Abstract:Uu is defined as the interface between UE and base station, and which undertakes the establishment, reconfiguration, release of wireless carrier service, guarantees the interconnection and intercommunication of equipment from different manufacturers. LTEAdvanced uu monitoring instruments based on the most complex protocol provide detection and inspection method for uu wireless signals and thirdparty evaluation tool for conformance testing of UE and networks. The conceptual design scheme for LTEAdvanced uu monitoring and analyzing instrument is introduced in the paper. Furthermore, the key technologies including multiuser random access, uu protocol analysis, security algorithm, carrier aggregation and multiuser behavior sensing are adopted to realize the function of L1, L2, L3 protocol analysis for LTEAdvanced uu, cross level correlation analysis, 5carrier aggregation analysis, 48user service tracing, signal processing and user behavior monitoring. All testing process accord with standard protocol flow, and the results are full agreement with the actual behavior of UE and base station. Therefore, the testing means for fault diagnostics and locating of equipment interconnection and intercommunication are provided, and it makes great sense for TDLTEAdvanced network construction and optimization.

Abstract:In order to improve the prediction accuracy of nonlinear time series, a prediction model based on particle swarm optimization Bspline network is proposed. In designing the structure of the network, the nodes of Bspline basis functions which are considered to be independent variables and every correlative weight parameter are to be optimized together in the network training process. And the forecasting error square sum is adopted to evaluate the training effect of the network. A particle swarm optimization algorithm with an appropriate search strategy is used as the training algorithm to search the distribution of optimal nodes of Bspline basis functions and find the optimal weight parameters, so that the structure of the network is optimized. Then, the nonlinear time series is predicted by the network. The simulation results indicate that the prediction model based on particle swarm optimization Bspline network has a fine generalization performance, and the algorithm optimizes the network effectively. The proposed prediction model is not only simple in structure, but also has higher prediction accuracy.

Abstract:Aiming at the problem that the lack of theoretical instruction lied in the design of the new type electrostatic loudspeaker based on graphene membrane, theoretical modeling and simulation analysis are conducted on electrostatic graphene loudspeaker. Based on the working principle of an electrostatic loudspeaker, the vibration characteristic of a graphene membrane is calculated with the massspringdamper system model. The influence of the size, thickness and stress of the membrane to the vibration characteristic is analyzed. A finite element model of the electrostatic graphene loudspeaker is built with COMSOL software. The comparison is made among loudspeakers with different size, thickness and stress in the simulation analysis, which verifies the accuracy of the theoretical model. The results show that the graphene membrane with larger radius, thinner thickness and smaller strain will contribute to better frequency response characteristic of the electrostatic graphene loudspeaker.

Abstract:A human machine system for astronauts virtual training was designed and evaluated in this paper based on the real ground training tasks. This system can simulate kinds of manipulations, such as grasping, pulling and pushing. It can help to build a high level of immersion and improve the fidelity of the simulation training. This system is mainly composed of a 7 DOF haptic device, hardware systems and software systems. The haptic device, with a hybrid structure, has a 3 DOF translational mechanism, a 3 DOF rotational mechanism and a 1 DOF finger mechanism. The hardware systems includes the signalsobtaining of photoelectric encoders and reception of Force feedback parameters. Unity3Dbased virtual scenes have been built in the software systems. The software systems also connect the scenes and the haptic device. A control methodology based on MR brake was proposed in this paper. A MR brake was designed and used as actuator in this virtual training system. The whole system has a large workspace, reaching 0.4 m×0.4 m×0.5 m. It also has features of high position tracking accuracy and high output capability.It can also be widely used in rehabilitation training, virtual surgery and other fields.

Abstract:The main disadvantage of currentmethods ofdynamic heart rate measurement is the low accuracy. In order to improve the problem, deep learning algorithm was introduced to extract the photoplethysmograph(PPG) of heart rate value. In this paper, the pulse signals of 15 healthy subjects participated in the experiment was acquired under the different velocityas the input of stacked autoencoders network (SAE). At the same time, electrocardiograph(ECG) signal as the label of that network was gathered by a standard ECG collector whichhas high antiinterference. Combining with the deep learning algorithm, SAE was trained,in which the pulse signal with strong interference was fitted to thesignalof sinelike wave with the characteristic of accurate heart rate, in order to realize the extraction of heart rate under the condition of serious disturbance under sports conditions.The experimental results show that compared with the output signal of SAE, the proposed method obtains smaller error value of the heart rate (1.165 8 bpm), which showsthe effectiveness of heart rate measurementusing deep learning algorithm, and provides a new way for the sportiveheart ratemeasurement.

Abstract:In order to improve theresolution of depth imagemore effectively, a deeper convolution neural network is constructed in this paper. The network directly adapts the lowresolution depth image as the initial input of the network,and learns the highorder representation of depth image through the convolution neural network to obtain the features with more expressive ability.At the same time,the subpixel convolution layer is introduced at the output layer of the network. Based on the extracted features, a set of sampling filter is learned to achieve the amplification operation. For a better performance of the convergence, the residual network is added to our network. The experimentsare conducted on four commonly used datasets, and the results show that our network is faster than other advanced ones at the convergence rate. The proposed method can effectively protect the edge structure of the depth image,solve the artifact problem,and reachesgreat performance both in qualitative and quantitative aspects.

Abstract:In order to solve the defects such as inaccurate target location, target drift and recognition error induced by influence of illumination change, target rotation, occlusion in complex environment, a motion recognition algorithm based on double feature fusion and adaptive boosting was proposed. Firstly, in order to reduce the influence of illumination variation and occlusion on behavior, spatiotemporal context was used to extract the image sequences feature based on spatiotemporal context and the visual system characteristics. At the same time, the convolution neural network was introduced to operate the image sequence’s features for obtaining the STC and CNN features. Secondly, the principal component analysis operator was introduced to effectively combine the STC features and features to form a more accurate and complete feature representation. Then, by the new features, the adaptive boosting algorithm was used for classification training, the decision making of action was completed. The tests on the current popular data set show that, compared with the current commonly used behavior recognition methods, the proposed algorithm can recognize and understand all kinds of action, recognition rate is greatly improved, able to adapt for complex background and behavioral changes. This algorithm has higher accuracy and practical value in video surveillance and humancomputer interaction.

Abstract:To solve the problem of line scale identification affected by line scale wear, line scale scratch, camera imaging out of focus, weak or strong exposure and uneven illumination of the line ruler, an algorithm for line scale identification and central line eigenvalue extraction is proposed. Firstly, the location of each gray extremum of line scale is acquired by region division method.Then,the eigenvalue of each center line of line scale is determined by extremum weighted method. Finally, the gray value on each center line of image is firstly smoothed, and then the gray jump terminal coordinate value of each line scale is searched, and the central line eigenvalue of the main line scale is derived according to the coordinate value. The experimental results show that the algorithm for line scale identification and central line eigenvalue extraction can eliminate the influence of the problems above,and the eigenvalue extraction of each central line is correct.

Abstract:Monitoring hydrogen sulfide is very important for controlling the risk of leakage and protecting the environment safety. Tunable diode laser absorption spectroscopy, which takes the advantages of high sensitivity, speed and specificity, has a very broad application prospect in the field of gas online detection. In this paper, the hydrogen sulfide gas detection method is proposed based on tunable laser absorption spectroscopy and the calibration method to eliminate the effect of pressure and temperature. The measuring system applies both temperature and current tuning protocol to achieve larger tuning range, higher tuning speed and stability. Based on the lockin technology, the second harmonic are extraction from the absorption signal of hydrogen sulfide. By analyzing the influence of gas pressure and temperature on the absorption spectrum, the calibration method based on parameter fitting is proposed. The results show that the detection system achieves high sensitivity in hydrogen sulfide detection. The calibration method is effective in eliminating the influence of gas temperature and pressure.

Abstract:Using the tunable diode laser absorption spectroscopy (TDLAS) technique, the average temperature along the laser path can be measured using two absorption lines. When the average temperature along the laser path is measured by using multiple absorption lines, the measurement accuracy is mainly determined by the deviation of the measured signals and the lower energy of the absorption lines. The laser with a central wavelength of 1 392.5 nm (7 181.32 cm-1) can rapidly scan four absorption lines of H2O, in which the wave number is 7 182.950, 7 182.209, 7 181.156, and 7 179.752 cm-1 respectively, and investigate the influence of different combinations of spectral lines on precision of average temperature along laser path. The experimental results show that the weight of the accuracy to the measured temperature of every absorption line can be quickly reflected from calculating the standard deviation of the measured temperature. The standard deviation of the average temperature along the laser path is less than 20 K by absorption line selection.

Abstract:Timeinterleaved analogtodigital convert (TIADC) is used to improve the sampling rate of ADC system effectively. However, the timing mismatch, gain mismatch and offset mismatch are unavoidable in the TIADC system, which degrades the signal to noise ratio (SNR) of the TIADC. In this paper, an adaptive estimation method based on VSSLSM algorithm is proposed, which can estimate the timing mismatch and offset mismatch simultaneously. According to the estimated values, the timing and offset mismatches are corrected. The simulation results show that the proposed algorithm requires less than 128 samples to estimate the timing and offset mismatches of each subADC. The effectiveness of the proposed algorithm is verified in an actual TIADC system. The SNR of TIADC system is improved by 20 dB. The simulation and experiment results show that the proposed algorithm improves the performance of TIADC system effectively.

Abstract:After pretreatment of 80 samples of maize, interval partial least square (iPLS), combination of interval partial least squares (SiPLS) and successive projections algorithm (SPA) is respectively used to optimize the best wavelength of moisture components, and the correction model is established. The results show that iPLS, SiPLS and SPA method reduces the modeling variables from 700 to 70, 140 and 2, respectively, which occupies 10%, 20% and 0.29% of the whole spectrum. And, the modeling accuracy is even better than that of the 700 full spectral variables. The modeling accuracy of SiPLS and SPA is matched. But the SPA method reduces variables from 700 to 2. The complexity is minimized, and the precision of the model is kept, which show that the SPA method is an effective feature extraction method of wavelength. This research method can be extended to the application of fat, protein and starch components detection of corn.

Abstract:Aiming at the issue of analog circuit fault diagnosis and location, a novel approach for analog circuit fault diagnosis based on continuous wavelet Tsallis singularity entropy (TSE) and extreme learning machine (ELM) is proposed to enhance the accuracy of fault diagnosis. Firstly, the fault response signals are preprocessed by the continuous wavelet transformation to obtain the timefrequency coefficient matrix, and the matrix is divided into 8 congruent timefrequency blocks. Then, the feature vector is obtained by computing TSE of each block. Finally, the feature vectors are used as the inputs of a kind of multiclass classifier, namely ELM. The simulation results demonstrate that the proposed fault diagnosis approach can not only extract the essential features of fault response signals with better performance, and also achieve higher diagnosis accuracy than other reported approaches.

Abstract:In order to realize the online monitoring and accurate identification of picks wear degree in the cutting process, a new method based on multifeature information fusion was proposed for identifying shearer pick wear degree. The vibration and acoustic emission signals in the cutting process of different picks wear degree were analyzed by time domain analysis and wavelet packet analysis. According to the features of two adjacent pick wear degrees, there were data intersections for characteristic samples, which increased the difficulty of system identification. The optimal fuzzy membership function for each characteristic signal was calculated by using the least fuzzy optimization model, the maximum membership degree of the feature sample was obtained. The backpropagation(BP) neural network recognition model was trained and learned by using multifeature data samples. The experimental results show that the results of network discrimination are consistent with actual wear level of test sample. It can accurately monitor and identify the type of picks wear. The research results have great significance to monitoring and replacement of picks in actual engineering.

Abstract:The wear of aviation engine bearing parts is one of main factors that cause engine failure and serious accidents. The wear degree of bearing parts can be judged by monitoring the wear debris of the aero engine oil. In the past, the detection method has the problem of difficulty operation, high randomness and large quantitative error. This paper presents a method for the detection of aircraft engine wear debris based on electrical capacitance tomography (ECT) and establishes an experimental system and establishes a mathematical model of the relationship between the size or the situation of the abrasive particle and the change of capacitance value. Through the ANSYS and test experiment, the ECT method can detect the size of metal particles. The experiment system is equipped with 36 mm internal diameter sensors, which can detect the particles’ size of 3~8 mm.

Abstract:Spray production line trajectory planning and spray robot self programming technology are based on the workpiece online measurement. As a costeffective threedimensional imaging device, TOF camera has been applied to workpiece measurement in recent years. Aiming at the problem that existing TOF camera in terms of limited field of imaging view and single image can only obtain local contour depth information. A method of 3Dobject modeling based on GPUaccelerated computing and TOF point cloud streaming is proposed. The main algorithm is signed distance function (SDF) point cloud fusion, then spatial hashing storage is used to manage massive point cloud data, meanwhile, fast odometry from vision (FOVIS) system for pose estimation is introduced to improve the efficiency and robustness of insitu 3Dobject modeling algorithms of workpices. The experimental results on simulation platform of spraying production line show that the average number of frames in the modeling process can reach 58 frames per second, failure rate less than 2%, graphic memory usage rate about 25%, provides complete point cloud data for subsequent 3D measurement and spray traajectory planning.

Abstract:A twostage guided cosaliency detection model based on interimage saliency propagation and intraimage manifold ranking is proposed to fully exploit the saliency bootstrap propagation mechanism of single image, and improve the accuracy of the cosaliency detection algorithm. For any pair of images in a group image containing N images, the first intersaliency propagation stage utilizes the similarity between a pair of images to discover common properties of the images and get N-1 initial cosaliency maps with the help of a single image saliency map. In order to effectively suppress the background disturbance, the efficient manifold ranking（EMR）algorithm is used to calculate the ranking scores of each initial cosaliency maps in the second stage. The ranking scores are then directly assigned to all pixels as their new saliency values. Finally, an integration algorithm is proposed in the Bayesian framework to get the final cosaliency map. Based on iCoseg and MSRC image databases, the experimental results show that the proposed algorithm is superior to the five existing cosaliency detection algorithms uniformly in Fmeasure and the area under ROC curve (AUC) value. The algorithm is further validated by the real context experiment from the general practicability principle.

Abstract:Aiming at the drawbacks of traditional dictionary learning methods, such as single sample signal and poor reconstruction effect, a new approach of subdictionary concatenate learning（SDCL）with label information was proposed to identify the power quality disturbances (PQD) signal. Firstly, the different types of testing and training of the PQD signal samples are dimension reduced feature extraction with principal component analysis (PCA), add the label information to train samples, then the different categories of power quality samples are trained into redundant subdictionary and concatenated into structured dictionary. Finally, using dictionary learning algorithm to optimize the structured dictionary and the object class is determined through minimizing the redundant error. The simulation results show that the recognition effect of SDCL method is better than that of SVM and SRC, and has good antinoise robustness, and the multiple PQD identification rate reaches above 91.4% in the noisy circumstance with the signal to noise ratio above 20 dB.

Abstract:To address the issue that principle component analysis (PCA)shows a poor ability to acquire the characteristic of nonlinearity data, a kernel principle component analysis(KPCA) temperature point optimization method is proposed. Firstly, nonlinearity mapping function is introduced to map the input temperature data into the characteristic space,and a Gaussian radial basis is selected to be a kernel function. Secondly, inner product operation in characteristic space is transformedinto kernel function operation in input space, eigenvalues and kernel eigenvectors are found. Finally, a comprehensive independent variable is formed. According to an experiment conducted on a CNC machine center,and comparedwith the PCA model, RMSE and Maximum residual error reduces by 36% and 29%, respectively.KPCA can preferably acquire the characteristic of temperature data, and the prediction ability of KPCA model has an obvious improvement.

Abstract:Aiming atthe variety of electric larceny means,the efficiency of electric larceny detection remains improvement.Firstly, the fuzzy C mean clustering algorithm is used to construct different load characteristic curves of the user, and the suspiciouselectric larceny user is preliminarily determined by comparing the curves to be detected with the corresponding characteristic curve.Secondly,the particle swarm optimization support vector machine regression model is adopted to detect the behavior of suspected power stealing users.The experiments show that this method can reduce the range of electricity larcenydetection and overcome the influence of less electricity larcenysamples, improve the efficiency ofelectricity larcenydetection, and increasethe mean square error and average absolute error by 00051 and 0.034 respectively.

Abstract:When the gasliquid twophase flow occurs, the drive energy of Coriolis mass flow transmitter based on singleended drive mode is insufficient, which leads to very low of the vibration amplitude of the flow tube,very large of the measurement error of the flowmeter, and even abnormal work. Therefore, the drive capacity and its enhancing space of the transmitter based on singleended drive mode are analyzed.A differential drive mode is proposed to enhance the drive energy and meet the requirements of intrinsically safe. The circuit is developed and used in Coriolis mass flow transmitter. The comparative experiments on two kinds of driving modes are performed under the gasliquid twophase flow, and the water flow calibration experiments are conducted. The experimental results show that the vibration amplitude of the flow tube is almost doubled in the same flow rate and gas volume fraction when the transmitter is based on differential drive mode. The measurement accuracy can still reach 0.1% in the case of the singlephase flow, so the differential drive mode can maintain the vibration stability of the flow tube.

Abstract:Aiming at the structural data cleaning, an improved structural repairing algorithm based on regular expression was proposed according to calculate the edit distance between strings. Firstly, the violation partial order edge from edge set of nondeterministic finite automata was extracted, then the edit distance for edge in it was only revised by priority queue. At the same time, others edge to satisfy the partial order relation could calculate by recursive formula instead of the complex priority queue. The experimental results show that the improved algorithm not only has obvious advantage in time complexity, but also the improvement rate is significant and stable comparted with the original algorithm.

Abstract:The electric characteristic of microgrid’s battery has obvious nonlinearity and irregularity at work, it is difficult to accurately estimate the using the traditional mathematical methods. According to the problem above, the topology of back propagation (BP) neural network is constructed, and the network model is trained with new adaptive algorithm to improve the traditional learning model, the weights among neurons in the neural network model are adjusted reasonably, and the error is reduced with higher efficiency. The simulation result shows that the estimated results are within the scope of preset accuracy, the average error is less than 4%. It indicates that the BP neural network using the optimized algorithm can accurately estimate the state of charge, the attempt is effective and feasible.

Abstract:The design of energy saving fuzzy control system for expressway tunnel lighting is discussed in this paper. Based on fuzzy control theory, an energy saving system for expressway tunnel lighting is designed. Firstly, the illumination intensity of the outside of expressway tunnel and the flow of tunnel traffic are measured on line. Then an adaptive control strategy is presented to adjust the lights in tunnel inlet and tunnel outlet, which can avoid “tunnel black hole effect” and “tunnel white hole effect”. Furthermore, a simple control method is proposed to adjust the lights inside tunnel. Finally, the practical running indicates that the proposed system meets the tunnel lighting design specification, and works stable. It can reduce 22% lighting energy compared with that of the same time interval, and has popularization value.

Abstract:In order to solve some problems that the signal is undetected in the low signaltonoise ratio(SNR), and the accuracy is not high of parameter estimation, the singular value decomposition (SVD) filtering is proposed on the basis of generalized S transform (GST) for multicomponent chirp signal (MLFM). On the basis of S transform, the generalized Stransform and inverse transformation formula are derived in the paper. The singular value of the generalized Stransform matrix is obtained by discrete singular value, and the multicomponent signal Timefrequency filtering is realized by selecting the appropriate singular value. The simulation results show that the method can effectively filter out the noise in the low SNR, and avoids the phenomenon of missed detection when the amplitude of each component signal is quite difference, the accuracy of the signal parameter estimation is optimized.