Abstract:Due to the mutual interference of the ultrasonic scattering echoes in the imaging process, it may cause speckle noise in the formed medical ultrasonic images and it is difficult to distinguish from the human body structure, such as organ, tissue, etc, so that brining about complications to the later clinical diagnosis and image subsequent processing. In order to process the speckle noise in ultrasound images, a noise reduction model is proposed for adaptive averagefiltering ultrasound image based on support vector machine (SVM). The method uses the classification characteristics of SVM to distinguish the noise signal and the nonnoise signal in the ultrasound image, then combines the SVM classification result and the averagefiltering to denoise for the noise image. This operate can ensure the tissue area and detail characteristic of the medically noisy image are maximumly retain while the noise area is maximumly smooth. In the experimental part, the method used on the physical body membrane and human ultrasound liver image respectively. The results show that the proposed method can effectively suppress and reduce the speckle noise in the ultrasound image, and retain its edge features, and the signalnoise ratio of denoised image is increased. It can prove that the proposed method is useful for medical ultrasound image denoising.

Abstract:In view of the current many image matching algorithms that mainly rely on threshold adjustment to complete feature matching, resulting in low matching accuracy and poor robustness of the algorithm. In this paper, an image matching method using ratio consistency constraints is designed on the basis of matrix. The box filter is used to approximate the Gauss partial derivative function, and the determinant is obtained by convolution of the box filter and the image on the basis of the matrix. The image features are extracted by the determinant. Hu invariant moments are computed in the neighborhood of feature points and their eigenvectors are obtained. Based on the vectors of feature points, the region energy of the neighborhood of feature points is introduced to obtain the matching results. Using the distance relationship between matching points in space, the ratio consistency constraint method is established. The Euclidean metric ratio of matching points is used to search for wrong matching and optimize the result of feature matching. Through experimental analysis, it is found that the matching results of the proposed method have better matching accuracy and robustness than those of the current methods.

Abstract:In order to take into account the transparency and antigeometric transformation ability of the watermarking system, a robust image watermarking algorithm based on visual saliency and quantization index modulation is proposed in this paper. Firstly, the Ripplet transform is used to process the host image for getting the feature map. Then, the Gaussian probability density model is used to calculate the visual saliency mapping corresponding to the feature map, and divide it into a series of nonoverlapping subblocks for calculating the saliency mean of each subblock. The nondown sampling Contourlet transform is introduced to decompose the host image for outputing the corresponding lowpass subbands and bandpass directional subbands. Subsequently, the low pass subband is divided into smaller nonoverlapping subbands and the energy of each subband is calculated. The saliency mean and energy are jointed to calculate the quantization step corresponding to the embedded subblock, which treat it as a key. The singular value decomposition is used to process each subblock of low pass subband for obtaining the corresponding diagonal matrix, and the maximum singular value is found out. A watermarking embedding method is designed based on the mean of the maximum singular value corresponding to all subblocks, and according to the quantization step corresponding to each subblock, the watermarking data is hidden into the carrier to get the watermarking image. Finally, according to the received key, the watermarking extraction mechanism is defined to detect the watermarking data in the watermarking image. The experimental data show that this algorithm has higher transparency than the current blockbased watermarking technology, and under the conventional geometric content operation, it shows stronger robustness, and the restoration of watermarking distortion is the smallest.

Abstract:In order to solve the problem as ringing effect induced by neglect of the image contrast feature in many current remote sensing image fusion methods, this paper uses the standard deviation information of the image to measure the contrast characteristics of the image, and then realizes image fusion. Hue, saturation, value（HSV） color model is introduced to extract V factor of multispectral image. With the help of NSCT, the different coefficients of V factor and panchromatic image are calculated. Then, the saliency factor of the image is obtained by Fourier transform, and combined with the regional energy characteristics of the image to form the fusion rules of lowfrequency coefficients, so as to realize the fusion of lowfrequency information. By using the standard deviation information of the image to measure the contrast characteristics of the image, and combining it with the average gradient information of the image, the fusion rules of high frequency coefficients are formed to realize the fusion of high frequency information. Finally, it is reconstructed by inverse NSCT to update the V factor. The updated V factor, combined with H factor and S factor of MS image, is reconstructed by inverse HSV color model, and the result is concordant. The experiments show that compared with the current remote sensing image fusion technology, this algorithm has higher spectral correlation coefficient and information entropy.

Abstract:In order to overcome the shortcomings of many current image fusion algorithm, such as discontinuity and ringing, which are mainly achieved by taking large image coefficients and ignoring the correlation between images, a multifocus image fusion algorithm based on nonsubsampled shearlet transform and guidance rule is designed in this paper. Firstly, the nondown sampling Shearlet transform (NSST) is introduced to calculate the multifocus image and obtain the different coefficients of the image. Secondly, the image correlation is measured by using the regional energy, standard deviation and spatial frequency characteristics of the image, and the measurement results are used as guidance information for selecting fusion rules, and the lowfrequency coefficient fusion is completed by constructing guidance rules. When high frequency coefficients are fused, the brightness and edge information of the image are measured by means of the mean value feature of the image and the Laplacian energy feature, respectively, in order to achieve the fusion of high frequency coefficients. The experimental results show that, compared with the current fusion algorithm, the fusion image quality of this algorithm is better and has better fusion performance.

Abstract:In order to improve the preservation of edge details and reduce false edges caused by noise in edge detection algorithm, an edge detection scheme based on the theory of Intelligence and direction αmean was designed. Firstly, based on the theory of ChiChi, the image is transformed into intelligence image, and the intelligence image was represented by three authenticity T, uncertainty I and false F members, which improves the expression ability of uncertain information such as noise. Then, in order to remove the noise effectively and keep the edge, the direction mask of the pixel was calculated, and a directionmean operator was defined by the direction average function. Then anisotropic filtering was performed on the image using the generated directionmean algorithm. Finally, an iteration equation was defined to determine whether a pixel was an edge pixel by judging the threshold of gradient. Experiments show that the proposed method can detect edges effectively and accurately compared with current popular algorithms. It can eliminate the influence of noise at different noise levels, reduce the generation of false edges and discontinuous edges, and provide a good basis for future industrial automation and intellectualization.

Abstract:This paper takes networked control systems(NCSs) with networkinduced delay as the research object, combined with the eventtriggered mechanism, which studies the stability of NCSs under denial of service(DoS) attacks. In the process of DoS attacks, the sensor of NCSs cannot receive the measurement information in time, or the actuator cannot get the control information, then there will be unstable subsystems. In order to improve the stability of NCSs, firstly, NCSs is modeled as a closedloop switched system with both stable subsystems and unstable subsystems. The model can handle of networkinduced delay and DoS attacks uniformly. Then, based on the analysis method of the switched system, a sufficient condition is derived from the concerned NCSs to be exponentially stable. Furthermore, the time ratio of the DoS attack is also analyzed, the concerned NCSs is always guaranteed to be exponentially stable, as long as the occurring probability of DoS attacks is in the effective range. Finally, the experimental results show that NCSs can obtain the corresponding exponential stability for 0<δ<3; the stability of NCSs with eventtriggered mechanism under the DoS attack is given to show the effectiveness of the proposed result.

Abstract:Aiming at the establishment of a complete description of the adaptive motion curve of finger touching keys, a measurement method based on a Leap motion 3D sensor is proposed. According to capturing characteristics of the measurement method, firstly finger touch kinematics model is established before coordinate system is aligned step by step. Then the motion capture and the selection of motion feature points are carried out. The deviation angle is first introduced to transform the finger motion model from twodimensional space to threedimensional space. Finally, different fitting models are used to process the motion data of piano finger with 95% confidence bounds, and an adaptive function is established to describe the motion characteristics of finger playing. The adaptive function proves that the calibration of human hand fitting with different lengths can provide correct initial posture for the robot, which provides theoretical support for the adaptive identification system of finger touch motion curve in the teaching platform of piano playing education service robot.

Abstract:The main technologies used in ultrawideband(UWB) based positioning systems are TOA and TDOA. However, the low positioning speed of TOA algorithm and the unstable positioning accuracy of TDOA algorithm limit the development of UWB positioning system. In order to balance the shortcomings of the two algorithms in positioning speed and accuracy, this paper proposes a new TOF and TDOA joint positioning algorithm combination of TDOA and TOF(CTT). The algorithm only needs three times of UWB communication to measure all time difference information and a TOF distance value required by the TDOA algorithm, and can calculate the distance between all base stations and a single tag in the positioning area. The positioning speed of CTT is at least 50% higher than the TOA algorithm. DW1000 is used to compare the positioning accuracy of TOA, TDOA and CTT algorithms respectively, the results show that the average positioning error of CTT algorithm in twodimensional case is less than 20 cm and the standard deviation is less than 10 cm, close to the accuracy of TOA and solve the instability of TDOA algorithm.

Abstract:The strong coupling between the bearing and other internal components of the equipment leads to nonlinear relationship between vibration signal and equipment state. Moreover, the single signal feature is difficult to describe the state of the equipment comprehensively, while multifeatures contain more status information, the signal redundancy caused by highdimensional features easily declines the classification accuracy of the model. Therefore, a rolling bearings fault diagnosis method based on generalized morphological difference filter (GDIF) and autoencoder network (AN) is proposed. This method uses the GDIF to reduce the noise of vibration signals, and obtains the lowdimensional intrinsic manifold from the highdimensional features of the signal by the max likelihood estimate (MLE) and AN algorithm, which alleviates the dimension disasters of highdimensional features. Finally, the extreme learning machine (ELM) fault diagnosis model is established to identify the bearing fault types. The experiments show that the method can effectively suppress the noise; and the classification accuracy of the ELM Model can reach 98.04% after dimension reduction of features by AN.

Abstract:Transmission line faults caused by wind deviation account for a large proportion in power grid operation. In order to monitor the change of the yaw angle of the insulated suspension insulators in real time, the fiberoptic sensing technology is applied by studying the load model of the overhang insulator wind bias. The online detection system for the wind deflector of the suspended insulator was built, and the online monitoring of the skew angle of the 110 kV pole suspension insulator in the Zhaotong area of Yunnan Power Grid was realized. According to the monitoring data, the change of the yaw angle of the suspended insulator string is consistent with the trend of the theoretical calculation model, and the variation range is roughly between -4° and +4°. The results show that the online detection method of windsounding angle of optical fiber sensing is accurate and reliable in practical application, which provides a certain reference for the safe operation of transmission lines.

Abstract:Aiming at the lack of adaptability as the traditional shallow bearing fault diagnosis method relying on artificial feature extraction and diagnosis expertise, a fault diagnosis method based on deep convolutional neural network is proposed to recognize twodimensional shapes. Firstly, in order to fully display the fault characteristic information of rolling bearing, the twodimensional time spectrum of rolling bearing vibration time series is obtained by using the shorttime Fourier transform. Secondly, different fault features are extracted by convolutional neural network adaptively. Finally, Softmax classifier is used to output the diagnosis results to realize bearing fault diagnosis. The results show that the accuracy of the measured bearing fault diagnosis is up to 999%, proves that the proposed method has a good generalization performance and feasibility.

Abstract:In this paper, a set of gymnastics motion recognition system based on MEMS inertial sensor is designed to solve the problems of complex background, limited range of activities and personal privacy leakage. The system mainly collects acceleration and angular velocity data of 11 positions when the human body performs gymnastics by constructing a sensor network. Based on the preprocessed two types of data, the parameters such as mean, standard deviation, information entropy and mean square error are calculated as classification features. The support vector machine (SVM) classification model is established and the actions of six gymnastics movements are effectively identified. The experimental results show that the SVM algorithm has better recognition effect than the machine learning algorithms such as Knearest neighbor, naive bayes and decision tree. The average recognition rate can reach over 97%.

Abstract:Series arc fault caused by aging damage of line insulation layer and poor electrical contact seriously threatens the power safety of low voltage distribution system. It is difficult to detect and extinguish series arc fault for its characteristics of small current, high temperature and strong concealment. Because of above reasons, a method based on wavelet feature and deep learning is proposed for detecting series arc fault. Firstly, series arc fault experimental platform was built to collect the current signals under typical resistive load, inductive load and resistiveinductive load. Secondly, after transformed by wavelet transform, collected signals were decomposed to construct training sets and test sets. Finally, the arc fault was identified by the improved AlexNet model, and the test results were output. The experimental results show that the accuracy of this method for serial arc fault identification is almost 9558%, about 1058% higher than using AlexNet model.

Abstract:Aiming at the problems of biogeographybased optimization algorithm (BBO) in solving complex job shop scheduling problems (JSP), an improved differential evolution biogeographybased optimization algorithm is proposed. By effectively combining the searchability of differential evolution algorithm (DE) with the utilization of biogeographybased optimization algorithm, at the same time, elite retention mechanism is adopted to retain individuals with higher fitness, and inertial weight strategy is introduced to adjust the proportion of mutation operation in hybrid migration operation to improve the global search ability of the algorithm, then increase the disturbance in the small probability in order to prevent the algorithm as the iteration progressed into a local optimal solution. Finally, different test functions and job shop scheduling problems are used for experiments. The results show that the improved algorithm has better performance in convergence speed and optimization results.

Abstract:Aiming at the problem of humidity sensor data distortion caused by temperature changes in utility tunnel, an improved Levy flight particle swarm optimization algorithm (ILPSO) is proposed to compensate data errors. Firstly, a neural network with prediction error is established, and the initial parameters of the network are found by PSO. Then, an improved levy flight is added in the PSO search process, the probability of particle flight is inversely proportional to the distance to the optimal particle, and when approaching the optimal particle, it reversely escapes from the optimal particle with a larger probability, thus overcoming the problem of particle premature. Finally, the network retrains with the output of PSO as the initial parameter. In the sensor error test experiment and stability experiment, the compensation effect of ILPSO algorithm is the best, the humidity value error after compensation is less than 5%, MSE is the lowest, and the stability is the best. In summary, compared with the traditional Levy flight, ILPSO algorithm has stronger adaptability to error prediction network, faster convergence, and improves the accuracy and stability of humidity sensor temperature compensation.

Abstract:After distribution network failure, a large number of alarm information will be generated. These information can be used to identify the faulty components quickly, and hence providing important decision support for dispatching center staff. Since the existing fault diagnosis methods of Petri nets are not applied in the distribution network, an improved hierarchical fuzzy Petri nets fault diagnosis method based on temporal constraints is proposed. An improved hierarchical fuzzy Petri net model for suspected fault component in distribution network is established, which can adapt to the change of network topology. The obtained alarm information is used to check the sequence of protection and circuit breakers by reverse and forward temporal reasoning, and the confidence of the transitions that does not meet the time constraints is corrected. The reasoning process and the matrix reasoning algorithm of the improved hierarchical fuzzy Petri net model are given. In the process of matrix reasoning, the modified probability of Gauss function is introduced to keep the probability at 0~1. Finally, the confidence probability of the fault components and its temporal point constraint are obtained. Through the comparison and analysis of the distribution network system examples, the correctness and rationality of the proposed method are verified, and the fault components of the distribution network can be effectively diagnosed.

Abstract:Anti-interference performance test is an important part of evaluating the performance of satellite navigation receiver. This paper focuses on two key indicators, interference rejection and interference recovery time, have been studied. Firstly, the receiver anti-interference test system based on navigation signal simulator is constructed. Reasonably determine the constraints of the receiver, by adjusting the interference signal to observe the positioning performance of the receiver to reach the constraint condition, to test the interference rejection of the receiver. By controlling the output time of the interference signal, to test the receiver’s delay time of restore normal positioning after interference signals, Then the interference recovery time of the receiver is measured. At last, two different types of receivers are tested in this paper. The performance of both receivers is degraded by the threat of interference signals, the test results of interference rejection and interference recovery time matched with their respective overall performance levels. It indicates that the test method in this paper can correctly evaluate the anti-interference performance of the receiver.

Abstract:Aiming at the problems of complex calculation and poor realtime performance of twodimensional MUSIC algorithm when searching for the spectral peaks, a twodimensional MUSIC spectral peak searching algorithm is proposed based on improved chicken swarm optimization (ICSO). This algorithm applies the ICSO to the spectral peak searching part. Firstly, the initial population is constructed by the theory of good point set. Secondly, the inertia weight function is composed of the feeding speed factor and the aggregation degree factor. Finally, the inertia weight is introduced into the position updating formula of the hen to make the algorithm quickly search the angle corresponding to the spectral peak. The results show that the proposed algorithm achieves the same searching accuracy as grid search method with lower time complexity. The time complexity is reduced by 648 times and the searching time is saved by 994%. Compared with the other three optimization algorithms, it has better convergence performance and higher searching precision.

Abstract:In devicetodevice(D2D) communication, when the distance between source node and destination node is too large, relay node can be introduced to improve the communication quality. When the communication decline is serious and single relay cannot improve the communication quality effectively, multirelay communication needs to be introduced. Aiming at multirelay communication in D2D communication, this paper proposes a multirelay selective communication mechanism based on Qlearning in machine learning. First, determine whether cooperative relay is needed for communication between source node and destination node. Secondly, the return value of Q function in qlearning algorithm is defined by considering the communication energy consumption in D2D network. Finally, the satisfaction function is obtained by calculating the transmission distance of D2D communication and the signaltonoise ratio of the communication receiver. Considering the return value and satisfaction, a cooperative relay set is obtained. Simulation results show that multirelay cooperation based on Qlearning algorithm can significantly reduce transmission delay and balance network load.

Abstract:In order to obtain the analysis signals reflecting the characteristics of the system, find the signal reconstruction method and autoregressive model that suitable for time series of the system. VMD filtering reconstruction method using harmonic coefficients to search the optimal penalty factor and decompose modulus is proposed to filter and reconstruct analysis signals of MR damper system, then ARMA and AR models of the analysis signals of the same order as the dynamic model of the system are established, compared with the reconstruction algorithm based on FFT and EMD signal filtering, the simulation accuracy of these models are analyzed. The results show that, among the three filtering reconstruction methods, the fitting accuracy of the simplified loworder model is lower than that of the nonsimplified highorder model, the simulation accuracy of the sameorder ARMA model is higher than that of the AR model, VMD filtering reconstruction method using harmonic coefficients to search the optimal penalty factor and decompose modulus has the highest simulation accuracy of the autoregressive model. Among them, ARMA (4,1) model based on VMD reconstructed signal has the highest modeling accuracy and is most suitable for system modeling and analysis.

Abstract:With regard to the problem that the traditional methods are difficult to precisely detect the demagnetization of permanent magnet synchronous motor (PMSM), an adaptive nonsingular terminalslidingmode control algorithm is discussed for PMSM. Firstly, the mathematical model of PMSM is established according to permanent magnet demagnetization condition. Then, adaptive observer and nonsingular terminalslidingmode observer (NTSMO) is constructed for permanent magnet demagnetization detection and the adaptive estimation of stator resistance is given. The stability of the sliding mode observer is proved by the Lyapunov stability theory. Based on the equivalent control principle of sliding mode variable structure, equation for permanent magnet flux linkage is constructed. Finally, it is verified by simulation experiments that after changing the stator resistance parameter, the adaptive highorder sliding mode permanent flux observer can detect the flux parameters accurately.

Abstract:Aiming at the current situation of realtime data stream transmission of different data bandwidths and fixed transmission speed of FPGA highspeed transceivers, this paper proposes a broadband adaptive serial data realtime transmission system. The sender first performs bitwidth conversion and crossclock domain processing on the transmitted data, then marks and encodes the data, and finally transmits data through the GTX transmission port. The receiving end first parses the valid data according to the control code and the flag bit, and finally performs bit width conversion and cross clock domain processing through the receiving buffer module. The system is applied to the acquisition data transmission of the oscilloscope. When the GTX line speed is 45 Gbps, the data of the 15 types of acquisition cards can be correctly received. The transmission system only needs to modify the write data bit width of the buffer FIFO at the transmitting end for different acquisition cards, which has the characteristics of strong code portability and strong hardware platform compatibility.

Abstract:While electric vehicles (EV) can improve driving range through wireless power transfer (WPT), changes in vehicle axle load or suspension vibration caused by road conditions can change the coupling coefficient between transmitter coil and receiver coil, leading to WPT system being far away from the maximum efficiency transmission point. To address the problem that WPT system efficiency deviates from the extreme point along with the fluctuation of coupling coefficient, a realtime estimation method of coupling coefficient for seriesseries (SS) topology is proposed, on this basis, a novel maximum efficiency tracking control scheme is proposed, which is regulating the DC link current of secondary side by PI feedback controller to maintain the highest efficiency of WPT system when coupling coefficient varies. Simulation and experimental results indicate the effectiveness of the proposed method.

Abstract:Aiming at the problem that the underground gas pipeline leakage positioning accuracy is not high, according to the characteristic of infrasound waves generated with leakage points when gas pipeline leaks, an infrasound source location algorithm of gas pipeline leakage is proposed. The algorithm builds positioning model, and the position of the leak point is obtained by the geometric relationship between the sensor nodes and the leak point. Firstly, the infrasound wave signals generated by the gas pipeline leaks are collected through a sensor array, and the signals are processed by wavelet denoising. Then the time delay between adjacent sensor nodes is calculated by the generalized crosscorrelation method, and finally the location of the leak point is calculated by combining the geometric model of the leak location. The signal preprocessing part of the algorithm is simulated, and the simulation experiment of leak location is carried out in the laboratory. The results show that the positioning error of the algorithm is within 1 m in the gas pipeline leakage location, and it has high positioning accuracy.

Abstract:The performance of the DC/AC inverter control strategy has a significant influence on the voltage level of the islanded microgrid system. The traditional PI control in voltage loop cannot meet the system performance requirement under load disturbance or parameter perturbation. Aiming at this problem, a voltage H∞ robust control method for the DC/AC inverter in islanded microgrid is proposed. A centralized secondary voltage regulation is used to restore the voltage deviation caused by conventional droop control. By considering the load current as the external disturbance input, the simplified mathematical state space model of the island microgrid is established. Then, a voltage H∞ robust controller is designed by selected weighting functionsbased on the mixed sensitivity optimization problem, which can improve the ability of DC/AC inverter so that its robust performance can be optimized under the system parameter perturbation. Finally, the numerical simulation result shows that the proposed strategy is feasible and the H∞ robust controller is of the stronger stability under disturbance.