Abstract:The objective of reducing 1 / f noise and achieving vectorized measurements by applying a modulation field is frequently used in atomic magnetometers. According to the relative magnitude of the modulation field frequency and relaxation rate, modulation atomic magnetometer (AM) can be divided into two categories, which are low frequency modulation AM and high frequency modulation AM. In this article, we summarize and classify high frequency modulation atomic magnetometers into two categories. The first is parallel modulation and the other is perpendicular modulation according to the relative position between the direction of the modulation field and that of the pump beam. Furthermore, seven different configurations are listed according to the main field direction, among which four configurations are analyzed in detail, namely perpendicular modulation X mode, perpendicular modulation zero-field mode, parallel modulation Z mode and parallel modulation zero-field mode. The measurement models for these four configurations are derived from the Bloch equation. In addition to combing and validating the theoretical analysis of the existing literature, we also obtain some measurement models that are not mentioned in the existing literatures. Numerical simulation is implemented by MATLAB Simulink module, which evaluates the rationality of some simplifications during analytical model derivation. Finally, a general Simulink model of Bloch equation is given, which can be used for numerical simulation of atomic magnetometers with different configurations.

Abstract:The single-channel absolute shaft encoder has advantages of high resolution, simple structure and strong reliability. To achieve high-precision and rapid angle identification and subdivision, this article proposes a phase-fitting-based absolute encoder angle subdivision method, which uses m-sequences to perform single-channel absolute encoding. First, the CCD sampling level signal is counted, and the combination of code values is judged to obtain the coarse code decoding data. Then, the Newton iteration method is used to implement trigonometric function fitting to obtain phase information, and a phase-based angle subdivision algorithm is proposed to obtain the subdivision angle. Finally, the angle information is obtained by combining the coarse code data and the subdivision angle. Experimental results show that the angle measurement standard deviation of the new proposed angle measurement method reaches 4. 57″, and the minimum graduation error is only 0. 23″. This method greatly improves the resolution and accuracy, and avoids the influence of the coarse error of the code disc on the angle measurement in principle.

Abstract:To meet the urgent needs of high-precision measurement of centering error at home and abroad, a laser differential confocal centering error of the lenses measurement system is developed. The system starts from the core fixed focus principle of center-offset noncontact measurement, and combines the laser differential confocal fixed focus technology to solve the problem of poor resolution of the legal focus accuracy. On the basis of the error analysis of the optical measurement system, the system is optimized again and the position precisely corresponds to the cat′s eye and the confocal point of the lens under test, which realizes the high-precision measurement of the centering error of lenses. Compared with the traditional sharpness legal focus measurement, experiments show that the measurement accuracy of the system is 0. 49% , and the measurement accuracy of the centering error of lens is effectively improved by 6 times. The system effectively applies the differential confocal fixed focus technology to the centering error of lenses measurement, improves the fixation ability of the catenary and confocal position of the measured mirror, and realizes the design of a high precision measurement system, which has a broad application prospect in the field of optical testing and lens processing and assembly.

Abstract:The signal demodulation of fiber optic hydrophone based on the 3×3 coupler method is generally affected by the asymmetric characteristic of coupler, which leads to inaccurate demodulation results and the measuring signals. To solve the problem of demodulation errors caused by both the amplitude and phase difference, which are commonly existed for a 3×3 coupler, a method of amplitude and phase demodulation correction based on differential cross-multiplying is proposed. Through the least squares ellipse fitting pairwise of three asymmetric interference signals, the amplitude and phase parameters of the interference signals are obtained, and the amplitude asymmetry is eliminated by DC removal and normalization. Further, the differential cross-multiplying operation is performed pairwise for the obtained three signals, which contain asymmetric phase. Then, the phase and the related correction coefficient of the signals are calculated by trigonometric transformation. In this way, the accurate demodulation results of the fiber hydrophone are achieved. Comparison of the proposed method with the existed method which only considers the asymmetry of amplitude is conducted. The demodulation errors of two methods varying with frequency and amplitude of the input acoustic signal are analyzed by numerical simulations. Results show that the proposed method could obtain the demodulated signal which is closer to the nominal signal waveform and has smaller demodulation errors. Moreover, demodulation experiments of fiber optic hydrophone are also implemented in the frequency range of 5 ~ 30 kHz in an anechoic water tank, which illustrate the effectiveness and stability of the presented correction method for demodulation.

Abstract:In the field of three-dimensional contour measurement, the complementary gray code phase shift method can obtain continuous unwrapped phase. However, the measurement accuracy of a single set of phase shift fringes is easily affected by the changes in the surrounding environment, and the traditional projection method increases the introduction of interference. To solve this problem, this article proposes a complementary gray code double-N-step phase shift profilometry based on the color-coded grating projection. First, the double-N-step phase shift fringe and the complementary gray code fringe are mixed and coded into color fringe, and then projected onto the target surface in turn. Finally, the phase information is extracted from the collected color fringe, and the phase is calculated and fused to obtain the three-dimensional surface topography. To evaluate the proposed method, a comparative experiment is carried out among this method, the traditional complementary Gray code N-step phase shift method and the three-frequency heterodyne method. Experimental results show that this method can obtain high-precision surface topography information, effectively reduce the measurement error, and improve the detection efficiency by 50% .

Abstract:Incremental optical encoders have grating line errors and subdivision errors inevitably during the manufacturing and installation. These errors reduce the accuracy of the angle measurement of encoders and cause instantaneous angular speed (IAS) signal fluctuations. Research on the compensation method for grating line error and subdivision error has great significance. However, the existing methods have limitations, such as low efficiency of error compensation and difficulty in field deployment. To solve the above problems, this article first analyzes the grating line error and the subdivision error of the incremental optical encoder and establishes the error model to reveal the relationship among the grating line error, the subdivision error and the IAS signal fluctuation. Consequently, a method to compensate for the grating line error and the subdivision error of incremental optical encoders using the IAS signal is proposed. The method has advantages of high efficiency and no need to modify the encoder. The correctness of the error model established in this article and the feasibility of the error estimation method are evaluated by simulation analysis. The grating line error and the subdivision error of the incremental optical encoder at the end of the servo motor are compensated on the RV transmission test rig. Finally, the error of the incremental optical encoder is measured by the optical rotating platform. The effectiveness of the proposed method is evaluated by comparative analysis.

Abstract:Concave grating has both imaging and dispersion capabilities, monochromators using concave gratings are well miniaturised and can be applied to the vacuum UV band below 200 nm. The spectral resolution and wavelength repeatability are important indicators of monochromators, for the spectral resolution of monochromators, this paper combines the intrinsic resolution of the grating and the broadening caused by the slit to derive a model for calculating the spectral resolution of monochromators. Experimental verification using our own micromotion slit, the monochromator resolution is in accordance with the theoretical model, with a limit resolution better than 0. 1 nm. For the wavelength repeatability of monochromator, the wavelength repeatability influencing factors are analyzed on the basis of the conversion of monochromator optical machine structure parameters, the wavelength repeatability calculation model of monochromator is derived, and the wavelength repeatability using mercury lamp as the light source is verified to be better than 0. 02 nm in accordance with the theoretical calculation model, which verify the validity of the structural design of the optical machine and the correctness of the theoretical analysis.

Abstract:Temperature-sensitive points significantly affect the performance of the thermal error model of machine tools. To solve the problem that many measurement points need to be arranged and key temperature difference information may be lost in the existing temperature-sensitive point selection methods, this article proposes a temperature-sensitive point selection method based on active construction of temperature difference variables. The temperature difference variables are obtained from constructing a limited number of the original temperature measurement points, and added as an extension of the original temperature variables. The temperature-sensitive points are selected, which are based on fuzzy clustering and correlation coefficient analysis, and used for the thermal error modeling. This method can make up for the lack of potential key temperature information in existing methods, which has higher accuracy and stability. Compared with the traditional temperature-sensitive point selection methods, experimental results show that the proposed method can reduce the average root mean square error from 11. 1、 10. 3 μm to 3. 6 μm, which shows the remarkable effect.

Abstract:Real-time monitoring of the rotor vibration signal is a key to ensure rotating machinery running steadily. The coupling source error of rotor roundness error and eddy current displacement sensor error is rarely considered in previous research and vibration monitoring, which causes distortion of the rotor vibration signal, and even causes misjudgment. Taking an actual rotor as an example, a measurement expression of the roundness error is proposed. The errors of two common eddy current displacement sensors are measured and analyzed. The expression of source error which is coupled by roundness error and eddy current displacement sensor error is constructed by Fourier fitting. The mapping relationship between coupling source error and rotor vibration signal is established. Three rotor vibration signal recognition methods are proposed, including the point-point method, the average value method and the max value method. All three methods can effectively recognize the rotor vibration signal. The point-point method is the most accurate, and its recognize error accounts for about 20% . The average value method is simple to calculate, and its recognize error accounts for about 10% . The maximum value method is conservative. However, it helps to effectively avoid misjudgment and its recognize error accounts for about 32% .

Abstract:An unbalanced M-Z interference system based on the 3×3 coupler is established. Combining the 3×3 coupler demodulation with the phase carrier, the output phase of the M-Z interferometer is dynamically demodulated by introducing carrier to adjust the output phase. The carrier wave is a cosine signal with a frequency of 2 kHz and a voltage of 0, 2, 6, 10 and 16 V. The output power curve of the interferometer with time is measured by the Throlabs optical energy meter. For the unequal splitting ratio of 3×3 coupler with 1 ∶ 1 ∶ 1, the phase difference is not 120°, the output interference light is decomposed by empirical mode, and a band-pass filter is constructed for denoising. The processed signal is fitted with Lissajous curve by the least square method to obtain the specific parameters of ellipse. The arctangent function is used to solve the phase difference. Finally, the voltage phase shift coefficient and the voltage wavelength shift coefficient are 0. 757 7 (°/ V) and 0. 725 8×10 -3 ( nm/ V), respectively. The first-order linear curve is fitted in MATLAB, and the goodness of fit R 2 is more than 0. 98. Finally, the demodulation system is applied to the stress load test of fiber Bragg grating (FBG) strings and compared with the micron optics (MOI) demodulation. The results show that the accuracy of this system is high, and the FBG demodulation effect at 1 560 nm is better than MOI.

Abstract:The Lagrangian interpolation filter can be applied to calibrate TIADC time error. This method has good calibration effect. However, the order of the Lagrange interpolation filter is not easy to adjust, which makes the filter not flexible in practical application. To address the problem, this article proposes a TIADC time error calibration method based on the Aitkin stepwise interpolation filter. To apply Aitkin stepwise interpolation algorithm to time error calibration, this article simplifies the calculation formula of Aitkin interpolation polynomial and designs a method for picking new interpolation nodes. The method can automatically adjust the order of the Aitkin stepwise interpolation filter according to the input signal frequency of the TIADC sampling system and the set precision requirement. Simulation results show that the calibration method can efficiently suppress the spurious components induced by the time error within 90% of the Nyquist frequency band, and can automatically adjust the filter order. When the normalized frequency of the input signal is 0. 35, the calibrated SNDR is improved by 40. 37 dB. The effectiveness of this calibration method is verified through experiments.

Abstract:The test of analog-to-digital converter (ADC) mainly includes two test processes of static parameters and dynamic parameters. With the improvement in the performance, the testing complexity and cost of ADC increases dramatically. Alternative test, which means obtaining two types of parameters from only one test process by analyzing the relationship between static and dynamic parameters, has been proven to be a major solution to reducing the complexity and cost of ADC test. In this article, the alternative testing is achieved by constructing a regression model based on artificial neural network. The model takes total Harmonic distortion (THD) as the prediction target, and takes the static performance parameters as the input features. For high-dimensional ADC nonlinear curves, statistical analysis and principal component analysis are combined to design a special feature extraction method, which greatly reduces the feature dimension and the loss of information. The prediction results on the test set show that the mean absolute error and R-squared between the predicted THD and the reference value reach 1. 15 dB and 0. 6, respectively, which are significantly better than those of other comparison models. In addition, SHAP (shapley additive explanations) model interpreter is used to analyze the dependencies between the prediction target and feature variables of the model, and meaningful results are obtained.

Abstract:Digital modulation technology is one of the key technologies of the high-speed communication transmission system. In this article, an architecture of high-speed digital modulation signal generation in parallel is proposed, which can be implemented by algorithm-level pipeline in field programmable gate array (FPGA) hardware platform. With theoretical analysis and derivation, DFT/ IDFT in parallel frequency domain forming filter can be cascaded by two levels of the low-complexity based-8 FFT algorithm, and the specific FPGA architecture and implementation method are given. In addition, a mixing-free digital orthogonal up-conversion architecture suitable for parallel implementation is analyzed and designed to further reduce hardware resource. The simulation experiments evaluate the algorithm of the high-speed parallel digital modulation architecture, and the FPGA hardware implementation results test the spectrum performance of the high-speed parallel digital modulation signal.

Abstract:With the increase of working time, the health status of analog circuit declines. The faulty parameter estimation in early fault state can accurately evaluate equipment health state and provide reference for fault prediction. Based on the transfer function of the circuit and the measured fault response, the possible fault parameters can be inversely derived. Due to the influence of tolerance, the same fault response can be generated by many parameter combinations. This article transforms the fault parameter estimation problem into a multi-objective optimization problem through mathematical analysis. In view of the problems of the huge difference in the optimization objective scale and the difficulty in generating a reasonable weight vector, it proposes to guide the population evolution based on the logarithmic distribution reference point and proposes a logarithmic distribution reference point-based decomposition multi-objective evolution algorithm. This method can accurately and stably find the optimal solution of the fault parameter estimation problem. Through simulation of the jump filter circuit, it is verified that as the tolerance increases, the parameter range becomes wider, and the maximum standard deviation is only 18. 616 Ω. In terms of time efficiency, there is no significant difference in the running time of 12 fault instances with three different tolerances, and the average running time is 0. 7 s. The correctness and robustness of the algorithm are verified by experiments, and compared with other multi-objective evolution algorithms in the forefront of this direction, the method proposed in this article is more accurate than the bi-objective evolution algorithm by 2~ 3 orders of magnitude in terms of accuracy, and has a wider fault interval than the method proposed by Tadeusiewicz, which verifies the higher accuracy and effectiveness of the method proposed in this article, reflecting the superiority and reliability of the method proposed in this article.

Abstract:The chip on PCBA is developing towards small size and high density, which make it much difficult to detect micro solder bump defects inside the package. To address the problems of difficulty and low efficiency in locating internal faults of ICs on industrial high-density integrated PCBA, a chip-on-board defect detection method combining the infrared thermal imaging and the deep learning algorithm is proposed, which realizes intelligent defect detection of ICs on PCBA suitable for industrial production scenarios. Taking the real DDR memory chip on FPGA as the target, the infrared defect detection model is formulated, and the test bench is established to conduct experimental research on the fault detection of solder bumps in the chip. The designed program realizes the chip data storage and readout. The infrared image sequence is collected to analyze the temperature evolution of different defect types in the process of DDR chip reading and writing. The thermal signals of different measurement areas are extracted for defects that are difficult to intuitively distinguish by infrared images. With the hyperparameter optimization, the CNN classification model realizes efficient and accurate detection of different defect types, including address, data, and bank address solder joint fault. Furthermore, after transfer learning, the other 9 different solder joint defects of the chip are accurately identified, and the accuracy is over 95% and over 92% under the conditions of 10 and 20 dB Gaussian white noise, respectively. It provides an efficient and effective method for microelectronics packaging and reliability analysis on industrial high-density integrated PCBA.

Abstract:Gas concentration measurement is crucial in the process industry. The measurement principle of non-dispersive infrared (NDIR) sensors is based on the difference in the absorption intensity of infrared light for different gases. It is widely used to measure CO2 and N2O concentrations for reducing the influence of environmental factors. But, the inhomogeneous gas diffusion in NDIR sensor chamber generates principle errors. It is necessary to determine the accurate time when CO2 and N2O reach stable diffusion within the gas chamber to reduce errors and improve the measurement accuracy. In this article, a Gaussian diffusion model is proposed to predict the gas stable diffusion time in the gas chamber of the NDIR sensor to correct gas concentration. The effectiveness of the proposed Gaussian diffusion model is evaluated by the simulation experiment. Fourteen sets of simulations for CO2 and N2O under different conditions show that the fitness of the proposed Gaussian diffusion model is 0. 82 and 0. 80, respectively, which verifies the robustness of the proposed model. The Gaussian diffusion model between gas stable diffusion time and inlet flow velocity provides theoretical support for error correction generated by inhomogeneous gas diffusion which finally improves the measurement accuracy.

Abstract:To accurately detect liquid level and pressure, a fiber optic diaphragm sensor and the measurement system are firstly prepared in this article. The sensor consists of fiber Bragg grating ( FBG), elastic diaphragm, microcavity, FBG fixture and silica gel tube. Secondly, the theoretical model of the sensor to detect the liquid level and pressure is formulated, and the response characteristics of the sensor to the liquid level, pressure and temperature are experimentally studied. Finally, to study the anti-interference ability of the sensor to changes in the external environment, the effects of temperature, inclination angle and liquid level velocity on the measurement results of the sensor are experimentally studied. Results show that when the liquid level velocity is in the range of 10~ 100 cm/ min, the sensor′s inclination angle is in the range of - 30° ~ 30°, and the tested liquid temperature is in the range of 20 ~ 60℃ . A linear relationship between the sensor output and liquid level 0~ 220 cm and pressure 0 ~ 22 kPa is obtained. The measurement results of the sensor are not affected by liquid velocity, sensor′s inclination angle, and temperature change. The liquid level, pressure and temperature sensitivity of the sensor reach 35. 16 pm/ cm, 359. 46 pm/ kPa and 10. 07 pm/ ℃ , respectively. The maximum relative error is 5. 6% .

Abstract:In the application scene of human-computer interaction in the process of upper limb rehabilitation training, it is difficult for a single sensor to track the trajectory of the upper limb accurately and stably. To address this issue, a multi-mode perception fusion scheme based on the inertial sensor and the Kinect sensor is proposed. To solve the problem of cumulative error of inertial sensors in trajectory tracking, an error correction algorithm based on polynomial compensation and zero constraint is proposed. To solve the problem that the imaging quality of the depth image of the Kinect sensor affects the effect of bone tracking, a variable parameter pixel filter is proposed to repair the depth image. Through the motion experiment of inertial sensor, the effectiveness of the cumulative error correction algorithm is verified, and the X-axis displacement error is 7. 8 mm. The depth image restoration results show that the loss of depth pixels is reduced by about 50%, and the image is clearer. Experiments show that the depth image can effectively track the bone trajectory. Finally, the feasibility and effectiveness of the upper limb trajectory tracking scheme based on the multi-mode perception are verified by the trajectory data fusion experiment, and the computing delay is only 1. 2 s.

Abstract:In the process of defect detection for various materials with complex curved surfaces or flat surfaces, conventional tactile sensors have disadvantages, such as small detection area and low detection efficiency. To address these problems, a fingertip type largearea optical tactile sensor is designed and prepared for multi-material surface defect detection. It is similar to the tip of human finger and has both finger-shaped curved and flat contact surfaces, which can meet the detection needs of various complex contact surfaces. A miniature actuator is designed in the sensor to drive the camera rotation to improve the imaging quality, and multiple images are collected by rotation and stitched together using the APAP image stitching algorithm to increase the effective area for single detection. A variety of material surface defects are simulated and a tactile image dataset is created, which is trained by the DeepLabv3 model. Experimental results show that, with a single acquisition, the effective detection area reaches 16. 3 cm 2 , and the model achieves 91. 2% MIoU through training, which enables the detection of defects on complex surfaces and planes of multiple materials.

Abstract:Transmission efficiency is an important index to measure the transmission performance of ball screw. The current researches on transmission efficiency lack in-depth analysis of the friction and lubrication of the contact surface. There are few experimental studies and most of them are for a certain type of sample. To obtain a transmission efficiency model that is more in line with the actual situation and analyze the influencing factors of transmission efficiency, based on the load distribution theory, the bearing and motion of the ball screw are analyzed, which consider the process error. The theory of electohydrodynamic lubrication is also combined to obtain the ball screw in different working conditions. According to the friction coefficient change rule under the conditions, a new transmission efficiency model is achieved. Then, the influence factors of transmission efficiency are calculated and analyzed. The final design test verifies the transmission efficiency model and analyzes the influence of different axial load, speed, preload, and lubrication viscosity on the transmission efficiency. The test results show that the maximum relative error between the calculated value of the new model and the test value is 6. 42% , which evaluates the validity and rationality of the new model.

Abstract:The hard spot detection is one of the key contents during quality evaluation of high-end steel plate products. In this study, a domestically developed instrument based on the principle of incremental eddy current method is presented for detecting hard spots in steel plates. The function and performance of the developed instrument are experimentally examined. First, a sensor of incremental eddy current and its detection circuit are developed by using AC-bridge, phase rotation and demodulation techniques. As a result, a specialized instrumentation of incremental eddy current is integrated. Secondly, several groups of experimental testing are performed on steel plate of Cr12MoV and pipeline steel of X70 with the developed instrument. The obtained results show that the spatial resolution of the sensor can reach around 10 mm. The peak-to-peak value (Up ) of the demodulated waveform of the incremental eddy current signal is used as feature parameter, which demonstrates approximately linear decreasing trend as the increase of the surface hardness. The value of Up is decreased by at least about 75% in the hard spot region in which the hardness is about 34. 8% higher than the baseline. Magnetic imaging of pipeline steel of X70 is realized by using the parameter of Up and the hard spots with a width around 10 mm is accurately identified from the imaging results. The above research results provide support in method and instrument for on-line, nondestructive testing of hard spots in high-end steel plates.

Abstract:The magnetic flux leakage detection is a common nondestructive detection method for ferromagnetic materials with high speed, high sensitivity and simple operation. The finite element model is widely used in magnetic flux leakage detection. The solution accuracy of the finite element model is high. But, the calculation time is long. Therefore, this article proposes a finite element modeling and solution method for fast calculation of 2D defect magnetic leakage field, which can reduce the calculation time while ensuring the accuracy. Through the theoretical derivation of finite element numerical calculation, the finite element model of magnetic leakage field is established. First, the six-node triangular element is used to divide the grid. Then, the sparse matrix is used to solve the equations, and the orthogonal basis of the zero space of the boundary condition matrix is used to solve the magnetic potential. The comparison with the calculation results of COMSOL simulation software shows that the formulated model is correct and feasible. Through the verification of actual testing data, this model has significantly improved performance compared with the traditional finite element model. The maximum accuracy is increased by 83. 49% , and the calculation time can be reduced by 95. 43% . A fast and effective forward model is provided for studying the magnetic flux leakage reconstruction of irregular defects.

Abstract:Low frequency vibrations threaten the safe operation of high-rising structures as their height rise. Passive dampers only have good damping effect at the tuning point and have a narrow damping bandwidth. Magnetic fluid provides a new solution to this problem. In this article, a semi-active tuned current controlled damper is proposed by using magnetic fluid as the working fluid. First, the theoretical basis for the inherent frequency is established. The theory shows that the current can change the inherent frequency. Secondly, the damper test system is established to evaluate the tuned damping capability under different excitation. Results show that the damping rate of the semi-active current controlled damper is about 22% , and the damping rate of the tuned liquid rolling ball damper is only about 12. 3% . The damping rate will be less than 10% when the frequency ratio between the tuned liquid rolling ball damper and the controlled structure is greater than 1. 11. The semi-active current controlled damper has good robustness, which can effectively reduce the displacement response of the structure and make up for the shortage of the traditional passive dampers with narrow damping bandwidth.

Abstract:The high-speed camera tends to produce noise when it works under an ultra-high frame rate (>10 000 FPS). The distribution of the produced noise is complex. Thus, it is very difficult to obtain the noisy-clean image pairs. To address this problem, a training scheme for the convolutional denoising networks based on the non-ideal noisy-clean pairs is proposed. Firstly, the cameras with high and low frame rates are leveraged to capture the images of the same scenes, through which the noisy images and the corresponding non-ideal paired clean images are achieved. Then, a deep denoising model based on a convolutional neural network is formulated to achieve supervised learning with the non-ideal noisy-clean image pairs by utilizing the brightness consistency and image alignment methods, contributing to removing imaging noise. Finally, a model quantization technique is introduced to quantize the values of parameters and activations, which helps them to transform the 32 bit floating-point number to an 8 bit fixed-point number, and thus greatly reduces the model size, memory consumption, and running time. Experimental results show that the proposed denoising method can effectively remove the imaging noise of a high-speed camera. Compared with other methods, the peak signal-to-noise ratio and structural similarity of the denoised image are improved by at least 1. 96 dB and 1. 95% , respectively. In addition, with the help of the model quantization technique, the model size is reduced by 4 times, and the memory consumption and running time are decreased by 45. 62% and 37. 5% , respectively.

Abstract:The end detection of guide wire in interventional surgery plays a vital role in ensuring the accurate control and safety of surgery. In this article, a method of end detection of guide wire based on the improved YOLOv4Tiny network is proposed for the clinical demand of end detection of guide wire during surgery. The YOLOv4Tiny network architecture is utilized in the proposed method. By optimizing the residual structure in the feature extraction network, and enhancing the attention mechanism and the hybrid expansion convolution network, the small target feature extraction ability and detection accuracy are greatly improved. The receptive field is also expanded, which ensures the image resolution without increasing the computation amount. To evaluate the effectiveness of the improved algorithm, it is tested in the constructed dataset and the actual surgical dataset. According to the experimental results, the average accuracy of the improved algorithm in the constructed dataset reaches 97. 6% , with the detection error of the guide wire end be less than 5% , and the average accuracy in the actual surgical dataset is 92. 8% . The improved algorithm is of great reference significance for the end detection of interventional surgical guide wire, which has broad application prospects in fields related to biomedical robots.

Abstract:To address the inaccurate extraction of 3D point cloud of the calibration plate encountered in the process of external parameter calibration of lidar in complex external environment, an optimization method of external parameter calibration of lidar and camera based on background clustering is proposed. The blind detection of the point cloud of the calibration plate is avoided in the whole threedimensional point cloud, which would lead to large error in the calibration results and the need to manually correct the wrong feature points. This method uses the density difference between the background point cloud without calibration plate and the target point cloud with calibration plate in some spatial domains, and obtains the difference coefficient K between the point cloud of calibration plate and the background point cloud through the adaptive spatial threshold model. Then, some three-dimensional points in the two-point cloud are clustered to complete the three-dimensional point cloud extraction of the calibration plate. Experimental results show that this method can accurately and efficiently extract 3D point cloud of calibration plate in complex environment to improve the accuracy of laser radar and camera external parameter calibration. On this basis, the correct projection proportion of point cloud can reach 97. 43% , and the projection error is reduced by about 25. 33% compared with other methods.

Abstract:In terms of the issues that the accuracy of ship lock water level sensor is easily affected by the water quality and the poor adaptability of the traditional image detection methods, a waterline detection method based on semantic segmentation and coarse-to-fine strategy is proposed, and the calibration model with subsection control points is established to calculate the water level. Considering the characteristics of long-range dependence of waterline, the strip atrous spatial pyramid pooling module is proposed. To address the problem of inaccurate segmentation boundary, the multi-path aggregation upsample module is proposed and online hard example mining is introduced to improve the segmentation accuracy of the model. The improved semantic segmentation model is used to perform coarse detection, separates water and non-water regions in the low-resolution image, which is compressed from the original image, and obtains the coarse detection result of the waterline according to the binary mask output from the model. Then, crops the neighborhood of the coarse detected waterline in the original image and performs fine detection to obtain the fine detection result of the waterline. Finally, establishes the calibration model with subsection control points to build the relationship between pixel coordinates and world coordinates, and calculates water level by fine detection result. The experiments are implemented on the constructed water level dataset. Experimental results show that the improved semantic segmentation model increases mIoU by 2. 58% and 1. 98% for coarse detection and fine detection, respectively. The average pixel error of the fine detection result is 1. 89 pixel, which is 52. 3% lower than the coarse detection result. With manual observation of the water gauge as the benchmark, when the working distance of the camera is 23 m, the uncertainty of the measurement result with a confidence level of 95% is 0. 026 m. The proposed method has well adaptability for a variety of outdoor environments such as sunny, cloudy, rainy and snowy days, which provides an available method for water level detection of ship lock.

Abstract:The tunable diode laser absorption spectroscopy tomography (TDLAST) is an important non-intrusive combustion monitoring technology. However, the ill-posedness of the inverse problem of TDLAST causes large errors in gas absorbance densities reconstructed by using traditional algorithms. In this article, the expected patch Log likelihood is introduced into TDLAST to solve the inverse problem. The Gaussian mixture model (GMM) regularized temperature reconstruction technique (GMMTRT) is proposed to image temperature distribution in the combustion field. This algorithm models the distribution characteristics of local absorbance density with GMM, and solves the inverse problem of TDLAST with GMM regularization via the half quadratic splitting method. Both the simulation with data generated from the fire dynamics simulator and the real experiment with the lab-scale TDLAST system show that GMMTRT can clearly indicate the temperature profile in the region of interest with correctly located flame peaks. Compared to the temperature reconstruction algorithm with Tikhonov regularization and the simultaneous algebraic reconstruction technique, GMMTRT can reduce the reconstruction error by15. 42% ~ 36. 16% and 23. 10% ~ 44. 79% , respectively.

Abstract:To address the problem that roller coasters are difficult to locate, this article proposes a roller coaster localization method based on time series alignment. The method firstly uses the dynamic time warping (DTW) to align the inertial measurement unit ( IMU) measured data with the simulated data to obtain the estimated position of roller coaster. Subsequently, the error state Kalman filter (ESKF) is used to update the IMU-based prediction by using the DTW estimation as observation. To improve the accuracy of the estimation, the segment recombination dynamic time warping (SRDTW) algorithm is proposed to solve the distortion problem of DTW. The localization experiments of a roller coaster are conducted using the method in this article. The results show that time series alignment using Z-directional acceleration and pitch angle can provide more accurate position estimation, and the average error after ESKF filtering can reach 0. 24 m, which is 45. 6% lower than average error of estimation alone.

Abstract:Vibration signals of rotating machinery driven by the variable-frequency converter have the characteristics of complex modulation components, wide frequency bands involved, and serious noise interference. It is difficult to extract the mono-component modulation components related to faults. To address these issues, a new variational nonlinear single chirp mode extraction (VNSCME) method is proposed, which formulates a variational optimization model with the combination constraints of the narrowest demodulation frequency band of a single target mode and the minimum energy of the residual component. A specific mono-component nonlinear modulation component is extracted by iteration. By being preset a priori knowledge about the instantaneous frequency of the target mode, VNSCME can independently extract the specific mono-component modulation component and accurately estimate its instantaneous frequency. Compared with the existing researches, VNSCME has advantages of without the resolution limitation of time-frequency distribution, simple initialization and high computational efficiency. The combination of VNSCME and order tracking is applied to the bearing fault diagnosis of the motor driven by variable-frequency converter. The simulated and measured fault vibration signals are processed respectively, and the results indicate that the relative error of instantaneous rotating frequency estimation is less than 0. 76% , and the calculation time for extracting a single target mode is less than 11. 9 s, verifying the effectiveness of the proposed method.

Abstract:The 802. 1CB protocol in the TSN protocol plays an important role in ensuring the reliability of the in-vehicle Ethernet network. It ensures the stable transmission of packets through link redundancy, and the inconsistency of transmission characteristics between different redundant links may affect subsequent traffic scheduling. Therefore, it is necessary to explore the impact of link redundancy on scheduling performance. This article starts from the topology redundancy structure of automobile network, analyzes the correlation between redundancy protocol and scheduling protocol, and discusses the influence of link redundancy on 802. 1Qbv timeaware shaper mechanism and 802. 1Qch cyclic queuing and forwarding mechanism. Finally, we establish a link redundancy scheduling performance evaluation model, based on OMNeT++ and hardware verification platform. We simulate and verify the above effects. The results show that the difference in the arrival interval of redundant traffic and the number of link hops caused by link redundancy has a significant impact on 802. 1Qbv and 802. 1Qch, respectively. In addition, it is necessary to pay attention to the impact of changes in the rate of bridge ports in redundant topology.