Search

Transient impulsive signal is usually related with the bearing or gear local defect. It is very difficult to extract those multi-transient features due to the non-stationary of the corresponding vibration signals of rotating machinery. Time-frequency analysis is a suitable tool for analyzing non-stationary signals. A multiple transient extracting transform has been proposed in this work, which can not only effectively detect the multiple transient information in the signal, but also achieve a more concentrated time-frequency representation. The results of numerical simulation show the effectiveness of this proposed method. The proposed multi-transient extracting transform can better locate the transient features and has a lower time-consuming and better noise robustness, compared with the traditional time-frequency analysis methods. Finally, multi-transient extraction algorithm is utilized to analyze practical bearing vibration signals. It has been well demonstrated that the proposed method is more effective than other advanced time-frequency methods in the field of transient feature extraction.

Impedance measurement is a practical method to obtain the impedance characteristic of the power electronic equipment (PEE), which is important for the small-signal stability analysis. In the impedance measurement, the measurement results may be influenced by the external noise. Therefore, it is essential to study how to avoid the measurement errors caused by the noise. To address this issue, this paper analyzes the method of suppressing the noise interference by enhancing signal-to-noise ratio (SNR) and concludes that there is limitation on improving the measurement accuracy by merely enhancing SNR due to the practical limitation on the perturbation amplitudes. And then this paper proposes an improved impedance measurement method, by which the measurement errors caused by the noise can be decreased effectively. By investigating the influence mechanism of the noise on the impedance measurement results, the measurement error can be divided into average component and fluctuant component. The average component can be eliminated by averaging two groups of measurement results with the phase difference of $\pi $ rad while the influence of the fluctuant component can be minimized by employing complex-linear fitting algorithm. Finally, experiments are conducted to verify the effectiveness of proposed impedance measurement method.

High-performance InAs/GaSb type-II superlattice infrared detectors and focal plane arrays (FPAs) are normally grown by molecular beam epitaxy (MBE). In this work, we demonstrate the first long-wavelength infrared InAs/GaSb superlattice FPA grown by metalorganic chemical vapor deposition (MOCVD) with clear image. High-quality superlattice material was obtained evidenced by sharp X-ray diffraction peaks and atomic flat surface. Electrical and optical measurements performed on single element detectors showed a 50% cut-off wavelength of $\sim 10.1~\mu \text{m}$ , a dark current density of $2.5\times 10^{-5}$ A/cm2, a peak responsivity of 0.88 A/W and a peak detectivity of $1.7\times 10^{11}$ cm $\cdot $ Hz1/2/W at 80 K. A $320\times256$ FPA with $30~\mu \text{m}$ pixel pitch was then fabricated. With an integration time of 1.9 ms and an applied bias of -0.1 V, the FPA shows an average operability of 96.96%, a non-uniformity of 4.97%, a noise equivalent temperature difference of 51.1 mK and a peak detectivity of $2.3\times 10^{10}$ cm $\cdot $ Hz1/2/W at 80 K without thinning down the substrate.

Existing text sentiment analysis methods mostly rely on a large number of language knowledge and sentiment resources. This paper proposes the Multi-channel convolution and bidirectional GRU multi-head attention capsule (AT-MC-BiGRU-Capsule), which uses vector neurons to replace scalar neurons to model text emotions, and uses capsules to characterize text emotions. In addition, traditional methods cannot extract the multi-level features of text sequence well. Multi-head attention can encode the dependencies between words, capture sentiment words in text, and using Convolutional Neural Network (CNN) and Bidirectional gated recurrent unit network (Bi-GRU) to extract local features and global semantic features of text respectively, the global average pooling layer is introduced to obtain the multi-level feature representation of the text sequence more comprehensively. This paper selects three English datasets and one Chinese dataset in the general corpus of sentiment classification to conduct experiments, and achieves better results than other baseline models.

Accurate fault location is crucial for finding and clearing faults in distribution networks. It can help to reduce the loss of power failure and ensure safe and stable operation. Without efficient observation points and the lack of phase angle information, traditional fault location methods of distribution networks usually have large errors. As the application of $\mu $ PMUs in the distribution network becomes more and more common, voltage and current waveforms, as well as phase angle can be recorded and provide more information for accurate fault location. In this paper, an intelligent location method is proposed to pinpoint the fault location based on the information of $\mu $ PMUs which are properly allocated. The fault section is firstly determined by comparing the zero-sequence current waveforms on both sides of the fault section. Then, a Stack Auto-Encoder (SAE) is modeled to provide an end-to-end solution to pinpoint fault point with the voltage and current phasors. Finally, the performance of the proposed method is tested by a simulated distribution network on the platform of PSCAD. The results show that the proposed method is effective in locating faults and can withstand the effects of transition resistance, fault type, and noise. Compared to another popular method, the proposed method shows better location accuracy.

To avoid the irregularities during the level set evolution, a fractional distance regularized variational model is proposed for image segmentation. We first define a fractional distance regularization term which punishes the deviation of the level set function (LSF) and the signed distance function. Since the fractional derivative of the constant value function outside the starting point is nonzero, the fractional gradient modular of the LSF does not approach infinity where the integer order gradient modular is close to 0. This prevents the sharp reverse diffusion of LSF in flat areas and ensures the smooth evolution of LSF. Then, we use the Grünwald-Letnikov (G-L) fractional derivative to derive the discrete forms of the conjugate of fractional derivatives and fractional divergence. To facilitate the calculation of fractional derivatives and their conjugates, we designed their covering templates. Finally, a numerical solution to the minimization of the energy functional is obtained from these discrete forms and covering templates. Numerical experiments of medical images with different modalities show that the model in this paper can well segment weak images and intensity inhomogeneity images.

For overhead long-distance high voltage direct current (HVDC) transmission lines, transients are produced due to complicated field conditions and lightning activities. To ensure reliable operation of protection devices, accurate recognition of faults and disturbances is quite critical. The most popular recognition methods include threshold-based ones which require the proper setting of the threshold value, and classifier-based ones that need suitable feature extractions. These methods depend heavily on the experience of engineers or experts and are ineffective in dealing with the variation of system parameters. In this paper, a transient recognition method based on stack auto-encoder (SAE) is proposed to characterize different transients and to avoid human interferences. A symmetrical ±500kv HVDC system is modeled to illustrate the performance of the proposed method. The effect of some factors, such as noises and conductors, are discussed and compared. The simulation results demonstrate that the proposed SAE-based recognition has excellent potential in transient recognition of practical HVDC systems.

Radar-based sensors do not require optimal lighting and atmospheric conditions and nonocclusion, making them a promising solution for human behavior analysis in complex environments. Existing radar-based models generally retrieve features from either the time-velocity domain or the time-range domain. Such two-dimensional representations cannot fully depict dynamic human motion features. In this paper, we propose a temporal range-Doppler PointNet-based method to analyze human behavior. We transform human echoes to 3D point sets and then feed them into the hierarchical PointNet model for classification. The proposed point network can learn structural features from the micromotion trajectory more effectively than directly processing the raw point cloud. To further improve our model’s robustness in practical applications, we design an outlier detection module for detecting anomalies such as in multitarget scenarios. The results of experiments on motion capture databases and range-Doppler radar measurements demonstrate that our method realizes outstanding performance in terms of the classification accuracy, noise robustness, and anomaly detection accuracy.

This paper studies a multi-objective optimization (MOO) problem in full-duplex (FD) networks with simultaneous wireless information and power transfer (SWIPT). In the considered networks, an FD base station (BS) communicates with multiple half-duplex (HD) uplink (UL) and downlink (DL) users simultaneously. The energy receivers (ERs) harvest energy from the ambient radio frequency (RF) signals and may act as potential eavesdroppers. Specifically, there exists two conflicting yet important system design objectives, i.e., secrecy energy efficiency (SEE) maximization and energy harvesting efficiency (EHE) maximization. An MOO design based on the weighted Tchebycheff approach is proposed to investigate the tradeoff between these two design objectives. The proposed design takes into account the quality-of-service (QoS) guarantees of secrecy rate and energy harvesting (EH) under the imperfect channel state information (CSI) of ERs. The formulated MOO problem is solved with a two-layer optimization algorithm, where the modified Dinkelbach method is applied to deal with the generalized fractional programming (FP) in the outer layer problem and semidefinite relaxation (SDR), successive convex approximation (SCA) as well as S-procedure methods are applied to transfer the inner layer problem into a convex iterative program. Moreover, we propose a suboptimal solution to the formulated problem and analyze the computational complexities. Finally, numerical results are provided to demonstrate the effectiveness of the proposed algorithms and reveal a tradeoff region between SEE and EHE.

In order to adapt to the rapid development of cloud computing, big data, and other technologies, the combination of data center networks and SDN is proposed to make network management more convenient and flexible. With this advantage, routing strategies have been extensively studied by researchers. However, the strategies in the controller mainly rely on manual design, the optimal solutions are difficult to be obtained in the dynamic network environment. So the strategies based on artificial intelligence (AI) are being considered. This paper proposes a novel routing strategy based on deep Q-learning (DQL) to generate optimal routing paths autonomously for SDN-based data center networks. To satisfy the different demands of mice-flows and elephant-flows in data center networks, deep Q networks are trained for them respectively to achieve low latency and low packet loss rate for mice-flows as well as high throughput and low packet loss rate for elephant-flows. Furthermore, with the consideration of the distribution of traffic and the limited resources of data center networks and SDN, we choose port rate and flow table utilization to describe the network state. Simulation results show that compared with Equal-Cost Multipath (ECMP) routing and Selective Randomized Load Balancing (SRL)+FlowFit, the proposed routing scheme can reduce both the average delay of mice-flows and average packet loss rate, while increase the average throughput of elephant-flows.

The security of Public-Key Infrastructure (PKI) for Internet-based communications has lately attracted researchers' attention because of Certification Authorities (CAs) crashes and consequent attacks. Google Certificate Transparency and subsequent log-based PKI proposals (e.g., AKI and ARPKI) have succeeded in making certificate-management processes more transparent, accountable, and verifiable. However, those proposals failed to solve the root CA generous delegation of trust to intermediate CAs, non-conformant certificate-issuance by them, and lack of rigorous authentication of domain ownership during certificate-issuance problems. This study presents Attack-Resilient TLS Certificate Transparency (ARCT) based on log servers to address these problems. ARCT enables root CA to enforce intermediate CAs to follow community standards through leveraging a log server at each root level. It also introduces a collaborative domain ownership verification method that deters false certificate-issuance and ensures that a set of CAs validates every certificate before any client will accept it. A certificate collectively approved by a set of CAs assures users that the certificate has been seen, and not instantly detected malicious, by a group of CAs. Finally, formal security and performance evaluations prove the reliability and effectiveness of ARCT.

Unidirectional transmission mechanism (UTM) is widely used in mechanical field. The UTM studied in this paper is the core component of the aircraft hatch system, which is used to transfer the power of the motor to drive the hatch to be retracted and locked automatically after the motor closes. Aiming at the failure mode exposed in the experiment, the matching relations of both geometric parameters and mechanical parameters are studied, and the force analysis model of the UTM is established. The results show that the load of the UTM reaches its maximum at 65°, and if the friction coefficient between the friction plates is less than its critical value, braking failure will occur, which is in agreement with the test phenomena. In order to further verify the failure mechanism and obtain the effects of the gradual damage on the performance, a parameterized simulation model is established, through which failure phenomenon in the test is recovered, and the correctness of the failure mechanism analysis is verified. Finally, combined with its failure mechanism, the maintenance advice is proposed, and the validity is verified through the simulation model.

In recent years, cycling has become increasingly popular globally, which takes up little space and leads to nearly no environmental damage. Bicycles permit daily commuters to travel in an efficient manner through frequent traffic congestion. Mixed traffic conditions and a complex physical environment pose difficulties to a bicyclist's activities and safety, especially when the bicyclist is engaged in more risky cycling maneuvers. To gain a better understanding of the risks inherent in various cycling maneuvers and assist in road safety assessments, an efficient system for identifying cycling maneuvers is needed. This paper proposes a new set of definitions of cycling maneuvers specific to Chinese bicyclists. The cycling maneuvers were categorized into passing, avoiding, carriageway-occupied, sidewalk-occupied, and regular riding maneuvers. In addition, a convolutional neural network (CNN)-based method was developed to classify these five cycling maneuvers. Field data from a video survey in the urban area of Xi'an, China (998 records) was used to evaluate the performance of the proposed model. The data includes human-related features, road-related features, and traffic-related features (for example, the gender of the bicyclist, cycling speed, vehicle-bicycle separation, bicycle-sidewalk separation, the width of the bicycle path, and traffic volume). A promising CNN model was identified by optimizing the model configuration and adjusting the model parameters. Five prevailing methods including multi-Logit, artificial neural network, support vector machine, random forest, and gradient boosting decision tree, were selected to conduct comparison analysis with the proposed CNN model. It is found that the proposed CNN model exhibited superior performance in cycling maneuver classification task.