Abstract:In order to warn the potential failure of wind turbines and enhance the safety of the unit output, a wind turbine fault warning method with abnormal data reconstruction is proposed based on the supervisory control and data acquisition (SCADA) system. Firstly, the SCADA data of the wind turbines in the same wind farm are fully utilized to reconstruct the two types of target unit data of the input and output, respectively, which overcomes the problems of partial data missing and data anomaly. Secondly, a fault warning model is established using the extracted representative data, which is closer to describe the dynamic characteristics of the unit. Thirdly, the improved deterioration degree is adopted to warn the potential failures and intuitively show the phased deterioration of the unit. In the case study, the SCADA fault data of a certain wind farm were used, and the parameter settings of the proposed strategy were determined with three criteria. The results show that the proposed method can predict the potential fault of the gearbox of the wind turbine at least 3 weeks ahead, which verifies the timeliness of the proposed early warning method.

Abstract:The operating conditions along longdistance oil and gas pipeline are completed, and the premise that the distribution of actual samples is consistent with that of standard samples in traditional method is destroyed. This situation results in the low identification accuracy of intrusion event for single identification model under different conditions. In order to improve the identification model deviation, this paper proposes a pipeline intrusion event identification method based on the deep transfer learning for domain invariant feature. The stacked sparse autoencoder network is utilized to adaptively extract the domaininvariant features for the intrusion events under different working conditions. Then, the transfer learning is introduced to achieve the accurate identification of pipeline intrusion events under complex conditions. The proposed method reduces the distribution difference between complex real scenes and typical scenes through scene difference evaluation, and obtains an effective domain invariant model. The experiment results show that the proposed method can obviously improve the recognition results of oil and gas pipeline intrusion events under complex conditions, and enhance the identification accuracy.

Abstract:Scientific and reasonable pipeline safety monitoring technology is of great significance to pipeline engineering operation. This paper carries out the experiment study on pipeline bending deformation monitoring based on distributed optical fiber sensing technology. Aiming at the deficiency of existing distributed fiber deformation calculation method, a calculation method of pipeline bending deformation monitoring using distributed optical fiber sensor is proposed, and the calculation program of pipeline bending deformation using distributed optical fiber sensor based on MATLAB is written. The study results show that the proposed pipeline bending deformation monitoring method based on distributed optical fiber sensing technology has high overall measurement accuracy. Within the bending deformation range of 180 mm, the absolute error is less than 4 mm and the average relative error is within 2%. When the bending deformation is getting larger, the absolute error increases, however the average relative error is below 32%. The pipeline force analysis based on distributed optical fiber sensing technology was carried out preliminarily. The results show that the simulated pipeline shear force pattern is in good agreement with the theoretical pattern and the actual situation. The proposed pipeline bending deformation monitoring method based on distributed optical fiber sensing technology possesses high measurement accuracy and small error, which can meet the requirements of pipeline bending deformation monitoring and has good application prospect. The method is an ideal deformation monitoring technology and can also be extended to the application of other safety analysis such as pipeline force analysis and etc.

Abstract:Aiming at the low accuracy of crack depth prediction in metal surface utilizing the empirical fitting method based on specific ultrasonic signal parameter and object characteristic, based on the divergence analysis and least parameter set principle, as well as training BP neural network, a surface acoustic wave (SAW) quantitative characterization technique of metal surface crack depth is proposed. This technique simulates the laser exciting SAW process with finite element method and extracts the characteristics of peak and mean values of the reflected and transmitted SAW signals caused by the surface crack, which is used to train BP neural network and predict the crack depth. The quantitative characterization of 20 groups of the opening cracks with crack depth of 01~20 mm on stainless steel specimen surface was realized. The simulation results show that the relative errors of the predicted crack depth are within 3%. Compared with the prediction result of empirical fitting curve, the accuracy is improved by more than 60%. Experiments adopted a 5 MHz SAW transducer to acquire 20 reflected SAW signals of two preprocessed cracks of the stainless steel specimens at surface depth of 10 and 15 mm, respectively. The relative errors of the crack depth predicted by BP neural network are within 01%, which verifies the feasibility and accuracy of the proposed quantitative characterization technique.

Abstract:To early discover the critical degradation status of rolling bearing and predict its degradation trend, Tdistributed stochastic neighbor embedding (TSNE) sample entropy state degradation feature index and the prediction method based on time convolutional network (TCN) are proposed. Firstly, the lowdimensional manifold features of original vibration signal are extracted by TSNE algorithm. Then, sample entropy of lowdimensional manifold features are calculated as the status degradation. Finally, the degradation trend of the bearing is predicted by the TCN algorithm based on features of historical status degradation. Compared with traditional feature index, experimental results show that the TSNE sample entropy feature index can detect the critical status of the bearing degradation significantly at least 50 minutes ahead of time and the prediction error of the TCN algorithm is only 045%. These results have high application values in engineering.

Abstract:Stochastic resonance is an important technique applied in weak signal detection. The timedelayed feedback ecological vegetation growth (EVG) system driven by weak periodic signals and additive white Gaussian noise is used as a model to carry out detailed stochastic resonance phenomenon analysis, and is applied to weak signal detection and bearing fault diagnosis in this paper. Firstly, the FokkerPlanck equation is used to deduce the equivalent potential function and further obtain the expression of the system signaltonoise ratio. Then the correlation curve diagrams are used to concretely analyze the influence of different system parameters on potential function and signaltonoise ratio. The study results show that adjusting the system parameters, signal amplitude and noise intensity all can induce the timedelayed EVG system to generate the stochastic resonance phenomenon. Finally, through adjusting the parameters and using the stochastic resonance method of timedelayed feedback EVG system, it is successfully detected that the target frequency of the weak signal is 001 Hz and its amplitude is 2 978; and the obvious peak values are detected at the fault characteristic frequencies of the inner and outer rings of the bearing.

Abstract:A large amount of labeled data is a necessary condition for model training of intelligent fault diagnosis methods. However, this condition is difficult to be met in some industrial application scenarios. It is difficult to collect enough labeled data, especially in the fault state. This limits the industrial application of the intelligent fault diagnosis method. To solve this problem, an intelligent fault diagnosis method for mechanical equipment based on feature knowledge transfer is proposed. The characteristic knowledge contained in the sufficient amount of labeled data collected by the experimental equipment or other related equipment is transferred to the intelligent model deployed by an industrial field device. The feature knowledge transfer of monitoring data between different mechanical devices is completed. In this way, the intelligent fault diagnosis of mechanical equipment under unlabeled data can be achieved. First, the proposed method constructs the onedimensional deep convolutional neural network to realize the depth mapping from the raw vibration signal to the mechanical equipment fault category. Then, the domain adaptation constraint is added to the deep convolutional neural network to realize the deep transfer adaptation of the feature knowledge between different mechanical equipment monitoring data. Finally, the health status of the mechanical equipment is identified by a fully connected neural network. To evaluate the effectiveness of the proposed method, the transfer fault diagnosis experiment is implemented by the monitoring data collected from the bearings of the two mechanical devices under different performance conditions. Experimental results show that the proposed method realizes the transfer and adaptation of the monitoring data feature knowledge between different devices. Compared with the traditional intelligent diagnosis method, the proposed method improves the recognition rate of migration fault diagnosis between two data sets more than 20%.

Abstract:Volatile organic compounds (VOC) in human exhaled breath reflects the health condition of body, which can be used for the assessment of disease. Diseaserelated exhaled biomarkers (e.g., acetone, nitric oxide, ammonia and ethanol) can be obtained by mass spectrometry. However, the source of most VOC is still unknown. Recently, the sensorbased exhaled breath analysis has made great progress. It is expected to achieve lowcost screening for early diagnosis and pathology of large population. The development of breath sensor has gained increasing interest. These studies are primarily concerned with specific sensors for the detection of individual compound. At present, the breath sensors include metal oxide sensors, carbon nanotube sensors, and colorimetric sensors. In addition, enose systems that consist of semiselective sensor array to simulate the mechanisms of smell sensing of human have been demonstrated in many medical applications. This article reviews and discusses the major mechanisms of different breath sensors and their recent applications in respiratory analysis.

Abstract:The hind wings of some twowinged insects, such as houseflies, utilize the stickshaped mechanosensory halteres, which detect Coriolis forces to achieve rapid course control during aerial maneuvers. In this paper, a kinematic model of stickshaped mechanosensory halteres navigation is proposed based on the structure of the bionic wing of the houseflies. Maneuvers navigation principle of this model is analyzed by MATLAB. The mechanosensory halteres of houseflies and the electromechanical coupling effect of resonant tunneling membrane structure (RTS) are combined. In this way, a novel bionic microstickshaped navigation sensor (BMSSNS) is designed. The processing technology, signal detection method and path solving method of BMSSNS are studied simultaneously. The structure of BMSSNS is simulated by ANSYS. Simulation results show that the motion path can be calculated by integrating the Coriolis and the driving force information of the stickshaped structure under the initial boundary conditions. The further path experiment results show that the BMSSNS can effectively detect the path and posture information (e.g., pitch, roll and yaw). Its horizontal and vertical path positioning accuracy can reach 15 mm and 05 mm in short time. The sensor enables precise path location in occasions where small spaces and high positioning accuracy are required.

Abstract:Regular high precision detection of the deformation and surface oxide film thickness of nuclear fuel assemblies has become an important measure to ensure safe operation of nuclear power plant (NPP). Aiming at the generally existed serious problems of poor passive adaption performance, insufficient contact and measurement flexibility, and urgent improvement of detection accuracy and efficiency of existing detection devices for nuclear fuel assembly, a passive adaptive flexible detection device for nuclear fuel assemblies that combines deformation and film thickness high precision detection functions together is developed through designing the adaptive aligning mechanism based on mutated hook joint, the active/passive flexible detection unit with dynamic feedback of contact force and the high precision detection mechanism based on serialparallel hybrid connection. On this basis, a measurement error compensation method based on parameter dynamic integration is proposed through deeply analyzing the deformation and film thickness detection mechanism of the device, based on the constructed serialparallel hybrid connection closedloop detection loop and integrating the probabilistic sensing error cooperative compensation strategy. The prototype experiment results show the developed device can adaptively align the nuclear fuel assembly under anisotropic random deformation, meet the requirements of flexible contact and flexible measurement in detection process. Combining the proposed measurement error compensation method, the developed device can realize the high precision detection of deformation and oxide film thickness of nuclear fuel assembly, effectively enhances the detection accuracy, efficiency and safety of nuclear fuel assemblies.

Abstract:In the process of robust Kalman filtering, in order to avoid the problem of gross error transfer due to the correlation among Global Navigation Satellite System (GNSS) observations, a partial gross error robust filtering algorithm for GNSS observations based on chisquare test is proposed. Firstly, the correlation among observations is analyzed based on the anomaly test of the observation model, and aiming at the problem of gross error misjudgment caused by the correlation among observations, a partial gross error robust filtering algorithm is proposed. According to the hypothesis testing theory, the overall test of the filtering model is constructed, which judges whether there exists an abnormality in the overall model based on chisquare test, and the overall flow framework of the partial gross error robust filtering algorithm for GNSS observations based on chisquare test is given. Finally, two sets of experiments are designed, and three methods are used for comparative analysis to verify the performance of the proposed algorithm. The experiment results show that the proposed algorithm greatly reduces the influence of correlation among observations, can accurately identify the location of gross errors, significantly reduces the false alarm rate of gross error detection and ensures the robustness of positioning.

Abstract:As the nondestructive testing (NDT) images have the characteristic of unbalanced grayscale distribution, fuzzy cmeans clustering algorithm cannot commonly used effectively separate the objects from background in NDT images. To solve this problem, an improved suppressed fuzzy cmeans (ISFCM) clustering algorithm is proposed to segment the NDT images. Firstly, the total membership degree of each cluster is incorporated into the objective function of the suppressed fuzzy cmeans (SFCM) clustering algorithm, which can equalize the contribution of object pixels and background pixels on the clustering results. Secondly, the iteration forms of the new membership degree and cluster center are deduced on the basis of newly built objective function. Thirdly, the convergence of the proposed ISFCM clustering algorithm is analyzed and the implementation steps are given. Lastly, the proposed ISFCM clustering algorithm was applied to carry out the segmentation experiment of the NDT images. The results demonstrate that the proposed ISFCM clustering algorithm can not only effectively segment the NDT images with unbalanced grayscale distribution, but also extend its application scope, and enhance the robustness and adaptability.

Abstract:The noncontact electromagnetic ultrasonic technology can be used to measure solidification shell thickness of slab in continuous casting. The casting technical parameters (e.g., rolling force, water sprays control, rolling location and rolling speed) can be adjusted in time to avoid central segregation and loose problems. The listed advantages have important application value for engineering. To improve the range resolution and signal noise ratio (SNR) of the electromagnetic acoustic transducer (EMAT) used in high temperature casting slab with coarse grains and surface oscillation marks, a finite element model for the testing process of a racetrack spiral coil EMAT with a chirp wave transmitting pulse is proposed. The influences of EMAT design parameters, pulse width and bandwidth of the chirp signal on the range resolution and SNR of the received ultrasonic waves after pulse compression are analyzed. Subsequently, simulation results are verified by experiments. Results show that the SNR of the received ultrasonic wave with chirp pulse compression can be improved by approximately 19 dB, and the pulse width is reduced more than 624%. Pulse width of chirp wave and magnet dimensions have significant influence on SNR. The range resolution of ultrasonic wave is affected by the bandwidth of chirp wave and magnet spacing and width. In addition, the conductor diameter of the racetrack spiral coil and its impedance matching parameters also affect the range resolution.

Abstract:Illumination plays an important role in the surface defect detection of large cabinet. The quality of cabinet surface image captured in uneven or low illumination condition is poor, which may lead to defect detection error. To solve this problem, an image enhancement method is proposed by combining cartoon texture decomposition and optimal hyperbolic tangent curve algorithm. Firstly, cartoon and texture maps are separated from cabinet images using an orientation filter. The image illumination model is also formulated based on the Gaussian scale space theory, and the uneven illumination is removed. Secondly, the hyperbolic tangent curve is used to enhance the lowillumination image by the weighted stretching. Finally, the performance of the proposed image enhancement method is evaluated using the contrast, brightness and grayscale variance product parameters. The method performance is also evaluated based on the comparison results of defect detection on the original captured image and the enhanced images. Experimental results show that the proposed method is suitable to enhance the cabinet image captured under the uneven and low illumination condition. The accuracy of defect detection on enhanced images is significantly improved. To be specific, the recall ratio is increased by 29% and the Fmeasure value is increased by 21%.

Abstract:The existing 3D imaging methods can hardly acquire both visual image and depth information of a scene using a single device and a single frame. They cannot have the merits of high efficiency, compact size, and low energy consumption at the same time. Therefore, this study proposes an innovative multispectral 3D imaging method. The method obtains depth from the defocused images, which are captured by an image capture system consisting of an optical imaging lens with longitudinal dispersion and a snapshot multispectral image sensor. Its basic principle is illustrated as follows. First, the inherent longitudinal chromatic aberration of an optical imaging system (especially the longitudinal dispersionenhanced one) makes the defocusing blurs vary with different spectral bands. Secondly, a snapshot narrowband multispectral image (SNBMSI) sensor is used to obtain multiple spectral images at each single imaging frame. Finally, the depth from defocusing algorithm is utilized to recover the 3D information from the edge gradients of multiple spectral images. Evaluation experiments are conducted using a chromatic dispersionenhanced optical imaging system and a SNBMSI camera to capture 450±10, 525±10 and 620±10 nm three spectral images to recover depth of objects that are within 5 m away from the camera. Experimental results suggest that error of depth recovery is no higher than 5 cm. The proposed multispectral 3D imaging method can realize depth estimation using a monocular and with a single frame image. The proposed method has capability of obtaining both visual and depth information without neither spatial registration nor precalibration of depth. The time of single frame 3D imaging time is about 0186 s. The size of the image capture system is 120 mm×77 mm×65 mm and the working power of the image capture system is about 10W. The advantages of the proposed method include high time efficiency, compact volume, and low energy consumption. Therefore, the proposed multispectral 3D imaging method can be widely used in unmanned driving and intelligent robots.

Abstract:The traditional piezoelectric ceramic actuator has the inherent contradiction of high precision and large working stroke. To solve this problem, a novel 2R1T parallel precision positioning platform is proposed. It can be driven by the external force directly. The high precision and large working stroke can be both realized by the piezoelectric linear motor. Firstly, the influence mechanism of the structure parameters on actuation stroke is analyzed. Based on the analysis result, the flexible cylindrical flexible hinge with large working stroke is developed. Its structure parameters are optimized by using the fuzzy optimization method. Secondly, the 2R1T parallel precision positioning platform is established with three parallel branches, which are consisted of the two translation pairs and the flexible cylindrical flexible hinge. The kinematics model of the parallel platform is discussed by the coordinate homogeneous transformation method. The forward and inverse position solutions of the 2R1T platform are also obtained. Finally, experiments are implemented to evaluate the performance of the parallel platform. Experimental results show that the 2R1T parallel platform has advantages of synchronization and repeatability of positioning. By using the step motion mode of the piezoelectric motor, the parallel platform can realize the precise positioning. Its translation resolution is 009 μm, and the rotation resolutions are 08 μrad, 09 μrad and 10 μrad, respectively. In the continues motion mode of the piezoelectric motor, the large working stroke of the translation and rotation are 120 mm, 618°, 674° and 658°, respectively. The key performance indicators of precision positioning and large working stroke of the parallel platform are both realized based on different motion mode of only one piezoelectric motor. All basic research contents have important theoretical value and experiment foundation for the further study of the dynamical performance, the realization of precise positioning and control law of the 2R1T parallel platform.

Abstract:Hydrogen maser and cesium atomic clock are the mainly precise frequency sources which can produce the international atomic time and national standard time. They have excellent features of shortterm and longterm stability. How to make full use of the shortterm stability of hydrogen maser and the longterm stability of cesium clock has become a key technology during the process of time generation. To improve the longterm and shortterm stability of the time scale, a method for generating the fusion time of hydrogen maser and cesium clock using VondarkCepek combined filtering is proposed. Firstly, AT1 method is used to generate a clock ensemble time scale for hydrogen masers and cesium clocks, respectively. Then, the key parameters of VondarkCepek combined filter are selected according to the least square principle. In further, the performance of the time scale of cesium clock ensemble is enhanced by the differential information of the time scale of hydrogen maser ensemble. In this way, the fused time scale of hydrogencesium is obtained. The calculation results show that the 1hour stability of the time scale is 336×10-15 and the 15day stability is 3×10-15. These results are better than the performance of single time scale of single cesium clock or single hydrogen maser in the same average time.

Abstract:An accurate prediction model plays a very important role in the quantitative spectrum analysis. Aiming at the problem of large model prediction error caused by information lacking and imbalanced information distribution in small sample set space, in this paper, based on traditional MDMTD (multidistribution mega trend diffusion) method, a HybridMega Trend Diffusion (HybridMTD) technique is proposed to construct virtual sample space, which further expends the training sample set and improves the information distribution of the sample set space, and then obviously reduces model prediction error. The spectrum data sets of total cholesterol and triglyceride in whole blood samples were utilized to carry out comparison and experiment verification. The experiment results show that the PLS prediction models established based on the reconstructed data set with virtual samples added can provide lower mean prediction mean square error MRmesp (mean of RMSEP). The values of MRmesp of total cholesterol and triglyceride are 041 and 045 mmol/L, respectively. Compared with traditional MDMTD method, the errors are reduced by 467% and 224%, respectively. The proposed HybridMTD method can construct an adequate number of highquality virtual samples; the prediction model corresponding to the sample set with the virtual samples filled obviously reduces the prediction error, and has superior prediction performance compared with the existing MTD method. The application of HybridMTD technique in blood spectrum analysis effectively enhances the evaluation quality of physiological indicators, speeds up screening speed for cardiovascular disease and reduces its risk.

Abstract:Muscle spasticity is a common disorder of motor function. The electromyogram (EMG) and mechanomyogram (MMG) dualmodel information synchronous analysis has significant importance for the quantitative assessment of muscle spasticity. Aiming at the difficulty of synchronously acquiring the EMG and MMG signals with high signaltonoise ratio (SNR), in this paper a synchronous analysis method of EMG and MMG dualmode information with antipower frequency interference and light accelerometer signal correction is proposed, and a wireless multichannel EMG and MMG signal synchronous acquisition system is designed. Compared with the commonly used commercial EMG and MMG synchronous acquisition system (Delsys system), the EMG signal to noise ratio performance of the selfdeveloped system is similar to that of the Delsys system (both about 20 dB), and the effective frequency band (0~20 Hz) energy of the MMG signal is obviously higher than that of Delsys. The clinic test of the selfdeveloped system was conducted. For the healthy subject, the EMG signal to noise ratio of active elbow flexion is about 20 dB. For the three patients (muscle spasticity levels are 1, 1+ and 2 respectively, according to the modified Ashworth scale), the normalized EMG indexes are 054±005, 059±004,062±001, respectively (mean±std). The flexor MMG RMS (root mean square, RMS) of the three patients are 269±104 m·s-2, 319±113 m·s-2 and 489±119 m·s-2, respectively. Using the selfdeveloped system, the EMG and MMG information can be effectively used to grade the level of spasticity. The test results demonstrate that the proposed method can be used for muscle spasticity assessment and limb movement function monitoring.

Abstract:Aiming at the problems of noise and direct current (DC) drift in strain signal measurement, a suitable DC drift cancelation method is proposed based on studying the DC drift characteristics of strain signal. The signal is segmented according to local extremum points, and then polynomial fitting is performed on each segment. The onedimensional bilateral filtering is adopted to perform noise reduction processing for the drift cancelled signal, and the optimal parameter selection method of the bilateral filtering is proposed, and the denoising performance of the bilateral filtering is evaluated with signaltonoise ratio (SNR) as the index. The effectiveness of the proposed method was verified with experiment, in which the experiment data was the strain signal acquired from a force sensor in the minimally invasive surgery robot. Experiment results show that the improved piecewise polynomial fitting method effectively remove the DC drift of the strain signal, and the bilateral filtering method not only is suitable for the dynamic filtering of the signal, but also ensures excellent filtering effect. With the proposed strain signal processing method the resolution of the force sensor is better than 2 g.

Abstract:In closerange photogrammetry, it is required that the adopted coded marks must have unique identification number and can be identified as well as located accurately in the image. In this paper, a kind of coded marks with Chinese character as encoding characteristic is designed, and a detection recognition method for the coded marks is proposed based on convolutional neural network. Firstly, a virtual camera method based on the camera imaging principle is used to automatically generate large amount of simulative images of the designed Chinese character coded marks, which are used as training samples. These samples are used to train the convolutional neural network that is used as the recognition network of Chinese character coded marks. The real captured Chinese character coded marks in the measurement images are detected with a series of cede mark sifting criteria, and the identification number is identified with the coded mark recognition network. Finally, the ceded mark centers are located with the center location algorithm. The experiment results show that the proposed recognition network has strong ability for recognizing the Chinese character coded marks, the recognition rate reaches 9767%. The proposed method is less affected by noise, projection angle, image contrast and brightness changes, and possesses strong robustness. The proposed method can accurately segment the Chinese character coded marks and the center location algorithm can accurately locate the mark centers.

Abstract:The actual underwater acoustic channel has the feature of clustered sparsity. To solve this problem, a hierarchical Bayesian model for underwater acoustic channel estimation is proposed, which is based on Markov chain Monte Carlo (MCMC) sampler in orthogonal frequency division multiplexing (OFDM) underwater acoustic communication. The prior distribution of the channel is formulated by the structural features of the clustered sparsity underwater acoustic channel. The posterior distribution of the channel model parameters is achieved by Bayesian inference and the likelihood function of the received signal. Finally, the MCMC sampler is utilized to sample the posterior conditional distribution of the channel model parameters. In this way, the maximum posterior estimation of the clustered sparsity underwater acoustic channel can be obtained. The performance of the proposed method is analyzed by simulation comparisons in terms of least squares, matching pursuit and stepwise orthogonal matching pursuit channel estimation methods under different received signaltonoise ratios. The lake test shows that the proposed method can achieve accurate OFDM underwater acoustic channel estimation, tracking and decoding without any channel prior information. The method realizes OFDM underwater acoustic communication with the communication distance of 600 m to 3 500 m. In addition, the transmission data rate can reach 608 kbps.

Abstract:Aiming at the problems of susceptibility to noise and low stability of fourrotor UAV with single attitude sensor, the gyroscope array is used to form multinode, antijamming and stable multiattitude system. A flight attitude measurement system of fourrotor UAV based on the new gyroscope array is proposed. The measurement array is composed of multiple micromechanical electronic systems with low accuracy, which improves the accuracy and stability of the system data. At the same time, a corresponding BP network data fusion algorithm based on neighborhood search is proposed, which solves the problem that the traditional BP neural network requires accurately giving the output value. The BP neural network model was used in the data fusion processing of the gyroscopearray. The experiment results show that the multigyroscope array system designed in this paper obviously improves the antinoise performance compared with the singlegyroscope system. Compared with the traditional linear weighted fusion algorithm, the proposed algorithm increases the support degree by 92% and reduces the residual error by 442%. The practicality experiment shows that the proposed method has practical significance in improving the flight stability of fourrotor UAV.

Abstract:The development of the traditional digital flicker meter is realized by converting the analog filter of each link into digital filter according to the flicker measurement principle provided by IEC. However, the effectiveness of the digital filter is limited by the sampling frequency and the transformation selection. To achieve accurate extraction of the voltage flicker envelope signal, the analytical mode decomposition is used to improve the filter link in the square detection method. A new KN mutual convolution window is proposed. The correction algorithm, which is based on the KN mutual convolution window and the threespectral line interpolation, is realized. The flicker parameter detection method based on improved square detection method and new KN mutual convolution window is proposed in this study. The flicker parameter detection platform based on virtual instrument is developed. Simulation experimental results show that the proposed algorithm can effectively realize the accurate detection of flicker parameters in the cases of single frequency modulation, multifrequency modulation, power grid frequency deviation, containing superposition harmonic interference and noise interference. This method is easy to realize and its accuracy is higher than the traditional detection methods.

Abstract:The driving control of the elevator permanent magnet traction machine needs to obtain the mechanical angle of the rotor in real time. How to accurately detect the rotor position during the starting and operation phases of the traction machine is a key issue to be solved. Aiming at this problem, this paper proposes a composite rotor position detection scheme combining absolute and incremental detection based on sincos composite encoder. Firstly, based on analyzing the principle of absolute and incremental rotor position detection, two rotor position detection algorithms are optimized to improve the rotor position detection accuracy. Then, combining with the rotor maximum vector control strategy, the motor starting and operation scheme is proposed under the rotor position composite detection scheme. Finally, the simulation and experiment results show that the proposed rotor position composite detection scheme can ensure the normal and stable operation of the motor.

Abstract:Aiming at the walking operation task of hexapod robot in complex terrain environment, a cooperative control strategy both considering the walking efficiency and stability margin of the robot is proposed. Based on the existing velocityposture cooperative teleoperation framework, the strategy incorporates the shared control strategy to improve the maneuverability and coordination ability of the control system. The method introduces dominance factor as the control weight, which makes the master operator have different control weights on the speed layer and posture layer operation subsystems of the robot. The dominance factor is obtained using the fuzzy reasoner that takes the stability margin as its input. The dominance factor is introduced into the master slave controllers, and the master robot guides the operator to iterate the control instructions through generating haptic force. The Vortex semiphysical simulation platform and the experiment platform built with operating handles are used to verify the proposed method and the method is compared with traditional cooperative control mode. The experiment results show that operating the hexapod robot with the proposed strategy can reasonably control the speed and posture of the robot while ensuring the stability margin.