performance within a designated frequency range, which can be determined by analyzing the available signal spectrum.

In order to analyze and predict the performance of mechanical structure and system, the tangential contact of solid-liquid interface under normal and tangential loads is studied, and the tangential stiffness model of rough surfaces is established by analyzing various contact states of the asperity under normal elastic-plastic deformation and tangential stick-slip. Based on the contact model of closed oil pits based on the influence of rough surfaces, on the basis of fully considering the influence of gradual application of load on the number of closed oil pit, and then changing the load distribution of solid-liquid interface, a liquid tangential stiffness model is established. The tangential stiffness of solid-liquid interface is obtained by two parts in parallel. Through simulation and experimental analysis, the effects of normal load, deformation and lubricant viscosity on the tangential stiffness of solid-liquid interface are revealed, and the tangential stiffness with or without lubricating medium is compared and analyzed. The results show that the tangential stiffness of solid-liquid interface and solid-solid interface increases nonlinearly with the increase of normal load and deformation. Due to the contribution of liquid stiffness, the tangential stiffness of solid-liquid interface is always slightly higher than that of solid-solid interface. The greater the viscosity of the lubricating medium, the greater the tangential stiffness of the solid-liquid interface. At low load, liquid contact is dominant, and the tangential stiffness of solid-liquid interface is low. Under moderate and heavy loads, the solid contact is dominant, and the tangential stiffness of the solid-liquid interface increases rapidly. Therefore, the tangential contact stiffness of the solid-liquid interface in mixed lubrication can be effectively improved by increasing the normal load and increasing the viscosity of the lubricating medium.

High-speed directly modulated distributed feedback (DFB) laser is crucial for high-speed optical communication systems. The modulation bandwidth and frequency chirp of such devices are primarily affected by the differential gain and linewidth enhancement factor (LEF) of the quantum wells (QWs) in the active region. In this work, the optical gain properties of long wavelength AlGaInAs-InP strained quantum wells (QWs) with different crystal orientations are numerically investigated using the multi-bands k.p theory, which considers both valence-band anisotropic and nonparabolicity. Compared to the QW laser grown on a conventional (001) substrate, a higher differential gain, and smaller LEF are observed for the laser grown on the (110) substrate with the ${\rm{(\bar{1}10)}}$ plane as the mirror facet. This is primarily due to the reduced effective mass of the valence band. The dynamic characteristics of (001)-and (110)-oriented DFB laser is theoretically studied using the one-dimensional traveling wave model (1D TWM) at ${\rm{25}}\;^\circ \text{C}$ and ${\rm{95}}\;^\circ \text{C}$. The simulation results show that the lower chirp and wider modulation bandwidth can be achieved for the QW laser grown on (110) substrate.

The macroscopic relative motion solid-liquid interface widely exists in the contact motion pairs of machine tools and other mechanical equipment. In order to accurately obtain the tangential contact stiffness and damping parameters, the Savkoor asperity adhesion-sliding friction contact model is used to analyze the contact area and the corresponding tangential force changes of a single pair of solid contact asperity in the four typical phases of contact growth, contact stagnation, crack adhesion and crack propagation. According to the hypothesis of Gaussian distribution of asperity on rough surface, the contact model of single-pair asperities is extended to the whole joint. The tangential contact stiffness and damping models of solid-solid interface in macroscopic relative motion are obtained. For the fluid contact part, the oil film pressure distribution and film thickness are obtained by solving part of the film Reynold’s equation, and then the fluid tangential stiffness and damping model is established. The tangential contact stiffness and damping of the whole solid-liquid interface are obtained by analyzing the stiffness and damping of solid and fluid parts, and the effects of normal load and moving velocity on tangential contact stiffness and damping are obtained by simulation. The results show that the tangential contact stiffness and damping of solid-liquid interface increase with the increase of normal contact load and decrease with the increase of moving velocity.

We propose a compact and high isolation dual-polarized antenna array based on plasmonic meta-structure operating at 2.58 GHz. The compact principle of the dual-polarized antenna is mainly based on spoof surface plasmon polariton (SSPP) radiation patch to support SSPP modes, which is fed by spatially coupled excitation. Due to the slow-wave feature of SSPPs, the size of dual-polarized radiation patch is reduced to $0.2\lambda _{0}\times 0.2\lambda _{0}$ , and the total antenna size is reduced to $0.39\lambda _{0}\times 0.39\lambda _{0}$ , which is much smaller than that of the traditional microstrip antenna with $0.5\lambda _{0}\times 0.5\lambda _{0}$ . The cross-polarization coupling between two polarization ports can be improved to below −32 dB, showing high isolation characteristics. Besides, a $3\times 3$ compact antenna array is also designed, fabricated, and measured to verify the feasibility of its extension to larger-scale antenna systems. Both simulated and measured results of the proposed compact antenna and the antenna array illustrate compact profiles, good ±45° radiation performance, and high isolation levels, which have potential application values for the massive multiple-input and multiple-output (MIMO) base stations.

An accurate and comprehensive fault knowledge representation is indispensable for an automated and intelligent processing of power grid failures. Current knowledge graphs are incapable of capturing the complex relations among power grid failures. This paper extends the current knowledge graph representation mechanism through temporal, spatial, and causal representations to facilitate the knowledge representation of power grid faults, allowing for ontology modeling of power event elements and event relationships. During the modeling process, this paper proposes an extraction method which includes Bi-directional Encoder Representation from Transformers (BERT), Bi-directional Long Short-Term Memory (BiLSTM), and Conditional Random Field (CRF) for entities and relationships in power grid faults. The innovation of the method lies in the clever combination of the three, BERT learning semantic representation, BiLSTM further learning semantic features, and CRF joint modeling of labels to improve accuracy, and the results verify the effectiveness and practicality of the method presented in this paper.

Infantile spasms (IS) is a typical childhood epileptic disorder with generalized seizures. The sudden, frequent and complex characteristics of infantile spasms are the main causes of sudden death, severe comorbidities and other adverse consequences. Effective prediction is highly critical to infantile spasms subjects, but few related studies have been done in the past. To address this, this study proposes a seizure prediction framework for infantile spasms by combining the statistical analysis and deep learning model. The analysis is conducted on dividing the continuous scalp electroencephalograms (sEEG) into 5 phases: Interictal, Preictal, Seizure Prediction Horizon (SPH), Seizure, and Postictal. The brain network of Phase-Locking Value (PLV) of 5 typical brain rhythms is constructed, and the mechanism of epileptic changes is analyzed by statistical methods. It is found that 1) the connections between the prefrontal, occipital, and central regions show a large variability at each stage of seizure transition, and 2) 4 sub-bands of brain rhythms ( $\theta $ , $\alpha $ , $\beta $ , $\gamma $ ) are predominant. Group and individual variabilities are validated by using the Resnet18 deep model on data from 25 patients with infantile spasms, where the consistent results to statistical analyses can be observed. The optimized model achieves an average of $79.78~\%$ , $94.46\%$ , $75.46\%$ accuracy, specificity, and recall rate, respectively. The method accomplishes the analysis of the synergy between infantile spasms mechanism, model, data and algorithm, providing a guideline to build an intelligent and systematic model for comprehensive IS seizure prediction.

Transfer learning (TL) has been applied in seizure detection to deal with differences between different subjects or tasks. In this paper, we consider cross-subject seizure detection that does not rely on patient history records, that is, acquiring knowledge from other subjects through TL to improve seizure detection performance. We propose a novel domain adaptation method, named the Cluster Embedding Joint-Probability-Discrepancy Transfer (CEJT), for data distribution structure learning. Specifically, 1) The joint probability distribution discrepancy is minimized to reduce the distribution shift in the source and target domains, and strengthen the discriminative knowledge of classes. 2) A clustering is performed on the target domain, and the class centroids of sources is used as the clustering prototype of the target domain to enhance data structure. It is worth noting that the manifold regularization is used to improve the quality of clustering prototypes. In addition, a correlation-alignment-based source selection metric (SSC) is designed for most favorable subject selection, reducing the computational cost as well as avoiding some negative transfer. Experiments on 15 patients with focal epilepsy from the Children’s Hospital, Zhejiang University School of Medicine (CHZU) database shown that CEJT outperforms several state-of-the-art approaches, and can promote the application of seizure detection.

The extraction of relations between drug-related entities from drug instructions is essential for clinical diagnostic decision-making and drug use regulations, which is a critical task. However, due to the complexity of the textual descriptions in drug instructions, it is challenging to extract fine-grained relations, even with a considerable amount of training data. Moreover, since manually-labeled, high-quality datasets in the pharmaceutical domain are typically expensive, obtaining an extensive and accurate training dataset could be challenging. To overcome the above challenges, this paper proposes a drug relation extraction framework that combines entity information enhancement and contrastive feature learning, which can better extract fine-grained relations with limited data. More specifically, a sample generator creates a group of different samples with role semantic information from the training set, an entity encoder embeds the entity role information and context information to enhance the semantic representation, and a contrastive learning module employs a hybrid loss function to learn inter-sample and intra-sample differences. Empirical study indicates that the contrastive entity enhancement approach can achieve higher extraction accuracy and has better generalization capability. More specifically, the experimental results show that the F1 value of the model can reach 0.8892, which provides a 7.13% improvement compared to the baseline pre-training method.

With an increase in the electrification of end-use sectors, various resources on the demand side provide great flexibility potential for system operation, which also leads to problems such as the strong randomness of power consumption behavior, the low utilization rate of flexible resources, and difficulties in cost recovery. With the core idea of “access over ownership”, the concept of the sharing economy has gained substantial popularity in the local energy market in recent years. Thus, we provide an overview of the potential market design for the sharing economy in local energy markets (LEMs) and conduct a detailed review of research related to local energy sharing, enabling technologies, and potential practices. This paper can provide a useful reference and insights for the activation of demand-side flexibility potential. Hopefully, this paper can also provide novel insights into the development and further integration of the sharing economy in LEMs.

Removing the noise and improving the visual quality of hyperspectral images (HSIs) is challenging in academia and industry. Great efforts have been made to leverage local, global, or spectral context information for HSI denoising. However, existing methods still have limitations in feature interaction exploitation among multiple scales and rich spectral structure preservation. In view of this, we propose a novel solution to investigate the HSI denoising using a multiscale adaptive fusion network (MAFNet), which can learn the complex nonlinear mapping between clean and noisy HSI. Two key components contribute to improving the HSI denoising: A progressively multiscale information aggregation network and a coattention fusion module. Specifically, we first generate a set of multiscale images and feed them into a coarse-fusion network to exploit the contextual texture correlation. Thereafter, a fine fusion network is followed to exchange the information across the parallel multiscale subnetworks. Furthermore, we design a coattention fusion module to adaptively emphasize informative features from different scales, and thereby enhance the discriminative learning capability for denoising. Extensive experiments on synthetic and real HSI datasets demonstrate that the proposed MAFNet has achieved a better denoising performance than other state-of-the-art techniques.

An artificial intelligent neural network improved particle swarm optimization algorithm is proposed for the inverse design of semiconductor optical amplifier. Seven input parameters, current-gain curve and saturation output power curve are selected to form the data set based on the physical model of semiconductor optical amplifier. The effectiveness of forecasting performance is improved by contrasting two back propagation neural network techniques (Scaled Conjugate Gradient and Levenberg-Marquardt) and operational settings (Central Processing Unit and Graphics Processing Unit). Higher accuracy is achieved through feedback analysis of neuron number optimization and test error. The addition of a unique backpropagation neural network can make the fitness of particle swarm algorithm mostly converge below $2\times 10^{-4}$. The relative difference between original performances and inverse predictions is close to 0%, which proves the effectiveness of parameter extraction. This method can take advantage of neural networks to improve accuracy and speed of particle swarm optimization algorithms for efficient semiconductor optical amplifier inverse design and multi-solution analysis.

Crop-type mapping using time-series remote sensing data is crucial for a wide range of agricultural applications. Crop mapping during the growing season is particularly critical in timely monitoring of the agricultural system. Most existing studies focusing on within-season crop mapping leverage historical remote sensing and crop type reference data for model building, due to the difficulty in obtaining timely crop type samples for the current growing season. Yet the crop type samples from previous years may not be used directly considering the diverse patterns of crop phenology across years and locations, which hampers the scalability and transferability of the model to the current season for timely crop mapping. This article proposes an innovative within-season emergence (WISE) phenology normalized deep learning model towards scalable within-season crop mapping. The crop time-series remote sensing data are first normalized by the WISE crop emergence dates before being fed into an attention-based one-dimensional convolutional neural network classifier. Compared to conventional calendar-based approaches, the WISE-phenology normalization approach substantially helps the deep learning crop mapping model accommodate the spatiotemporal variations in crop phenological dynamics. Results in Illinois from 2017 to 2020 indicate that the proposed model outperforms calendar-based approaches and yields over 90% overall accuracy for classifying corn and soybeans at the end of season. During the growing season, the proposed model can give satisfactory performance (85% overall accuracy) one to four weeks earlier than calendar-based approaches. With WISE-phenology normalization, the proposed model exhibits more stable performance across Illinois and can be transferred to different years with enhanced scalability and robustness.

Objectives: Eye blink artifact detection in scalp electroencephalogram (EEG) of epilepsy patients is challenging due to its similar waveforms to epileptiform discharges. Developing an accurate detection method is urgent and critical. Methods: In this paper, we proposed a novel multi-dimensional feature optimization based eye blink artifact detection algorithm for EEGs containing rich epileptiform discharges. An unsupervised clustering algorithm based on smoothed nonlinear energy operator (SNEO) and variational mode extraction (VME) is proposed to detect epileptiform discharges in the frontal leads. Then, multi-dimensional time/frequency EEG features extracted from forehead electrodes (FP1 and FP2 channels) combining with the improved VME (IVME) threshold are derived for EEG representation. A variance filtering method is further applied for discriminative feature selection and a machine learning model is finally learned to perform detection. Results: Experiments on EEGs of 16 subjects from the Children’s Hospital of Zhejiang University School of Medicine (CHZU) show that our method achieves the highest average sensitivity, specificity and accuracy of 95.04, 89.52, and 93.01, respectively. That outperforms 5 recent and state-of-the-art (SOTA) eye blink detection algorithms. Significance: The proposed method is robust in eye blink artifact detection for EEGs containing high-frequency epileptiform discharges. It is also effective in dealing with individual differences in EEGs, which is usually ignored in conventional methods.

Due to the different imaging mechanisms between optical and polarimetric synthetic aperture radar (PolSAR) images, determining how to effectively use such complementary information has become an interesting and challenging problem. Convolutional neural networks (CNNs) and other deep neural networks have achieved good experimental results in remote sensing land-cover semantic segmentation. However, the CNN convolution structure can extract only the features within the receptive field in the spatial dimension without focusing on the relationship between multiple channels; therefore, it is impossible to realize fusion and complementarity between multiple channels. In this article, we propose a novel spatial dense channel attention fusion network (SDCAFNet), which takes optical and PolSAR images as different inputs and completes feature fusion and semantic segmentation within a neural network. First, SDCAFNet uses a two-stream siamese CNN network to realize the preliminary feature coding of optical and PolSAR images. Then, a spatial dense channel attention module (SDCAM) is proposed. The channel activation values obtained at different positions are combined in the spatial dense matrix, which can describe the attention in the feature fusion process. Finally, we introduce the fused features into the symmetric skip-connection decoder composed of multiple symmetric decoder blocks to realize end-to-end land-cover semantic segmentation. Experimental results show that SDCAFNet can effectively learn the correlation between optical and PolSAR channels and has a better segmentation accuracy than other methods.

The autonomous vehicles (AVs) need to share the driving environment with the human driving vehicles (HDVs) on expressway in the future. The non-humanlike lane changing (LC) behavior of AVs can mislead human drivers, which brings potential risks. Stronger humanlike ability requires a more complex algorithm. However, the requirement of the on-board vehicle computation resources limits the humanlike ability of LC algorithms. In this context, considering environmental risks, driver speed requirements and driver focus shifting process, this paper proposes a new type of LC algorithm based on artificial potential field (APF) aiming to improve the humanlike ability. The coupling relationship between the longitudinal and the lateral potential field forces is analyzed theoretically based on environmental risk APF. Based on the relationship, we proposed a conversion mechanism of the driver between the lateral and longitudinal potential field forces to mimic the driver’s speed requirement. Then a target lane selection strategy is designed to trade-off the driver’s speed requirement and safety between multiple lanes. Finally, to mimic the driver’s focus transfer process, the lateral forces of the ego vehicle are analyzed theoretically to further propose a moving virtual lane lines algorithm to build the lateral space varying potential field. The algorithm proposed in this paper is validated by comparing with other LC algorithms in the real traffic scenarios. The results indicate that the proposed algorithm has a better ability to mimic human driver’s LC decision-making and provides a smoother and safer humanlike trajectory.

Synthetic aperture radar (SAR) image change detection is a vital yet challenging task in the field of remote sensing image analysis. Most previous works adopt a self-supervised method which uses pseudolabeled samples to guide subsequent training and testing. However, deep networks commonly require many high-quality samples for parameter optimization. The noise in pseudolabels inevitably affects the final change detection performance. To solve the problem, we propose a graph-based knowledge supplement network (GKSNet). To be more specific, we extract discriminative information from the existing labeled dataset as additional knowledge, to suppress the adverse effects of noisy samples to some extent. Afterward, we design a graph transfer module to distill contextual information attentively from the labeled dataset to the target dataset, which bridges feature correlation between datasets. To validate the proposed method, we conducted extensive experiments on four SAR datasets, which demonstrated the superiority of the proposed GKSNet as compared to several state-of-the-art baselines. Our codes are available: at https://github.com/summitgao/SAR_CD_GKSNet.

Benefited from the rapid and sustainable development of synthetic aperture radar (SAR) sensors, change detection from SAR images has received increasing attentions over the past few years. Existing unsupervised deep learning-based methods have made great efforts to exploit robust feature representations, but they consume much time to optimize parameters. Besides, these methods use clustering to obtain pseudolabels for training, and the pseudolabeled samples often involve errors, which can be considered as “label noise.” To address these issues, we propose a dual path denoising network (DPDNet) for SAR image change detection. In particular, we introduce the random label propagation to clean the label noise involved in preclassification. We also propose the distinctive patch convolution for feature representation learning to reduce the time consumption. Specifically, the attention mechanism is used to select distinctive pixels in the feature maps, and patches around these pixels are selected as convolution kernels. Consequently, the DPDNet does not require a great number of training samples for parameter optimization, and its computational efficiency is greatly enhanced. Extensive experiments have been conducted on five SAR datasets to verify the proposed DPDNet. The experimental results demonstrate that our method outperforms several state-of-the-art methods in change detection results.

The response of the global ionosphere to solar flares is an important topic in the field of space weather. The global ionospheric response to solar flares is comprehensively analyzed from the perspectives of total electron content (TEC) and very low frequency (VLF) signals by using solar flare data on the eruption days of X-class flares from 2006 to 2019, including flare level, flare duration, geographical location, and local time. In addition, the relationship between X-ray flux and VLF phase variation is studied through correlation analysis. The concepts of disturbance intensity (DI) and disturbance angle are defined, and a DI evaluation model is established to help detect the difference in sensitivity to solar flares between TEC and VLF signals. Results show the following. (1) The higher the flare level and the longer the duration, the greater the disturbance to the ionosphere. Simultaneously, a phenomenon exists in which the disturbance caused by low-level and long-lasting flares is greater than that caused by high-level flares. (2) VLF phase variation and flare level exhibit a good correlation, and they are also closely related to geographical location and local time. The disturbance degree of a station facing the sun is more evident than that of a station facing away from the sun. The TEC disturbance of stations in the morning (local time) is more obvious than that in the afternoon, and disturbance increases along the direction of Earth’s rotation. (3) When DI is at the same level, the lowest flare level that causes TEC response is higher than the lowest flare level that causes VLF signal response. The disturbance angle of TEC is unevenly distributed within the interval [0°, 90°], and that of VLF signals is more than 85°. The sensitivity of VLF signals to flare response is considerably higher than that of TEC, and the difference between the two even stride across DI level.

Existing semi-supervised anomaly detection methods usually use a large amount of labeled normal data for training, which have the problem of high labeling costs. Only a few semi-supervised methods utilize unlabeled data and a few labeled anomalies to train models. However, these kinds of methods usually encounter two problems: (i) since anomalies usually have different behavior patterns or the internal mechanisms that produce anomalies are complex and diverse, a few labeled anomalies cannot cover all anomaly types; and (ii) the amount of unlabeled data in the training set is substantially greater than the amount of labeled data, which leads to that unlabeled data with contamination often dominates the training process. To solve these two problems, we propose the semi-supervised anomaly detection method named ConNet and a new loss function named concentration loss. Specifically, ConNet consists of two stages. Firstly, we obtain the prior anomaly score of unlabeled data via prior estimation module and attach the prior anomaly score to unlabeled data as the training weight. Then, an anomaly scoring network is training to assign anomaly scores to data instances, which can ensure that the anomaly scores of anomalies significantly deviate from those of normal instances. We have conducted experiments on thirteen real-world data sets and tested the performance of our method in terms of detection accuracy, utilization efficiency of labeled data, and robustness to different contamination rates. The experimental results show that the performance of our method is significantly better than those of the state-of-the-art anomaly detection methods.

Synthetic aperture radar (SAR) image change detection is a critical yet challenging task in the field of remote sensing image analysis. The task is nontrivial due to the following challenges: First, intrinsic speckle noise of SAR images inevitably degrades the neural network because of error gradient accumulation. Furthermore, the correlation among various levels or scales of feature maps is difficult to be achieved through summation or concatenation. Toward this end, we proposed a siamese adaptive fusion (AF) network for SAR image change detection. To be more specific, two-branch CNN is utilized to extract high-level semantic features of multitemporal SAR images. Besides, an AF module is designed to adaptively combine multiscale responses in convolutional layers. Therefore, the complementary information is exploited, and feature learning in change detection is further improved. Moreover, a correlation layer is designed to further explore the correlation between multitemporal images. Thereafter, robust feature representation is utilized for classification through a fully connected layer with softmax. Experimental results on four real SAR datasets demonstrate that the proposed method exhibits superior performance against several state-of-the-art methods. Our codes are available at https://github.com/summitgao/SAR_CD_SAFNet.

This study presents an automated registration method based on optical flow for an adaptive optics scanning laser ophthalmoscope. The method was designed to align and average images to obtain a higher signal-to-noise ratio image. A correlation based optical flow image registration method, which has large registration degrees of freedom, is adopted as a local registration method. By comparing the images before and after image registration, we show the effectiveness of our method. Furthermore, the advantage of our method, which is the containment of large registration degrees of freedom, is confirmed.

Stability and reliability are the most important indicators of drive systems. In this article, a robust nonlinear control method is advanced for dual three-phase permanent-magnet synchronous machine (dual-PMSM) drives, based on the synergetic control theory. This control is implemented by a new synergetic controller (SGC) integrating with an extended sliding mode observer (ESMO), and an optimization method is applied to obtain the optimal parameters of the SGC. This control scheme not only improves the dynamic performance even with operating parameters change, but also can diagnose the speed sensor fault and realize the fault tolerant control (FTC). The proposed control method is compared with the PI control, the sliding mode control (SMC) and the predictive control, as well as the simulation and experimental results confirm the superior dynamic performance and the effectiveness on the speed sensor fault diagnosis (FD) and FTC.

Geometrical analysis of cone photoreceptor cells is important not only for ophthalmic diagnosis, but also for research on eye diseases. In this study, an automated cone photoreceptor cell segmentation and identification method based on morphological processing and watershed algorithm is presented for adaptive optics scanning laser ophthalmoscope images. Our method includes steps for image denoising, rough segmentation, fine segmentation, small region removal, and identification. The effectiveness of the proposed method was confirmed by comparing its results with those obtained manually yielding precision, recall, and F1-score values of 93.6%, 98.0% and 95.8%, respectively. The performance of our method is further verified by processing images with different cone photoreceptor cell densities from healthy retina and an image from an eye with diabetic retinopathy. The experimental results show that our algorithm achieved high accuracy in cone photoreceptor cell segmentation and identification in healthy retinas as well as in retina with diabetic retinopathy.

Object detection plays an important role in automatic driving systems. Considering the characteristics of classical and deep learning algorithms, a fusion logic is proposed to combine the advantages of these two kinds of object detectors. The relationship of detection performance among different detectors is established theoretically. According to the established theoretical relationship, the improvement of detection performance by fusion is further studied numerically. Furthermore, an optimization method is proposed to guide the design of the sub-detectors to achieve a better comprehensive performance. The effectiveness of this combined approach is validated by application to the detection of pedestrian, in which a support vector machine trained by the HOG feature of pedestrian is adopted as the classical detector and a comparatively simple transfer convolutional neural network (CNN) based on AlexNet structure acts as the deep learning detector. Several comparative tests with the classical and CNN detectors on the training dataset and other totally different dataset have been conducted to show the advantage of the combined one in ensuring detection performance with simpler network and adaptability to new application conditions.

In open traffic environments, humans still have to remain in the control loop of vehicle due to the insufficient of the existing technologies and their high costs. For the realization of cooperation between the human and the automatic driving system, the determination of the time when automatic driving is necessary is very important. To avoid unnecessary intervention when the driver has the control authority of vehicle, a new driving capability-based transition strategy was proposed, which comprehensively considers the driver's correction ability and the driving risk. The transition time from the human driver to the automatic driving system is determined by an unreliable domain (UD), whose boundary is modeled according to the driving data recorded by a driving simulator and statistically described by a log-normal distribution. Furthermore, an adaptive algorithm is designed to update the parameters of UD boundary online to make this strategy suitable for different drivers. This UD-based transition strategy is validated by several tests on the driving simulator. The bench test results show that the individual driving characteristic can be identified by the adaptive algorithm in time, the transition time determined by UD is more accurate, and sufficient time is reserved for the correction carried out by the automatic driving system.

High-resolution remote sensing image object detection plays an increasingly important role in image processing and interpretation. The application of region-based convolutional neural network (R-CNN) greatly enhances the performance of object detection. However, the attributes of remote sensing images such as overlarge image size, similar background, disequilibrium distribution of categories make this task more challenging. The previous works have focused on extracting multi-scale features of region proposals, often ignoring the quality of region of interest (ROI). In this work, we proposed a patch-based three-stage aggregation network (PTAN) for object detection in high-resolution remote sensing images. It consists of a three-stage cascade structure that sequentially improves the quality of candidate regions by increasing the IoU threshold stage by stage, and adopts a resampling strategy to obtain sufficient region proposals. At the same time, we also proposed patch-based strategy and applied it to the framework during training and inference. Ablation experiments and comprehensive evaluations on a communal remote sensing image object detection dataset DOTA demonstrate the effectiveness and robustness of the proposed framework, which obtained a mean average precision (mAP) value of 0.7958 on validation dataset and a front-rank mAP of 0.7858 on testing dataset. On another remote sensing image object detection dataset NWPU VHR-10, the proposed PTAN obtained a mAP value of 0.9187, outperforming other five object detectors.

A narrow-strip single-longitudinal-mode gain-coupled distributed feedback semiconductor laser with a central wavelength of approximately 670 nm is demonstrated. We fabricate a novel micro-scale regrowth-free gain-coupled grating with an island-like shape without nanoscale gratings by i-line lithography only. The maximum output power from the fiber of the device with butterfly package is 48.1 mW at 100 mA and its slope efficiency is 0.79 W/A. The laser can be continuously tuned from 670.750 to 670.784 nm with a side-mode suppression ratio of over 40 dB, covering the absorptive peak of lithium atoms. Our easily fabricated devices have great potential as light sources in many applications such as lithium atom detectors.

It is well acknowledged that the functional near-infrared spectroscopic (fNIRS) imaging of brain functions can be quantitatively enhanced by diffuse optical tomography (DOT). Unlike the widely-used conventional optical topography (optical mapping), an fNIRS-DOT instrument needs to possess both high sensitivity and large dynamic-range so that measurements at multiple source-detector separations are reliably and synchronously collected for improved depth-resolution, quantitation, and superficial signal reduction, the performances particularly crucial to the occipital cortex probing. For the goal, we implemented a novel 3-wavelength, 240-channel (20 sources and 12 detectors) continuous-wave fNIRS-DOT instrument. The system features the square-wave modulation lock-in photon-counting scheme that combines the multi-channel parallelization of the lock-in detection, the ultra-high sensitivity, and unprecedented dynamic-range of the photon-counting technique, as well as the reduced complexity of the hardware architecture. We have systematically investigated the key specifications of the instrument such as the detection linearity, stability, channel crosstalks and so on, under the practically-viable conditions of the incident power and signal-to-noise ratio, and accordingly proposed wavelength-switched multiple-field illumination paradigms for both the non-overlapping and overlapping DOT measurements, respectively. Finally, DOT experiments using the semi-three-dimensional image reconstruction has been conducted on a phantom with a practically low-contrasted and deep-sit absorption target, to evaluate the overall imaging performance of the system and to demonstrate the superiority of the model-based image reconstruction over the topographical mapping.

A line-scanning ophthalmoscope (LSO) is a retinal imaging technique that has the characteristics of high imaging resolution, wide field of view, and high imaging speed. However, the high-speed imaging with rather short exposure time inevitably reduces the signal intensity, and many factors, such as speckle noise and intraocular scatter, further degrade the signal-to-noise ratio (SNR) of retinal images. To effectively improve the image quality without increasing the LSO system's complexity, the post-processing method of image super-resolution (SR) is adopted. In this paper, we propose a learning-based multi-frame retinal image SR method that directly learns an end-to-end mapping from low-resolution (LR) image sequences to high-resolution (HR) images. This network was validated on down-sampled and real LSO image sequences. We evaluated the method on a down-sampled dataset with the metrics of peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and perceptual distance. Moreover, the power spectra and full width at half maximum (FWHM) were used as the no-reference image quality assessment (NR-IQA) algorithms to evaluate the reconstruction results of the real LSO image sequences. The experimental results indicate that the proposed method can significantly enhance the SNR of LSO images and efficiently improve the resolution of LSO retinal images, which has great practical significance for clinical diagnosis and analysis.

High-power single-longitudinal-mode regrowth-free gain-coupled distributed feedback laser diode based on ridge waveguide with periodic current injection is achieved at 990 nm. Our device is fabricated only by standard i-line lithography with micron-scale precision, obtains an excellent performance at high injection current. A continuous-wave power of over 0.681 W is achieved at 3 A. The maximum continuous-wave power at single-longitudinal-mode operation is up to 0.303 W at 1.4 A. Narrow linewidth emission has been reached with a 3 dB spectrum width less than 1.41 pm. The high side mode suppression ratio is over 35 dB. The lateral far field divergence angle is only 15.05°, the beam quality factor M2 is 1.245, achieving a laterally near-diffraction-limit emission. It is more beneficial for single-mode fiber coupling as pumping sources and other applications which require high beam quality at high power with easy fabrication technique.

Rock materials used as building stone and construction materials are usually subjected to the corrosion of hydro-chemical solutions. To investigate the effect of chemical corrosion on the change of pore structure and mechanical properties of sandstones, the nuclear magnetic resonance (NMR) technique was used for the measurement of porosity and T2 spectrum distribution. Series of triaxial compression tests with the confining pressure of 0, 5, 10, and 20 MPa were then conducted for rock specimens treated with chemical corrosions. The test results revealed that, compared with the rock specimens in their natural state, after chemical corrosion the porosity increased, the T2 spectrum distribution was broadened successively, and the pore structure became more complex with the fractal dimension increasing. The proportions of pores with different sizes were various with the change of chemical solutions and the immersion time. Moreover, under a certain confining pressure, the peak strength, elastic modulus, internal friction angle, and cohesion of sandstone after chemical corrosions decreased, but the corresponding peak strain increased. A chemical damage variable was proposed based on the porosity of sandstone and the order of effect of chemical solutions on pore structure and mechanical properties of sandstone was obtained as H2SO4 solution > NaOH solution > distilled water. Finally, a decay model considering both of chemical damage and confining pressure was established to predict the triaxial compressive strength of sandstone treated with chemical corrosion.

A two-segment regrowth-free gain-coupled distributed feedback (DFB) laser monolithically integrated different gain contrast segments was proposed, realizing both wavelength-tunable (5 nm) and wavelength stabilization characteristics, controlled by a single electrode only. At current larger than 320 mA, side mode suppression ratio reached >30 dB with 3 dB linewidth

A low-cost regrowth-free narrow-strip single-longitudinal-mode diode laser, based on periodic anodes defined only by i-line lithography is proposed. With the periodic anodes separated by surface insulated grooves acting as a 44th-order gain grating, a high gain contrast is achieved without introducing effective index coupling. The CW output power of the device without facet coatings is up to 100.9 mW/facet at 980.63 nm in single-longitudinal-mode operation with a single facet having a power conversion efficiency of up to 18.6% and a slope efficiency of 0.53 W/A. The highest side-mode suppression ratio is 45 dB with a linewidth below 889 MHz, at 350 mA. The beam parameter product, M2, in the slow axis is 1.85, achieving a near diffraction-limited beam quality. A novel method for easy surface gratings fabrication is first presented, wherein, 4-μm-width periodic surface grooves, acting as waveguide ridges, are defined once by i-line lithography only without regrowth, nanoscale gratings fabrication, or facet coating. Owing to the low-cost fabrication methods and good performances, our single-longitudinal-mode devices can be used for widespread practical applications.

Soil drainage is a widely used agricultural practice in the midwest USA to remove excess soil water to potentially improve the crop yield. Research shows an increasing trend in baseflow and streamflow in the midwest over the last 60 years, which may be related to artificial drainage. Subsurface drainage (i.e., tile) in particular may have strongly contributed to the increase in these flows, because of its extensive use and recent gain in the popularity as a yield-enhancement practice. However, how evapotranspiration (ET) is impacted by tile drainage on a regional level is not well-documented. To explore spatial and temporal ET patterns and their relationship to tile drainage, we applied an energy balance-based multisensor data fusion method to estimate daily 30-m ET over an intensively tile-drained area in South Dakota, USA, from 2005 to 2013. Results suggest that tile drainage slightly decreases the annual cumulative ET, particularly during the early growing season. However, higher mid-season crop water use suppresses the extent of the decrease of the annual cumulative ET that might be anticipated from widespread drainage. The regional water balance analysis during the growing season demonstrates good closure, with the average residual from 2005 to 2012 as low as -3 mm. As an independent check of the simulated ET at the regional scale, the water balance analysis lends additional confidence to the study. The results of this study improve our understanding of the influence of agricultural drainage practices on regional ET, and can affect future decision making regarding tile drainage systems.

Our world and our lives are changing in many ways. Communication, networking, and computing technologies are among the most influential enablers that shape our lives today. Digital data and connected worlds of physical objects, people, and devices are rapidly changing the way we work, travel, socialize, and interact with our surroundings, and they have a profound impact on different domains, such as healthcare, environmental monitoring, urban systems, and control and management applications, among several other areas. Cities currently face an increasing demand for providing services that can have an impact on people's everyday lives. The CityPulse framework supports smart city service creation by means of a distributed system for semantic discovery, data analytics, and interpretation of large-scale (near-)real-time Internet of Things data and social media data streams. To goal is to break away from silo applications and enable cross-domain data integration. The CityPulse framework integrates multimodal, mixed quality, uncertain and incomplete data to create reliable, dependable information and continuously adapts data processing techniques to meet the quality of information requirements from end users. Different than existing solutions that mainly offer unified views of the data, the CityPulse framework is also equipped with powerful data analytics modules that perform intelligent data aggregation, event detection, quality assessment, contextual filtering, and decision support. This paper presents the framework, describes its components, and demonstrates how they interact to support easy development of custom-made applications for citizens. The benefits and the effectiveness of the framework are demonstrated in a use-case scenario implementation presented in this paper.

When multiple distributed converters are integrated, the high frequency harmonics will randomly accumulate at the point of common coupling (PCC). This paper proposes a new fast global synchronous discontinuous pulse width modulation (GSDPWM) method of three-phase inverters to effectively attenuate the high frequency current harmonics at PCC. Firstly, the basic principle and the realization method of GSDPWM for three-phase inverters are explained, which can be employed for different modulation types. Then a fast calculation method, which can equally derive the minimized total harmonic distortion (THD) of total current, is proposed to release the calculation burden. Finally, MATLAB simulations and experimental results are presented to verify the performance of GSDPWM.

In this paper, a high-temperature fiber sensor based on an optical fiber Fabry-Perot interferometer is fabricated by splicing a section of simplified hollow-core fiber between two single-mode fibers (SMFs) and cleaving one of the two SMFs to a certain length. With the superposition of three beams of light reflected from two splicing joints and end face of the cleaved SMF, the modified Vernier effect will be generated in the proposed structure and improve the sensitivity of temperature measurement. The envelope of spectrum reflected from the proposed sensor head is modulated by the ambient temperature of the sensor head. By monitoring and measuring the shift of spectrum envelope, the measurement of environment temperature is carried out experimentally, and high temperature sensitivity of 1.019 nm/°C for the envelope of the reflected spectrum was obtained. A temperature measurement as high as 1050 °C has been achieved with excellent repeatability.

This paper proposes a novel approach to compensate buses voltage and current harmonics through distributed generation (DG) interfacing converter in a multi-bus microgrid. The control approach of each individual DG unit was designed to use only feedback variables of the converter itself that can be measured locally. In the proposed approach, the adjacent bus voltage is indirectly derived from the measured DG converter output voltage, DG line current and line impedance. A voltage closed-loop controller and a current closed-loop controller are designed to achieve both functions of DG real power generation and harmonics compensation. Therefore, the traditional harmonic measurement devices installed at the bus as well as the long distance communication between the bus and the DG converter are not required. The proposed approach can compensate the current harmonics, mitigate the buses voltage distortion and enable the customer devices to be operated in normal conditions within the multi-bus microgrid, and meanwhile relieve the burden of power quality regulator installed at the point of common coupling. Matlab simulations and experimental results are presented to show the operational effectiveness of the proposed approach.

This paper presents control methods for hybrid AC/DC microgrid under islanding operation condition. The control schemes for AC sub-microgrid and DC sub-mi-crogrid are investigated according to the power sharing requirement and operational reliability. In addition, the key control schemes of interlinking converter with DC-link capacitor or energy storage, which will devote to the proper power sharing between AC and DC sub-microgrids to maintain AC and DC side voltage stable, is reviewed. Combining the specific control methods developed for AC and DC sub-microgrids with interlinking converter, the whole hybrid AC/DC microgrid can manage the power flow transferred between sub-microgrids for improving on the operational quality and efficiency.