Search

Microvibrations of spacecraft are mainly generated by control moment gyroscopes (CMGs), which seriously reduces the effectiveness of spacecraft attitude control. Considerable work about vibration isolators has been conducted for isolating microvibrations generated from CMGs and minimize their effect. However, deformations of isolators are coupled with the motion of CMGs, which influences the output performances. We focus on identifying how the vibration of dynamic and isolator parameters affect the output characteristics of a CMG with an isolator. A theoretical model is established and the coupling principle of the CMG-isolator system is described systematically. We verify the numerical simulation using MATLAB/Simulink software, analyze the stability of this model and estimate the domain of attraction by Monte Carlo method. Finally, an SGCMG-isolator prototype is fabricated and an integrated test system is assembled to get properties of the prototype; a series of confirmatory experiments are designed and conducted. Model correction is verified via a comparison of the numerical and experimental data.

Earthquakes pose a significant threat to urban areas, necessitating accurate forecasting models to mitigate their impact. This study focuses on earthquake forecasting in Los Angeles, a region with high seismic activity and limited research. We established a feature matrix for forecasting earthquakes within a 30-day period by analyzing the most predictive patterns from recent studies. Our model developed a subset of features capable of forecasting the highest magnitude of an earthquake. Using advanced machine learning algorithms and neural networks, our model achieved an accuracy of 69.14% in forecasting the maximum magnitude earthquake as one of the 6 categories. We aim to provide a useful guideline for future researchers.

This study investigates the application of machine learning (ML) algorithms to enhance the precision of wine quality assessment, focusing specifically on Portuguese red wine. Amidst the growing interest in leveraging artificial intelligence (AI) for sensory analysis, our research distinguishes itself by employing a rigorous methodological framework. Our approach, named the ‘Incremental Analysis of Baseline Accuracy,’ identifies the chemical variables most predictive of wine quality. This framework aims to streamline the predictive process by pinpointing key variables that significantly influence quality assessments. In this paper, we demonstrate the feasibility of a methodology that precisely determines the criticality of chemical inputs, both their exact values and their correct order, to identify which inputs significantly contribute to the quality assessment of a sensory perception, such as taste. The centerpiece of our paper is a vibrant 3D pie chart that illustrates the percentage criticality of different input variables for perceiving the quality of red wine. This chart symbolizes the essence of our paper: a ‘pie’ representing the empirical conclusion, not mere conjecture. Through this paper, we have shown that it is possible to quantify a qualitative, perceptual aspect like taste perception, which is often believed to be assessable only through subjective conjecture. Moreover, our findings, facilitated by the Incremental Analysis of the Baseline Accuracy method, demonstrate that this perception can be systematically quantified, challenging traditional assumptions about sensory analysis.

This paper examines the potential of utilizing a Faraday anomalous dispersion optical filter (FADOF) to enhance noise reduction in a FMCW LiDAR (Frequency-Modulated Continuous-Wave Laser Detection And Ranging) system. While the FMCW LiDAR itself demonstrates a certain level of robustness, it may fail to operate effectively in the presence of significant background noise interference. To address this challenge, we have proposed a validation experiment utilizing FADOF technology. Specifically, we have introduced a FADOF component at the optical receiving end of the FMCW LiDAR, and we have implemented a dynamic frequency stabilization method that combines frequency stability with frequency modulation for the laser source in the transmitter. By integrating FADOF into the FMCW LiDAR system to mitigate external noise, we conduct indoor and outdoor validation experiments. In the outdoor experiment, we introduce real sunlight background noise into the system, allowing it to directly illuminate the receiver. We then filter the light using FADOF and a 10 nm filter, separately. The experimental results confirm that the FMCW LiDAR system, when equipped with FADOF, can operate normally in intense background light noise, with a noise resistance capability at least twice as effective as the 10 nm filter. Additionally, the system achieves a detection range of at least 17 meters under direct sunlight conditions. This study has the potential to extend the application of LiDAR systems to more extreme environments with strong background noise, enabling their functionality in a broader range of scenarios.

The stock market is playing an increasingly important role in the global economy. Accurate stock price forecasting not only aids the government in predicting economic trends but also helps investors anticipate higher expected returns. Nevertheless, hurdles such as nonlinearity, complexity and high volatility make it a daunting task to predict stock prices. To address this issue, this study proposes a new hybrid model, termed Hierarchical Decomposition-based Forecasting Model (HDFM), to decompose and forecast stock prices in a hierarchical fashion. The model utilizes complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) for the initial decomposition of stock price time series. To enhance the predictive efficiency, sub-series with similar sample entropy from the decomposition were combined using the K-means clustering method. Through a thorough analysis, it was found that the first combined sub-series contained more high-frequency signals. Therefore, the first combined sub-series is subjected to a second decomposition with variational mode decomposition (VMD). Subsequently, the gated recurrent unit (GRU) model was used to individually predict each sub-series. The final results were obtained by merging the prediction outcomes. The proposed model was evaluated on three different stock markets. The experimental results show that the proposed model outperforms other forecasting methods across all stock indices. Moreover, ablation studies demonstrated the effectiveness of each component within the proposed model.

False or missed detection may happen in the process of pavement defect detection due to cracks with different shapes and sizes and interference from complex pavement background. To solve this problem, we proposed a Cascade R-CNN detection method based on the MS-Feature Pyramid Network (MS-FPN). First, we introduced a deformable convolution module in the backbone network ResNet101, so that it can adaptively change depending on the pavement defect. Second, we utilized the MS-FPN for the cross-scale bi-directional fusion of feature maps output by the backbone network,in which the multi-branch hybrid dilated convolution (MCE) generates feature maps with multi-scale receptive fields while expanding the receptive field. The dual-channel attention fusion algorithm (ST-A) was used to improve the identification between the background and the object, so that more attention was paid to the location and features of the pavement defect object. The improved Cascade R-CNN can better adapt to the detection of various pavement defects. On the open-source dataset and self-built dataset, the detection precision has improved by 5.02% and 5% respectively compared to that of the original Cascade R-CNN.

The application of wind power is greatly restricted due to the volatility and intermittency of wind. It is a challenging task to quantify the uncertainty of wind speed prediction. To tackle such a challenge, an adaptive interval construction-based gated recurrent unit (GRU) model is proposed for directly generating short-term wind speed prediction intervals in this article, using the two phase search strategy to search the model parameters. Different from the traditional interval prediction techniques, in the proposed model an adaptive interval construction method is designed, where the target values of wind speed are characterized by two interval width adjustment variables which are used to determine the lower and upper bounds of the interval of wind speed. A two phase search strategy is designed to optimize the parameters. In Phase I, the dynamic inertia weight particle swarm optimization algorithm is used to search the two interval width adjustment variables. In Phase II, the GRU networks are trained using the root mean square prop (RMSProp) algorithm (an effective gradient-based optimizer) to fit the upper and lower bounds of the constructed intervals, respectively. The two phases are executed alternately, so as to obtain optimal prediction intervals. The performance of the proposed method is compared with eight other machine learning and deep learning methods, and the experimental results show that the proposed method outperforms the compared methods. It indicates that the proposed method can generate satisfactory and better prediction intervals compared with other methods.

This study proposes a novel image registration algorithm to solve the problem of low registration accuracy caused by excessive difference in resolution and spectral differences between infrared and visible images. First, we use a VI-CycleGAN network to translate visible images into fake infrared images. Furthermore, we incorporate a normalization-based attention mechanism (NAM) into each residual block to capture the global information of the image and preserve its details. Meanwhile, we consider higher-level semantic information and introduce a hybrid loss function that better preserves the content features of the original image. We then use guided filtering to process the infrared image and fake infrared image to eliminate complex background noise. Subsequently, coarse registration was performed using the Speeded-Up Robust Features (SURF) algorithm. Finally, we used the random sample consensus (RANSAC) algorithm to remove the mismatch points and achieve accurate registration. This study conducted experiments on two different VIS-IR image datasets and compared Gaussian field estimator with manifold regularization (GFEMR), radiation-variation insensitive feature transform (RIFT), and locally normalized image feature transform (LNIFT) algorithms. The experimental results show that compared with LNIFT algorithms, the registration accuracy of the proposed method was improved by approximately 5%.

Patients with early-stage Parkinson’s disease (PD) exhibit various but subtle motor symptoms, especially postural instability and gait disorders (PIGD). Patients show deteriorated gait performance at turns as the complex gait task requires more limb coordination and postural stability control, which may help to discriminate signs of early PIGD. In this study, we firstly proposed an IMU-based gait assessment model for quantifying comprehensive gait variables in both straight walking and turning tasks from five domains: respectively gait spatiotemporal parameters, joint kinematic parameters, variability, asymmetry, and stability. Twenty-one patients with idiopathic Parkinson’s disease at the early stage and nineteen age-matched healthy elderly adults were enrolled in the study. Each participant wore a full-body motion analysis system with 11 inertial sensors and walked along a path consisting of straight walking and 180-degree turns at a self-comfortable speed. A total of one hundred and thirty-nine gait parameters were derived for each gait task. We explored the factor effect of group and gait tasks on gait parameters using a two-way mixed analysis of variance. The discriminating ability of gait parameters between PD and the control group was evaluated using receiver operating characteristic analysis. Sensitive gait features were optimally screened (AUC $>$ 0.7) and categorized into 22 groups to classify PD and healthy controls based on a machine learning method. Results demonstrated that PD patients exhibited more gait abnormalities at turns, especially on the RoM and stability of the neck, shoulder, pelvic, and hip joints compared to the healthy control group. These gait metrics have good discriminating abilities to identify early-stage PD (AUC $>$ 0.65). Moreover, the inclusion of gait features at turns can significantly improve the classification accuracy compared to that only used parameters during straight walking. We show that quantitative gait metrics during turning have great potential to be used for enhancing early-stage PD detection.

With the use of sensorless control strategy, mechanical position sensors can be removed from the gearbox, so as to decrease the maintenance costs and enhance the system robustness. In this paper, a switching PI control based model reference adaptive system (MRAS) observer using Fuzzy-Logic-Controller (FLC) is introduced for sensorless control of permanent magnet synchronous motor (PMSM) drives. The main work and innovation of this paper include: 1) A disturbance observer based voltage compensation method is proposed to solve the problems of voltage distortion in voltage source inverter (VSI) fed PMSM systems. 2) A double closed-loop FLC is designed for speed and current control. The response performance of speed regulation and the accuracy of the command voltage are improved, and the difficulty of manual adjustment of control parameters is avoided. 3) A sensorless control strategy of a switching PI control based MRAS observer is proposed for stator resistance tracking. By setting different thresholds, the proposed MRAS observer can choose the appropriate PI adaptive mechanism based on the amplitude of the current error variable, which can improve the accuracy of resistance estimation and improve the dynamic performance of sensorless control. 4) The experimental results are given to verify the availability of the proposed scheme.

Accurate transient stability assessment (TSA) and effective preventive control are important for the stable operation of power systems. With the superiorities in precision and efficiency, data-driven methods are widely used in TSA nowadays. Data-driven TSA model can be adopted in the stability constraints of preventive control optimization, but existing methods are mostly iteration-based ones, which may result in low efficiency, sometimes even non-convergence. In this paper, an analytical representation method of data-driven transient stability constraint is proposed based on a non-parametric regression model built for TSA. Key feature extraction and dominant sample selection are proposed to reduce the scale of the TSA model, and bi-level linearization is applied to further modify it. Optimal preventive control model is then formulated as a mixed-integer linear program (MILP) problem with the linearized analytical data-driven transient stability constraint, which can be solved without iterations. An overall procedure of data-driven TSA and preventive control is finally developed. Case studies show that the proposed method has high accuracy in TSA and can achieve effective preventive control of power system with high efficiency.

The full-bridge pulse-width-modulation (PWM) inverter is wildly applied to drive the electromagnet and enables advanced technologies, such as the active magnetic bearing (AMB) and the permanent-electro magnetic suspension (PEMS). However, the characteristic relationship between the current through the electromagnet and the duty cycle of the PWM signal has strong nonlinearity around the zero current. Moreover, the electromagnet possesses the induction that results in the time constant and significantly hinders the change of the current. Hence, the open-loop control of the full-bridge PWM inverter cannot accurately or timely tune the current through the electromagnet, especially around the zero current. This work proposes a closed-loop current control approach by three steps: (1) inserting a current-sensing resistor into the middle of the electromagnet, (2) obtaining the current signal with an analog signal-processing circuit, and (3) generating the PWM signal with the bang-bang control circuit. The proposed approach innovatively arranges the current-sensing resistor to take advantage of the symmetry and to minimize the influence from the high-frequency switching of the inverter, so that the measured current signal is comparable to the hall-effect current sensor. The experimental results demonstrate the effectiveness and efficiency of the proposed closed-loop current control approach, though the weak charging capability of the full-bridge PWM inverter still hinders its performance.

Adding the auxiliary frequency control function to the wind turbine generator (WTG) is a solution to the frequency security problem of the power system caused by the replacement of the synchronous generator (SG) by the WTG. The auxiliary frequency control using rotor kinetic energy is an economical scheme because the WTG still runs at the maximum power point during normal operation. In this paper, the functional optimization model of the auxiliary frequency control strategy of WTG is established. The optimal auxiliary frequency control strategy is obtained by solving the model numerically. As for the practical realization of the control strategy, the coordination of the auxiliary frequency control with the maximum power point tracking (MPPT) control is studied. The practical auxiliary frequency control strategy is modified to adapt to different power disturbances in the system, and the parameter setting method is also proposed. The sensitivity of system frequency to control parameters is studied. Finally, the simulation results verify the effectiveness and practicability of the proposed control strategy.

The permanent-electro magnetic suspension (PEMS) technology takes advantage of the attractive magnetic force between the magnet and the iron core and reduces the power consumption eventually to zero. However, the active current in the electromagnet of the zero-power PEMS system fluctuates around zero due to external disturbances and suffers from the electronic nonlinearity of the driving circuit, such as the full-bridge pulse-width-modulation (PWM) inverter. This work presents that the 2 $\mu \text{s}$ turn-off delay (one electronic defect) of the integrated circuit L298N (one commercial full-bridge PWM inverter produced by STMicroelectronics) leads to the nonlinear current-duty cycle characteristic, which undermines the control stability and limits the PWM frequency of the zero-power PEMS system. Moreover, the nonlinear mechanism is experimentally and theoretically analyzed for the critical PWM frequency and the sensitivity transition with respect to the $2~\mu \text{s}$ turn-off delay. Besides, the critical PWM frequency is of great significance for the energy efficiency and the dynamic performance of the high-speed PEMS transportation system.

Heat assisted interlaced magnetic recording (HIMR) is a promising candidate for the next-generation of magnetic recording technology to further increase the area density beyond 1Tb/in2. Specifically, the high temperature and low temperature tracks are written in an interlaced order to improve the recording performance. However, the inter-track interference (ITI), inter-symbol interference (ISI) and thermal jitters brought by the increased recording density and Curie temperature variations are severe in HIMR, which degrade the bit error rate (BER) performance obviously. In this study, we propose a multitrack detection scheme with joint intertrack interference and media noise mitigations. Here a multi-task neural network (MTNN) is designed to simultaneously predict ITI pattern and residual media noise, then the 2D variable equalizers corresponding to different ITI patterns are implemented and predicted residual media noises are embedded into the branch metrics of modified Bahl-Cocke-Jelinek-Raviv (BCJR) detector to mitigate ITI and whiten media noise. The simulation demonstrates that the proposed MTNN with variable equalizer and modified BCJR detector (MTNN+VE+MB) algorithm mitigates the ITI and media noise effectively. At the channel bit density of 3.10 Tb/in2, it provides 2.6 dB signal-to-noise ratio (SNR) gain compared to that of conventional 2D fixed equalizer with pattern dependent noise prediction detector (FE+PDNP) for the low temperature (LT) tracks with 4% Curie temperature variance.

We presented a machine learning-based method to recognize suspension by distinguishing dispersoid-dependent speckle patterns using a convolutional neural network. The dispersoid size and concentration-related transmissive speckle patterns were recorded by a lensless camera when a coherent He-Ne laser irradiated the suspension. Firstly we realized the recognition of the polystyrene microspheres-dispersed suspensions with different particle sizes and the recognition of several common suspensions including protein powder and milk powder with similar concentration. Further recognition from three different food suspensions with unknown concentration was achieved with high accuracy of 99%. The experimental results confirm that this non-contact method could find applications in the measurement and classification of suspension liquid containing micrometers-sized particles.

The automatic recognition of traffic signs is essential to autonomous driving, assisted driving, and driving safety. Currently, convolutional neural network (CNN) is the most popular deep learning algorithm in traffic sign recognition. However, the CNN cannot capture the poses, perspectives, and directions of the image, nor accurately recognize traffic signs from different perspectives. To solve the problem, the authors presented an automatic recognition algorithm for traffic signs based on visual inspection. For the accuracy of visual inspection, a region of interest (ROI) extraction method was designed through content analysis and key information recognition. Besides, a Histogram of Oriented Gradients (HOG) method was developed for image detection to prevent projection distortion. Furthermore, a traffic sign recognition learning architecture was created based on CapsNet, which relies on neurons to represent target parameters like dynamic routing, path pose and direction, and effectively capture the traffic sign information from different angles or directions. Finally, our model was compared with several baseline methods through experiments on LISA (Laboratory for Intelligent and Safe Automobiles) traffic sign dataset. The model performance was measured by mean average precision (MAP), time, memory, floating point operations per second (FLOPS), and parameter number. The results show that our model consumed shorter time yet better recognition performance than baseline methods, including CNN, support vector machine (SVM), and region-based fully convolutional network (R-FCN) ResNet 101.

Although performing tunnel excavation using tunnel boring machines (TBMs) is rapid, it is susceptible to the unexpected geological changes located in front of the tunnel face. Seismic detection methods are useful to detect local geological conditions in front of the tunnel face. However, manual detection is time consuming and may interfere in realizing rapid excavation by TBMs. Therefore, this study proposes an automated seismic detection system that is mounted on a TBM for rapidly imaging anomalous zones in front of the tunnel face. The system consists of an automated data acquisition system and a data processing system. The automated data acquisition system allows reducing the data acquisition time. To reduce the data processing time, the data processing system images the anomalous zones based on a modified Coppens' method. The proposed system has been mounted on TBMs and verified based on field tests.

Transient stability assessment (TSA) is of great importance in power systems. For a given contingency, one of the most widely-used transient stability indices is the critical clearing time (CCT), which is a function of the pre-fault power flow. TSA can be regarded as the fitting of this function with the pre-fault power flow as the input and the CCT as the output. In this paper, a data-driven TSA model is proposed to estimate the CCT. The model is based on Mahalanobis-kernel regression, which employs the Mahalanobis distance in the kernel regression method to formulate a better regressor. A distance metric learning approach is developed to determine the problem-specific distance for TSA, which describes the dissimilarity between two power flow scenarios. The proposed model is more accurate compared to other data-driven methods, and its accuracy can be further improved by supplementing more training samples. Moreover, the model provides the probability density function of the CCT, and different estimations of CCT at different conservativeness levels. Test results verify the validity and the merits of the method.

Based on a Double Deep-Q Network with deep ResNet (DDQN-ResNet), this paper proposes a novel method for transmission network expansion planning (TNEP). Since TNEP is a large scale and mixed-integer linear programming (MILP) problem, as the transmission network scale and the optimal constraints increase, the numerical calculation and heuristic learning-based methods suffer from heavy computational complexities in calculation and training. Besides, due to the black box characteristic, the solution processes of the heuristic learning-based methods are inexplicable and usually require repeated training. By using DDQN-ResNet, this paper constructs a high-performance and flexible method to solve large-scale and complex-constrained TNEP problem. Firstly, we form a two-objective TNEP model, in which one objective is to minimize the comprehensive cost, and another objective is to maximize the transmission network reliability. The comprehensive cost takes into account the expansion cost, the network loss cost, and the maintenance cost. The transmission network reliability is evaluated by the expected energy not served (EENS) and the electrical betweenness. Secondly, from the TNEP model, the TNEP task is constructed based on the Markov decision process. By abstracting the task, the TNEP environment is obtained for DDQN-ResNet. In addition, to identify the construction value of lines, an agent is establish based on DDQN-ResNet. Finally, we perform the static planning and visualize the reinforcement learning process. The dynamic planning is realized by reusing the training experience. The validity and flexibility of DDQN-ResNet are verified on RTS 24-bus test system.

Applications of connected vehicles (CVs) will be widely deployed in the near future owing to the rapid development of vehicle-to-everything (V2X) communication technology. However, when CVs are running in a real traffic environment, they may encounter some problems (i.e., traffic adaptability problems) that are not noticed in simulations, hardware/software-in-the-loop tests, and closed area tests. This leads to an unexpected poor performance of CV technology in a real traffic environment. Therefore, there is a need to test and assess CV technology in a real traffic environment before the deployment of CV applications. This study concentrates on the platoon driving scenario to explore the testing and assessment methods for the traffic adaptability of CV applications. The concept of traffic adaptability, which includes the aspects of efficiency, safety, and comfortableness, is defined, and an indicator system of traffic adaptability is established as the approval standard of CV traffic adaptability. A dedicated and decentralized field test system that can provide a simple CV environment was developed and used for field tests in real urban traffic environments in Beijing. Based on the data acquired from the field test, a traffic adaptability assessment of CV under platoon driving scenario utilizing the relevant indicator system is executed. The results show that with the assistance of CV technology provided by the developed field test system, that is, CV-DAVS, the traffic adaptability values of the CV platoon on the efficiency, safety, and comfort aspects are 109.98%, 113.07%, and 103.85%, respectively. Therefore, the traffic adaptability of CV under platoon driving scenario can be approved, and its overall traffic adaptability is 108.97%.

The increasing use of renewable generation, power electronic devices, and nonlinear loads in power systems brings more severe interharmonic tones to the measurand, which can increase estimation errors of P-class PMUs, cause misoperation of protection relays, and even threaten the stability of the power systems. Therefore, the performance of the P-class synchrophasor estimator under interharmonic interference should be evaluated and new estimation schemes that can improve the operational robustness are needed for various protection and control applications. In this paper, a synchrophasor estimator design for P-class PMUs that introduces singular value decomposition to the least square algorithm based on the Taylor series is proposed. By constructing an optimization with proposed adjustable parameters, finite impulse response filters with high transition band attenuation are designed. Numerical and experimental tests verify the proposed dynamic synchrophasor estimator performance and show an effective improvement in rejecting interharmonic tones, especially when a short time window and light computational burden are required.

Ground subsidence caused by coal mining may lead to serious damage to ecological environment system and economic and social development. SAR image interferometry technology has been widely used for monitoring coal mining subsidence. However, at present monitoring fast and large subsidence using SAR images is still a huge challenge. In order to obtain fast and large gradient subsidence using SAR images, some scholars have resorted to other methods and technology, which do not just depend on the interferograms, and they have obtained better results. In summary, these methods and technology can be grouped into three main categories: (i) Combining InSAR with coal mining subsidence prediction models; (ii) Using SAR image amplitude information offset tracking technology; (iii) Combining InSAR with terrestrial 3D laser scanning technology. In this paper, the technology features, research progress, applications and limitations of these methods are analyzed; and some potential prospects of using SAR images to monitor fast and large subsidence in coal mining areas are discussed in the end.

To design a deep learning network model for automatic human blastocyst quality assessment with multifocal images, a total of 11,275 images of 1,025 normally fertilized blastocysts underwent traditional in vitro fertilization (IVF) treatment between May 2017 and August 2018. Images captured 116± 1 hours after insemination were classified according to embryo grade as follows: good, poor. Multiple images with different focal depths were designed as a batch input to the proposed deep learning network so that the network had more information about the blastocysts and higher accuracies were expected. In our dataset, the area under the curve (AUC) of the best model reached 0.936. Using the Grad-CAM algorithm to visualize all blastocyst stage images on the test set, it was found that key features relied on by the classification model were trophectoderm (TE), inner cell mass (ICM) and zona pellucida (ZP). A deep learning approach provides a new method for assessing blastocyst quality. The multichannel combination network showed the best performance among the three proposed models, in which the clearest images from multiple focal depths were chosen and fetched into the VGG-16 network as different channels.

To address the problem that the traditional human motion attitude detection process is easy to ignore the data calibration, which leads to the problems of long running time, low accuracy and poor detection effect, a human motion target attitude detection algorithm based on semi-supervised learning in the Internet of things environment is proposed. Firstly, human motion target images are collected using the Internet of things (IoT), human motion attitude features are extracted based on the eight-star model, and multi-features are fused to form image blocks of 17-dimensional feature vectors. Then, random fern classifiers are optimized and semi-supervised learning is used to calculate a large number of uncalibrated data in time domain, spatial domain and data. The classifier is trained to complete image block classification. Finally, the classifier parameters are updated iteratively to complete the attitude detection of human motion target. The results show that the proposed algorithm has high accuracy in human motion attitude extraction and multi-feature fusion, and has a high correct classification rate for different feature poses, as high as 92.5%. The average F value of human motion attitude detection is 0.95, the overlap ratio is high and the time is short. The overall performance is good.

A new X-band retrodirective antenna based on superheterodyne mixer and phase conjugating technology is proposed for inter-satellite data transmission in this paper. The novel phase conjugating circuit, utilizing the reciprocity of the passive mixers, achieves high RF-IF and LO-IF isolations, high communication link gain, and a very compact low profile layout with all antenna elements and circuit components arranged at one common layer of a double-sided printed circuit board (PCB). Its mathematical principles of circuit design are derived, and the effect of up and down conversion loss deviation of mixers on the amplitude and phase error of the circuit, thus on the beam pointing error and the side-lobe of the antenna is discussed. A prototype of circularly polarized one dimension retrodirective antenna is fabricated, and its bistatic RCS measurement in an anechoic chamber is implemented to verify the effectiveness of the proposed antenna array. The results show that, within the scanning range from -30° to +30°, the array has good retrodirective performance. The proposed antenna would be a good choice for the inter-satellite data transmission inside a distributed satellite cluster.

Multilabel learning handles the problem that instances are associated with multiple labels. In practical applications, multilabel learning often suffers from imperfect training data. Typically, labels may be noisy or features may be corrupted, or both. Most existing multilabel learning models only consider either label noise or feature noise. Theoretically, ignoring any kind of noise in the learning process may lead to an unreasonable model, and thus affect the multilabel learning performance. In this paper, we propose a robust multilabel learning model, Tri-structured-Sparsity induced Joint Feature Selection and Classification (TriS-JFSC), to handle the data with hybrid noise. Specifically, the proposed TriS-JFSC model employs the tri-structured-sparsity regularization bridged with a label enhancement matrix to simultaneously smooth the feature and label noise, and embed a feature selection scheme that can simultaneously learn label-shared features and label-specific ones to boost the multilabel learning performance. Furthermore, by employing Alternating Direction Method of Multipliers (ADMM) method, a simple but efficient optimization algorithm is designed to solve the proposed TriS-JFSC model. Finally, the extensive experiments performed on several benchmark datasets demonstrate that our TriS-JFSC model outperforms other state-of-the-art learning methods.

This paper proposes a novel tiered type permanent magnet spherical motor (T-PMSpM), which has the advantage of relatively larger torque ability compared with most existing spherical motors. The proposed T-PMSpM is capable of continuously rotating while the rotating shaft can tilt in all directions. Considering that it is usually much difficult to measure the multi-DOF rotor orientation of a spherical motor in an easy and convenient way, this paper also proposes a specific rotor orientation measurement method, aiming at providing an idea of simple solution for detecting the multi-DOF rotor orientation based on phase difference. In order to clearly reveal the intrinsic relationship between the phenomenon of phase shift and the tilting motion, a simplified back-EMF model of the T-PMSpM is put forward according to the geometric changes as the rotor tilts. Based on the simplified back-EMF model, both the initial phase and the magnitude of the back-EMF are calculated analytically and verified by the 3D FEM. The full region forward relationship between the rotor tilting position and the phase difference is further obtained. Therefore, the rotor tilting position can be easily identified based on the inverse solution. Finally, a prototype motor of the investigated T-PMSpM has been manufactured. The experimental results validate that the multi-DOF orientation can be detected effectively with enough accuracy.

This paper concerns a novel data-based policy iteration control method for discrete-time systems with disturbances. To avoid the need of system dynamic and state, a new system equation is constructed by using system operation data. Based on this, we derive a new data-based Bellman equation that can be applied to the design of model-free policy iteration algorithm. For discrete-time system with disturbances, we transform H∞ control into a zero-sum game problem, and then use the proposed new data-based PI method to solve it. Wind turbine control is a typical control problem with time delay and disturbance. Mathematical simulation results are given to demonstrate the effectiveness of the proposed data-based PI control method.

In recent years, with the development of machine learning, especially after the rise of deep learning, time series clustering has been proven to effectively provide useful information in cloud computing and big data. However, many modern clustering algorithms are difficult to mine the complex features of time series, which is important for further analysis. Convolutional neural network provides powerful feature extraction capabilities and has excellent performance in classification tasks, but it is hard to be applied to clustering. Therefore, a similarity measurement method based on convolutional neural networks is proposed. This algorithm converts the number of output changes of the convolutional neural network in the same direction into the similarity of time series, so that the convolutional neural network can mine unlabeled data features in the clustering process. Especially by preferentially collecting a small amount of high similarity data to create labels, a classification algorithm based on the convolutional neural network can be used to assist clustering. The effectiveness of the proposed algorithm is proved by extensive experiments on the UCR time series datasets, and the experimental results show that its superior performance than other leading methods. Compared with other clustering algorithms based on deep networks, the proposed algorithm can output intermediate variables, and visually explain the principle of the algorithm. The application of financial stock linkage analysis provides an auxiliary mechanism for investment decision-making.

A control moment gyroscope (CMG) is a key actuator for spacecraft's attitude and stability control. But there exist micro-vibrations from high-speed flywheel, which will affect the performance of the spacecraft greatly. Due to the unbalanced mass and the supporting structure, the high-speed rotating flywheel could generate disturbance torque, which will be passed to the gimbal to deteriorate the gimbal servo system. At the same time, the rotation of the gimbal will increase the disturbance torque for gyroscopic effect. According · to the bearing theory, the stiffness of the bearing is affected by the load that applied on it, which results the influence on the flywheel vibration and the generated disturbance torque. This means that there is a coupling effect in the CMG. In this article, we focused on this coupling relationship between the flywheel vibration and the gimbal rotation in a single gimbal control moment gyroscope (SGCMG) system. Firstly, a vibration model of the flywheel with variable stiffness supporting is established. And the rotation model of the gimbal is also developed. The flywheel model and the gimbal model are coupled with each other through the variable stiffness supporting and gyroscope effect. Based on the numerical simulations the interactions among the bearing stiffness, the flywheel vibration and the gimbal rotation are analyzed. Finally, experiments are carried out on the prototype. The experimental results have verified the theoretical analysis and the simulations, and which would support the design of high-performance control system.

Open shop scheduling problems have many practical applications, in which the jobs can be in any order with the only restriction that their durations do not overlap with each other. In some cases, the order of a subset of jobs dictates the flow shop model while the remaining jobs can be processed in any order according to the open shop model, and the corresponding problem is called mixed shop scheduling. For every job, if its processing times on all machines are the same, then the shop is called proportionate shop. In this article we investigate the three-machine proportionate open shop and mixed shop minimum makespan problems proposed by Koulamas and Kyparisis. Our result improves the previous work in three ways. First, for both the open shop and the mixed shop scheduling problems, we derive additional sufficient conditions under which the corresponding problems are solvable in polynomial time. Second, for the non-solvable cases of the open shop scheduling problem we present an improved 13/12-approximation algorithm. Third, for the non-solvable cases of the mixed shop scheduling problem we present an improved approximation algorithm with the worst-case ratio bound of (7/6 + ϵ). All the algorithms proposed in this article run in polynomial time.

Metabolic pathways refer to the continuous chemical reactions in the metabolic process in vivo. Compounds are the major participant for most metabolic pathways. It is essential to determine which compounds can constitute a metabolic pathway. This problem can be converted to the identification of the metabolic pathways of compounds. Although traditional experiments can provide solid results, they are always of low efficiency and high cost. To date, several machine leaning models have been proposed to address this problem. However, almost all models only identified metabolic pathway types of compounds rather than actual metabolic pathways. This study proposed a novel model for predicting actual metabolic pathways for given compounds. The pairs of compounds and metabolic pathways were termed as samples, thereby modeling a binary classification problem. With the concept of “similarity”, each sample was represented by seven features, extracted from seven associations of compounds, which measure compound linkages from different aspects. The model adopted random forest as the classification algorithm. Two types of ten-fold cross-validation were adopted to evaluate the performance of the model, indicating its utility. A feature analysis was also performed to determine which compound association was highly related to the identification of metabolic pathways of compounds.

Transient stability assessment (TSA) is of great importance in power system operation and control. One of the usual tasks in TSA is to estimate the critical clearing time (CCT) of a given fault under the given network topology and pre-fault power flow. Data-driven methods try to obtain models describing the mapping between these factors and the CCT from a large number of samples. However, the influence of network topology on CCT is hard to be analyzed and is often ignored, which makes the models inaccurate and unpractical. In this paper, a novel data-driven TSA model combining Mahalanobis kernel regression and ensemble learning is proposed to deal with the problem. The model is a weighted sum of several sub-models. Each sub-model only uses the data of one topology to construct a kernel regressor. The weights are determined by both the topological similarity and numerical similarity between the samples. The similarities are decided by the parameters in Mahalanobis distance, and the parameters are to be trained. To reduce the model complexity, sub-models within the same topology category share the same parameters. When estimating CCT, the model uses not only the sub-model which the sample topology belongs to, but also other sub-models. Thus, it avoids the problem that there may be too few data under some topologies. It also efficiently utilizes information of data under all the topologies. Moreover, its decision-making process is clear and understandable, and an effective training algorithm is also designed. Test results on both the IEEE 10-machine 39-bus and a real system verify the effectiveness of the proposed model.

In recent years, a surging development of vehicles and continuous enhancement of transportation infrastructures have been witnessed worldwide, leading to a remarkable growing of traffic flow data. The traffic data is highly valuable in today's society, accurate modelling of traffic flow for the concerned areas can significantly benefit the government agencies, related commercial departments and individuals. Specifically, road users are allowed to make better traveling decisions, avoid traffic congestion, reduce carbon emissions and improve traffic operation efficiency. In order to estimate the possible traffic flow scenarios within a specific area for multiple horizons, we propose a scenario generation model based on sequential generative adversarial networks (LSTM-GAN) where the long short term memory (LSTM) network is incorporated to capture the temporal dynamics involved in traffic flows. Through game training, the spatiotemporal scenarios of traffic flow in line with the characteristics of observed road network traffic flow can be well generated. These traffic scenarios can be applied in the design and planning of road traffic system, as well as in the virtual training cases of intelligent driving.

With the development of the EV industry, the growing demand for EV fast charging load has brought enormous impact on the power system. Due to the unbalance of charging time distribution and the high power of EV fast charging, the difference between peak and valley of the load curve will widen. In this paper, a hierarchical navigation strategy (HNS) based on dynamic traffic/temperature data is proposed to decrease the EV fast charging load at peak hours and the time and energy cost during the charging process. The upper layer of the HNS is charging time selection. The optimal selections of charging time, which is based on the habits of EV users, is proposed in this layer. It aims at providing efficient time slots for charging, which can decentralize the fast charging demand and decrease the EV users' time cost. The underlayer is the route selection layer, which is based on the priority coding genetic algorithm. It proposes the optimal charging routes to decrease EV users' energy cost and time cost. At the same time, the peak charging load can also be shaved due to the decline in energy cost. The case study under the scene with realistic traffic, temperature, and power grid information shows that the proposed HNS can shave the peak load of the power grid and decrease the energy/time cost during the EV fast charging process. Therefore, the effectiveness of the HNS is proved.

In this paper, a new clustering algorithm is proposed based on cross clusters without using membership functions. In light of the cross clustering data transformation, the spatial distribution of data is changed while the original data dimension simultaneously is maintained. Combining with the performance index and visual technology, an explanation of the performance improvement of the classification model is presented in accordance with the proposed algorithm. This approach was evaluated on UCR time series datasets, the experiments showed that the algorithm can improve not only the accuracy and the performance of the fully convolutional network and nearest neighbor algorithm, but also the time complexity in time series classification model. It is worth well to apply this method to further research and popularization.

In this paper, factors that account for the disc brake noise were explored with a bench test to fix the methods of reducing the brake squeal noise. The influence of the stiffness and damping of insulator on the noise of the braking system was analyzed by establishing a mathematical model for the 6-DOF braking system with internal and external insulators. It is proposed that the insulator with a multilayered structure of different materials has a better restraining effect on brake squeal. Four different insulator schemes were designed, the test results of which were compared with those of the original scheme. The comparative analysis shows that the stiffness and damping of insulator have an important influence on the braking system noise, which can be reduced by improving the material parameters of the structure layer of the insulator. The test results show that the insulator with one layer of rubber and one layer of steel has a good effect on restraining the low-frequency noise, while the insulator with two layers of rubber and one layer of steel has a good inhibition on the high-frequency noise around 7~8 kHz. The final conclusion shows that a rubber-steel-rubber structure insulator made of a 0.5-mm-thickness steel sheet can reduce the overall incidence of the brake system noise from 19.27% to 1.27%. An appropriate selection of material structure parameters of the insulator can minimize the occurrence of brake squeal.

Driven by the increasing number of connected vehicles and related services, powerful communication and computation capabilities are needed for vehicular communications, especially for real-time and safety-related applications. A cellular network consists of radio access technologies, including the current long-term evolution (LTE), the LTE advanced, and the forthcoming 5th generation mobile communication systems. It covers large areas and has the ability to provide high data rate and low latency communication services to mobile users. It is considered the most promising access technology to support real-time vehicular communications. Meanwhile, fog is an emerging architecture for computing, storage, and networking, in which fog nodes can be deployed at base stations to deliver cloud services close to vehicular users. In fog computing-enabled cellular networks, mobility is one of the most critical challenges for vehicular communications to maintain the service continuity and to satisfy the stringent service requirements, especially when the computing and storage resources are limited at the fog nodes. Service migration, relocating services from one fog server to another in a dynamic manner, has been proposed as an effective solution to the mobility problem. To support service migration, both computation and communication techniques need to be considered. Given the importance of protocol design to support the mobility of the vehicles and maintain high network performance, in this paper, we investigate the service migration in the fog computing-enabled cellular networks. We propose a quality-of-service aware scheme based on the existing handover procedures to support the real-time vehicular services. A case study based on a realistic vehicle mobility pattern for Luxembourg scenario is carried out, where the proposed scheme, as well as the benchmarks, are compared by analyzing latency and reliability as well as migration cost.

In approximate processing on stream data, most works focus on how to approximate online arrival data. However, the efficiency of approximation needs to consider multiple aspects. Generally, customers submit their requests with specific quality requirements (e.g., maximum error). This raises a critical problem that online quality control is required to meet the desired quality of service. Since the continuous arriving data may not be entirely stored and needs to be processed immediately, it brings the difficulty of acquiring knowledge online which significantly affects the quality of results. To address these problems, we present an online adaptive approximate processing framework with a delicate combination of data learning, sampling, and quality control. We first design an online data learning strategy for stream data. With the real-time learning results, we propose a dynamic sampling strategy that switches to different sampling methods based on the change of the load. Finally, we present a double-check error control strategy to monitor and correct large errors. Each operation module is correlated through online learning and feedback. The experiments with both synthetic and real-world datasets show that the proposed approximate framework is not only applicable to different data distributions but also provides a customized error control.

Accurate fault diagnosis of high-voltage circuit breakers is crucial for the safety of power grids. Mechanism action time, though can reflect the state of high-voltage circuit breakers, is usually difficult to obtain, as multiple sensors are required for acquisition. To solve this problem, a novel method that can extract mechanism action time from vibration signals was proposed in this paper. The method involved enhancement of vibration signals using short-time zero-crossing rate and determination of mechanism action time by the double threshold method. Then, time parameters were utilized to calculate singular spectrum entropy as the input vector for a classifier. Finally, a hybrid classifier that involved support vector data description and extreme learning machine was developed to identify faults. The hybrid classifier can not only classify known mechanical states but also detect unknown mechanical faults of high-voltage circuit breakers. The effectiveness of the proposed method was verified using a 35-kV high-voltage circuit breaker.

Parallel three-level simplified neutral point clamped inverters (3L-SNPCIs) are applied in this paper because of fewer switching devices, low harmonic, larger power capacity and more flexibility. However, the challenge of zero-sequence circulating current (ZSCC) is inevitable, which causes the output current distortion and system instability. The conventional control and modulation methods introduced by far cannot solve them properly due to the limitations of fewer voltage vectors selection. To overcome these limitations, a modified model predictive control (MMPC) strategy is proposed based on the discrete mathematical model of parallel 3L-SNPCIs. The MMPC strategy realizes circulating current suppression and neutral-point voltage balancing by selecting P-type or N-type small voltage vector. In addition, the control objectives are optimized and the implementation is simplified, which can reduce the burden of calculation and reserve the advantages of the conventional model predictive control (MPC) algorithm. The proposed control strategy can be flexibly extended to parallel multilevel inverters. The efficiency of the proposed method is proved by simulation and experimental results under various conditions.

Automatic segmentation of skin lesions in dermoscopy images is a challenging task due to the large size and shape variations of the lesions, the existence of various artifacts, the low contrast between the lesion and the surrounding skin. In this paper, we propose a novel Attention Based DenseUnet network (referred as Att-DenseUnet) with adversarial training for skin lesion segmentation. Att-DenseUnet is a Generative Adversarial Network which contains two major components: Segmentor and Discriminator. In the Segmentor module, we propose an architecture which is similar to DenseNet in the down-sampling path to ensure maximum multi-scale skin lesions information transfer between layers in the network at dense scale range, meanwhile, we design an attention module to automatically focus on the skin lesion features and suppress the irrelevant artifacts features in the output feature maps of the DenseBlocks. In the Discriminator module, we employ adversarial feature matching loss to train the Segmentor stably, force the Segmentor to extract multi-scale discriminative features, and guide the attention module focusing on the multi-scale skin lesions. A novel loss function of the Segmentor is proposed which combines the jaccard distance loss with the adversarial feature matching loss introduced by the Discriminator. We trained the proposed Att-DenseUnet on ISBI2017 dataset. The test results show that our approach gains the state-of-the-art performance, especially for JAC (0.8045) and SEN (0.8734) scores which are significantly improved by 2.2% and 1.9%, respectively, also our network is robust to different datasets, and gains the lowest time cost which make our network suitable for clinical application.

Recently, it has become a main consensus that early diagnosis of benign prostatic hyperplasia (BPH) is meaningful in delaying disease progression. Although ultrasound has been well recognized as the most useful early screening technique, the lack of a standard observation plane makes it difficult to determine the BPH stages in a way that is operator-independent and satisfies clinical expectations. Strong enthusiasm has fueled intense research on ultrasound coupled with artificial intelligence (AI) for automated or semiautomated classification of different clinical diseases. Therefore, it is crucial to formulate standard planes for ultrasound imaging with the implementation of AI to evaluate BPH. This study was a preliminary examination of AI-assisted ultrasound in the prostate disease diagnostic classification task, which aimed to define and assess the clinical value of the length of the proximal prostatic urethra (Lppu), detected via transrectal ultrasound within the uniquely calibrated standardized plane, in 643 patients suffering from BPH between January 2015 and December 2016. International Prostate Symptom Score (IPSS), quality of life (QoL) and PSA levels were also recorded. The mean (standard deviation) of Lppu was 28.54 mm (9.19 mm), and there were significant differences in the mean Lppu among men who were over 50 yrs old. Lppu was colinear with IPSS, voiding-IPSS, storage-IPSS and QoL with individual r values of 0.13, 0.09, 0.04, and 0.03, respectively (P

Energy storage equipment is an important part of integrated energy systems, but the construction and operational costs of it are great. Therefore, it's difficult to apply energy storage equipment on a large scale. As an alternative measure, adjusting controllable loads including electric vehicles (EVs) and air conditioning loads jointly can transfer loads across time periods and reduce variation difference between heat and electric loads. These controllable loads can be regarded as “virtual energy storage system”. Managing the charging of EVs and heat storage of buildings, a joint virtual energy storage system including electric energy storage and thermal energy storage is proposed in this paper. Then, these models are integrated into a scheduling model for regional integrated energy system (RIES). The charging/discharging power management of joint virtual energy storage systems can be realized by arranging the charging of EVs based on vehicle-driven rules and by adjusting building indoor temperatures within the temperature comfort range. Finally, simulations are carried out to demonstrate the effectiveness of the scheduling method. Simulation results show that the RIES optimal scheduling model considering joint virtual energy storage can reduce operational costs. The stability of RIES is improved with the combination of virtual electrical and thermal energy storage.

Local community detection, only considering the regional information of the large network, can be used to identify a densely connected community containing the seed node in a network, aiming to address the efficiency problem faced by global community detection. However, most existing studies in local community detection did not account for the higher-order structures crucial to the network, but rather have simply focused single nodes or edges. Moreover, existing higher-order solutions are not purely local methods, as they still use global search to find the best local community, which leads to a global search problem. Furthermore, the quality of the detected community depends on the location of the seed node, which leads to a seed-dependent problem. Thus, in this paper, we proposed a fuzzy agglomerative algorithm (FuzLhocd) for local higher-order community detection based on different fuzzy membership functions. To solve the global search problem, we introduce a novel, purely localized metric called local motif modularity. Based on this local metric, FuzLhocd only needs to visit a limited number of neighborhoods around the seed node. To solve the seed-dependent problem, we systematically studied the formation of the local community, divided the process of local community detection into three stages and employed various fuzzy membership functions at different stages. Our extensive experiments based on both real-world and synthetic networks demonstrated that FuzLhocd not only runs efficiently locally but also effectively solves the seed-dependent problem and achieves a high accuracy as well. We concluded that our local motif modularity metric and FuzLhocd algorithm is highly effective for local higher-order community detection.

An embryo develops from a single-celled zygote, which produces a multi-cellular organism by mitosis. Due to the complication of processes and mechanisms, research on embryo cell clusters in different early embryo developmental stages with significant phenotypic differences is still lacking. In this work, we identified some gene characters and expression rules to classify these individual cells using several advanced computational methods. The single cell expression profiles of embryo cells were analyzed by the Monte Carlo feature selection (MCFS) method, resulting in a feature list. Then, the incremental feature selection (IFS) method, incorporating support vector machine (SVM), applied on such list to extract key gene characters. These gene characters include KHDC1, HMGN1, DCP, GDF9, RNF11, DNMT3L, and CDX1. Furthermore, a rule learning algorithm, Repeated Incremental Pruning to Produce Error Reduction (RIPPER), was applied to the informative features yielded by MCFS method, producing a group of classification rules. These rules can clearly uncover different expression patterns on cells in different stages. This study provided a group of effective gene signatures and rules for embryo cell subtyping and presented an applicable computational tool to further dig into the regulatory mechanisms of embryo development.

Dynamic voltage restorers (DVRs) are devices that compensate for voltage sags/swells in power distribution systems effectively. However, the performance of DVRs may be easily endangered by large load-circuit currents. This work proposes a dual-functional DVR (DFDVR) that solves multiple problems by controlling the trigger angle of antiparallel thyristors, which are connected between the power converter and output filter. The DFDVR not only compensates for voltage fluctuation in the upstream grid but also limits fault current in the downstream grid. Moreover, the combined application of the DFDVR with a recloser is introduced. The DFDVR flexibly regulates the amplitude of the faulty current to coordinate the recloser. In addition, the coordination sequence diagrams of the DFDVR and the recloser under instantaneous and permanent faults are analyzed in detail. Simulations and experimental results verify the correctness and feasibility of the operations and coordination of the DFDVR and the recloser.

Membrane protein is an important type of proteins and has been confirmed to play essential roles in various cellular processes. Based on their intramolecular arrangements and positions in a cell, they can be categorized into several types. However, it is time- and cost-consuming to recognize the type of a given membrane protein via traditional biophysical methods. In view of this, several computational models have been proposed in recent years. Most models adopted various information of membrane proteins, such as their sequences, domain profiles, physiochemical properties, etc. to extract different features, which were fed into downstream classification algorithms. In this study, we built two novel prediction models, which incorporated novel feature extraction methods, i.e., network embedding methods. To this end, several protein networks were constructed using the protein-protein interaction information retrieved from STRING. Among these models, one model was constructed based on features obtained by applying Mashup on seven protein networks, another model was built using features yielded by Node2Vec on one comprehensive protein network. Each model adopted random forest as the classification algorithm and employed the Synthetic Minority Over-sampling Technique (SMOTE) to overcome the influence yielded by the great difference on sizes of different membrane protein types. Furthermore, two models were integrated into one model to improve the predicted quality. The test results shown that the integrated model had good performance and was superior to any individual model. Also, we compared our models with some previous models, suggesting that our models were competitive.

Due to strong volatility and intermittent characteristics, the fluctuations of the photovoltaic (PV) output is inevitable which cause negative impacts on power system operation. However, such an issue can be resolved in a combined cascade hydro-PV-pumped storage (CH-PV-PS) generation system through an appropriate control. Based on the optimal base power of variable-speed pumped storage station (VSPSS), a smoothing method of fuzzy complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is proposed in this paper. To prevent smoothing insufficient and excessive, the fuzzy CEEMDAN algorithm is used to obtain the target power of the photovoltaic-pumped storage (PV-PS) generation system and the control signal of cascade hydro stations (CHSs) for eliminating large amplitude fluctuations of the PV output. Furthermore, in order to overcome the drawback that too frequent conversion of unit operating mode would reduce the service life and smoothing effect of the VSPSS unit, the optimization model of the base power of the VSPSS is established and solved by the grid adaptive direct search method (MADS) to obtain the control signal of the VSPSS for suppressing the short-term fluctuations of the PV output. Finally, the performance of the method is evaluated through using the PV station data in Xiaojin County, Sichuan province, China. The simulation results reveal that compared with the fixed base power smoothing method, this method can effectively suppress the PV power fluctuations and improve the smoothing effect and energy efficiency.