Abstract:Actuators are essential components of flight control systems, and their performance directly affects flight safety. However, most components can only get the performance data 2~ 3 times during their life cycle, and the samples of performance degradation parameters are extremely small, which poses challenges for predicting actuator performance. To solve this problem, a performance prediction method combining statistical analysis and a physical model is proposed. Firstly, statistical analysis is carried out on the batch-type component data, and statistical distribution rules of different stages of the actuator are established. Then, based on the physical model and statistical law of the actuator degradation, a actuator degradation function with probability distribution is established. The function parameters are calibrated based on AMESim simulation to obtain the probability density function under different time and health parameters. Finally, the update method of probability density function based on posterior probability is given for the health parameters obtained by any component. To verify the effectiveness of the method, multiple samples containing 3 data points were used for validation. The results show that the probability of the measured values in the 3σ range of the predicted density function is 92. 27% , which proves that the predicted density function can characterize the degradation rule of actuators with high confidence.

Abstract:Machine learning models have achieved remarkable success in intelligent fault diagnosis, but are mainly applied in static environments. In practical scenarios, new fault category data arrives continuously in the form of streams, and the distribution of the data changes due to changes in the operating conditions of the machinery and equipment, resulting in a continuous stream of data characterized by non-independent homogeneous distribution. This diagnostic problem of non-independently and identically distributed continuous stream data is called the continuous transfer diagnostic problem. To solve this problem, a continuous transfer learning system (CTLS) fault diagnosis method is proposed. The method includes a domain-adaptive learning loss function and a continuous transfer learning mechanism, which can efficiently handle industrial streaming data and learn new categories without replaying old category data. Moreover, a mechanical failure case evaluations validate the performance of the method, and analysis results show that CTLS can effectively handle industrial streaming data under different working conditions and is a promising tool for solving real industrial problems.

Abstract:Anomaly detection based on telemetry data is a key technology for the on-orbit operation and maintenance management of satellite. However, most of the existing methods only use normal samples to build models, while the anomaly detection results are sensitive to the detection threshold, resulting in a high false positive rate. To address this problem, this paper proposes an anomaly detection method based on contrastive time-series reconstruction satellite of telemetry data, which makes full use of the prior knowledge of limited abnormal telemetry samples to enhance the differences between normal and abnormal samples. First, variational autoencoders is used to extract the time-series evolutionary characteristics of telemetry data, specifically the contrastive learning method is introduced to establish an encoder with differentiated outputs of abnormal and normal telemetry data, which uses a large amount of normal telemetry data to further train the whole model to achieve precise time-series reconstruction of normal telemetry data and form a time-series data reconstruction model sensitive to abnormal data. Then the anomaly detection threshold of satellite telemetry data is deduced based on the kernel density estimation method to further improve the detection rate of abnormal samples. Experimental verification was conducted using real satellite telemetry data and the results show that the proposed method can effectively use historical abnormal samples to establish an anomaly detection model, effectively reduce the false positive rate ( all below 0. 002) of anomaly detection and maintain a high detection rate at the same time, keeping a good practical application level.

Abstract:To address the problem that the traditional milling surface roughness prediction method relies excessively on signal processing knowledge to extract features and has low prediction accuracy, a surface roughness prediction method based on a deep residual shrinkage network improved by the inception module ( IDRSN) and a bidirectional long-short-term memory network ( BiLSTM) is proposed. Firstly, the input signal is noise reduced using the soft thresholding structure and attention mechanism in the deep residual shrinkage network. Secondly, the Inception module is introduced to build IDRSN to enhance the multiscale information acquisition capability of the network for adaptive multiscale feature extraction. Then, a bidirectional recurrent network structure is introduced to construct a BiLSTM prediction network, which utilizes both positive and negative LSTM to improve the network′s ability to capture complete information about the past and the future. Finally, experiments verify the effects of four methods of extracting features, IDRSN, DRSN, BiLSTM and manually extract features, and the prediction accuracy of four surface roughness prediction models, BiLSTM, CNN, DRSN, and CNNLSTM, are compared respectively. It is shown that the proposed method has a high prediction accuracy and establishes a method basis for surface roughness prediction of milling machining.

Abstract:To investigate the damage mechanism of carbon fiber reinforced plastics ( CFRP) and monitor its manufacturing process quality, high-resolution non-destructive testing was conducted on the impact-damaged areas of CFRP. An all-optical non-contact photoacoustic microscopy ( AONC-PAM) imaging system was constructed, utilizing a self-developed dual-contrast imaging mode combining optical absorption with backscattering, for high-resolution non-destructive testing of CFRP′s damage areas under different impact energies. Experimental results showed that the spatial resolution of the AONC-PAM system was 2. 9 ± 0. 5 μm. The dualcontrast imaging strategy enabled acquisition of images based on optical absorption and surface scattering characteristics simultaneously, as well as their overlay, at a rate of 2 seconds per frame. The AONC-PAM system revealed more imaging details compared to the conventional brightfield microscopy system, including carbon fiber distribution and other microscopic defects such as fiber breaks and delamination, missing bundles, and wrinkles, with detectable defect sizes ranging 10 ~ 20 μm, facilitating precise quantification of damaged areas.

Abstract:This paper proposes an ultrasonic multi-feature fusion bolt stress measurement method based on extreme learning machine (ELM) to address the non-linearity and ill-posedness issues in traditional ultrasonic bolt stress measurement. Firstly, based on the theory of acoustic elasticity and scattering theory, ultrasonic feature parameters such as the acoustic time difference of ultrasonic waves and the attenuation coefficient of longitudinal waves in polycrystalline materials within the Rayleigh scattering range are extracted from ultrasonic echo signals. Then, by vector dimension reduction, the acoustic time difference, attenuation coefficient and effective load length are selected as the input feature vector to establish an ELM-based ultrasonic multi-feature fusion bolt stress measurement model. A bolt axial stress ultrasonic measurement experimental platform is set up to measure the stresses of bolts of different materials and specifications. The results are compared with those of traditional ultrasonic measurement methods to verify the limitations of traditional ultrasonic detection methods. The measurement results and precision of ELM are compared with other machine learning methods, including back propagation ( BP) and support vector regression ( SVR). The results show that the method proposed in this paper effectively overcomes the shortcomings of traditional ultrasonic measurement methods, which can measure the stress of bolts of different materials and specifications, and has higher measurement accuracy (with an average relative error of 3. 86% ) and better generalization abili K ty.

Abstract:Fan blade is a key component of wind turbine, and its crack fault is particularly common. The presence of cracks can cause damage to the blade or unit. Therefore, based on the tip timing principle and analysis method, a method of fan blade crack fault identification is proposed. Firstly, according to the principle of tip timing, the influence of blade crack on tip offset under load is analyzed, and the mathematical model between tip offset and tip offset time is established. Secondly, through the simulation analysis of blade tip offset degree in different states, combined with the mathematical model between different working condition parameters and tip offset time, the crack characteristic signal is identified. Finally, the results show that the recognition method proposed in this paper can successfully extract more than 92% of the characteristic signal of the crack, and can complete the extraction and analysis of the crack signal in real time, indicating that this method can realize the real-time recognition of the crack fault.

Abstract:Surface conformal circuit based on the 3D printing technology has broad application prospects. However, due to the influence of machining errors in curved substrates and positioning errors in conductive circuits, it is difficult to accurately control the height of the 3D printing head for surface circuits, which affects printing accuracy. Therefore, this article takes pneumatic direct writing printing as the research object and proposes a print head height error compensation method of “ on machine measurement + line model reconstruction”. A curvature adaptive measurement point selection method is designed, which uses the inverse NURBS curve algorithm to reconstruct the measurement data and restore the real line model. By generating the print trajectory of the real circuit and replacing the original theoretical print trajectory, the print head height error is compensated. A 3D printing platform for surface conformal circuits is established independently by integrating with an on-machine measurement system. A comparative experimental plan is also designed on circuit printing before and after compensation with three shapes of circuits: straight lines, arcs, and NURBS-free curves. The results show that before compensation, in areas with high line curvature, the wire drawing phenomenon occurs at the peak of the line on the substrate, and the stacking phenomenon occurs at the valley of the line. The wire width after compensation is relatively uniform. Further measuring the resistance of each wire shows that the resistance values of each line fluctuated significantly before compensation, and the resistance values of the compensated circuit are uniform, with an average decrease of about 65. 99% compared to before compensation, and a maximum resistance decrease of 85. 75% .

Abstract:The accuracy of automated assembly heavily depends on the accuracy of automated pose measurement. The commonly used pose measurement methods are based on non-contact measurement methods, which have low robustness due to lighting, distortion, and other factors. To improve the flexibility and robustness of pose measurement, this article proposes a contact-based pose measurement method. This method is based on a multi-probe contact measurement system and decouples the pose adjustment into independent measurements of pose, center, and phase. Firstly, a visual coordinate system is established by using visual guidance to achieve flexible measurement through self-updating of point measurement paths. Subsequently, coordinate points are obtained by measuring the workpiece plane, circumference, and hole circumference. Finally, the least squares method is used to fit the plane and project the circumference points to fit the center of the circle. The vector space angle formed by the center of the positioning hole is solved to obtain the adjustment of posture, phase, and center. The proposed method can improve measurement efficiency, and the decoupling adjustment method greatly reduces the mutual influence of pose tuning, improving the flexibility and robustness of pose measurement. The effectiveness of the proposed method is evaluated through experiments. The experimental results show that the relative position deviation of the workpiece is less than 0. 075 mm, the attitude angle deviation is less than 0. 02°, and the phase angle deviation is less than 0. 055° after the adjustment of the space pose flexible contact measurement method.

Abstract:In the wind tunnel experiment, high-speed airflow will cause bending deformation and torsion deformation in the wing. This paper proposes a method for measuring the wing bending and torsion deformation based on the error correction model. First, the camera calibration method based on photogrammetry is used to obtain the camera distortion parameter with the use of digital image correlation method to locate and track the distortion-free pixel coordinates of fluorescent points. Then an airflow calibration plate with coordinate axis system is established based on photogrammetry technology, and the airflow coordinate axis calibration plate is used to calibrate the external parameters of the camera and obtain the Y coordinate of the fluorescent marker point on the wing. Finally, the three-dimensional reconstruction models of the single and binocular measurement systems and the error correction model for bending and torsion deformation are established with the known Y-axis constraints of the marker point. When the angle of attack is 0° horizontally, the bending and torsional deformation of the wing in the blowing state is calculated as the reference state with the model. After experimental verification, the measurement error of wing torsional deformation proposed in this article is less than 0. 01°, and the measurement error of wing bending deformation is less than 0. 15 mm / m. This method can provide reliable and robust experimental data for aircraft design.

Abstract:Rapidly obtaining the three-dimensional error information for all tooth surfaces of gears is the key and prerequisite for characterizing the quality of complex tooth surfaces. In this paper, based on the principle of laser triangulation, a three-dimensional measurement model with line laser is established for the gear, and the developed measuring instrument can be used to quickly obtain the error information of three-dimensional tooth surface and evaluate the quality of gear. The instrument adopts the vertical structure consisting of base, precision spindle, circular grating sensor, control system and software system, etc. The precision spindle adopts the dense bead shaft system to achieve high-precision rotation, which ensures the high-precision positioning and rotation of measured gear. Two high-precision line laser sensors are arranged in the circumferential direction of the precision spindle, and the positional state is adjusted according to the parameters of measured gear. Specifically, the rotation angle of precision spindle is acquired by the circular grating in real time, and the geometrical information of gears’ left and right flanks is collected and recorded by the collector in real time. In conclusion the three-dimensional measurement and evaluation software for the gear has been developed, which can realize the measurement and evaluation of gear′ s tooth profile deviation, pitch deviation, topological deviation etc. , and meet the detection requirements of grade 5 precision gear Ke . ywords:gear; gear measurement; gea

Abstract:The current calibration methods of broadband ADCP cannot effectively realize the measurement of flow rate stratification. An acoustic response calibration method based on frequency shift processing by signal resampling is proposed in this paper. This method utilizes a group of response transducers to receive the signals transmitted by each transducer of the ADCP to be calibrated, then some simulated echo signals which contain Doppler shift information are generated by the process of signal segmentation and time-window compensation. In order to be consistent with the real water scattering situation, a calculation model of the echo signal amplitude is given based on the active sonar equation. A broadband ADCP calibration system for multi-cell is established, which could be applied in onland and underwater, and two ADCP devices are calibrated by designing some experiments. The measurement uncertainty of the calibration system is less than 0. 3% v+5 mm/ s (v denotes the flow velocity) (k = 2) in the flow velocity range of 0. 01 ~ 10 m/ s. The proposed method and calibration system are also verified by comparing to the flume towing method through experiment. Keywords:acoustic response calibration method; signal resampling; flow rate stratification; time window compensati

Abstract:The cantilever beam structure serves as a prevalent platform for micro-mass measurements. Conventional measurement methodologies necessitate a stable temperature environment, posing practical challenges. Temperature fluctuations profoundly impact measurement outcomes and pose difficulties in direct decoupling from the cantilever beam′s characteristic equation. This paper introduces a temperature decoupled mass sensing method, leveraging CBAM-CNN and a piezoelectric cantilever beam. Initially, a temperaturecontrolled measurement platform employing a resonant piezoelectric cantilever beam is established to capture impedance response signals across varied mass loads. An adaptive weighted preprocessing method is tailored to augment structural features and accentuate critical information within confined samples. Subsequently, a CBAM-CNN network, incorporating a hybrid domain attention mechanism, is devised to evaluate the relative relationships of multiple resonance peaks in the signals, achieving concurrent temperature decoupled mass sensing. Experimental findings underscore the method′s prowess, attaining an impressive 99. 70% accuracy in mass measurements ranging from 0. 1 g to 1 g within a temperature range spanning 25℃ to 55℃ . Moreover, the method exhibits precise mass sensing across a broad temperature spectrum, obviating the need for temperature compensation.

Abstract:In response to the current situation where the current measurement of three core cables directly uses sensing signals to measure the position and current of multiple conductors, the number of sensors is uncertain and the number of sensors used is large. This not only makes the multi-conductor current measurement system lack a determined sensor usage strategy but also leads to large measurement errors. This article analyzes and provides the minimum and optimal number of magnetic sensors required for the non-invasive measurement of three-core cable current. The principle of non-invasive multi-conductor current measurement is proposed. Based on the magnetic field distribution relationship of the three-core cable current, 6 magnetic sensors are used to surround the cable distribution. The optimal estimation of 2 eigenvalues is obtained from each sensing signal, and 12 independent equations are established to achieve the detection of conductor position and measurement of conductor current in the cable. A three-core cable current measurement platform was built and tested for verification. The test results show that the maximum amplitude error of the method for measuring three-phase currents ranging from 15 to 100 A is less than 1. 00% , the phase angle error is less than 2°, and the conductor position error is less than 0. 2 mm. The proposed measurement principle greatly reduces the online measurement error of multi-core cable current, and the measurement system has a simpler structure and accurate position and current waveform measurement results.

Abstract:According to the challenges of human pose analysis and assessment in domains such as human-computer interaction and medical rehabilitation, this paper introduces a Transformer-based methodology for 3D human pose estimation and the evaluation of action achievement. Firstly, key points of human pose and their joint angles were defined, and based on the deep pose estimation network (DPEN), a Transformer-based 3D human pose estimation model (TPEM) is proposed and constructed, the incorporation of Transformer facilitates better enhanced extraction of long-term sequential features of human pose. Secondly, the TPEM model′s outcomes in 3D human pose estimation are utilized to formulate a dynamic time warping algorithm, which focuses on weighted 3D joint angles. This algorithm temporally aligns pose keyframes for different individuals performing the same action and subsequently introduces an assessment method for action accomplishment to provide scores for the degree of action fulfillment. Finally, through experimental validation across various datasets, TPEM achieves an average joint point error of 37. 3 mm on the Human3. 6 M dataset, while the dynamic time warping algorithm based on weighted 3D joint angles yields an average error of 5. 08 frames on the Fit3D dataset. These results demonstrate the feasibility and effectiveness of the proposed approach for 3D human pose estimation and action accomplishment assessment.

Abstract:In recent years, tool-medicated haptic feedback for virtual surface textures has become a hot topic in the field of haptics. In view of the problems of narrow application range, weak generalization ability and low interactive realism of the existing haptic texture rendering methods, a new texture haptic rendering model is constructed in this paper based on the improved MelGAN. This model takes texture image and real-time user action information as inputs, which can generate vibrotactile signals with high fidelity and has better generalization ability for common texture images. Furthermore, this paper designs a pen-type device with real-time action data acquisition and vibrotactile expression. After collecting vibrotactile signals from real texture surfaces outside the database, this paper compared the performance differences between the proposed model and existing methods in signal generation. The results indicate that the model in this paper achieved the lowest root mean square error ( 0. 173), verifying its ability to perform haptic rendering on unmodeled textures. Finally, this paper conducted two user experiments using a pen-type device. A subjective similarity score of 6. 01 on average indicates that even for new textures outside of the database, our model can provide users with a high level of texture interaction realism.

Abstract:Blood flow velocity profile is utilized to calculate hemodynamic indicators such as wall shear rate, etc. , which are closely related to the progression of atherosclerosis. Ultrafast ultrasound speckle tracking is widely employed for the blood flow velocity profile estimation. However, motion artifacts from blood flow scatterers in multi-angle plane wave compound imaging have an adverse impact on the flow velocity estimation. A block-to-block motion compensation method for the multi-angle plane wave compound imaging is proposed to improve the accuracy of the flow velocity measurements by eliminating motion artifacts at different radial positions by performing BMoCo for every two neighboring frames in the time series of radio frequency signals. Compared with the direct coherent compound, the BMoCo method reduces the normalized root mean square errors of the flow velocity estimation in the simulation and in-vitro experiments by an average of 10. 37% and 37. 82% , which has demonstrated the effectiveness of the B-MoCo method. The in-vivo experiments based on rabbit skeletal arteries have further demonstrated the clinical feasibility of the proposed method. In summary, the B-MoCo method can effectively improve the measurement accuracy of blood flow velocity profile, which is beneficial for the early diagnosis of related cardiovascular diseases K .

Abstract:Accurately and rapidly identifying the switching states in continuous lower limb movements is crucial for natural humanrobot interaction ( HRI ) with exoskeletons. The switching state sEMG signals contain both pre-and post-switching movement information, as well as transient information related to the switching, making them difficult to directly use for recognition. In order to quickly and accurately identify the switching states, this paper proposes a real-time recognition method called FMICMD-LACNN. An adaptive multi-component instantaneous frequency estimation method is proposed to improve the computational efficiency of the multivariate intrinsic chirp mode decomposition ( MICMD) , and a component energy penalty factor is proposed to enhance the decomposition accuracy of MICMD, thus forming the fast multivariate intrinsic chirp mode decomposition (FMICMD) algorithm. For the sEMG signals decomposed by FMICMD, a LACNN recognition model was established to achieve fast and accurate switching states identification. This study collected sEMG signals from 10 subjects in 8 common lower limb continuous motion switching states for experimental verification. The results show that for these 8 switching states, the average recognition accuracy of this method is 98. 35% , and the average recognition time is only about 8 ms, which is better than the CNN-LSTM, E2CNN and CNN-BiLSTM methods. This method has high accuracy and real-time performance, and can meet the needs of fast and natural interaction between the exoskeleton and the human body.

Abstract:Failures of the photovoltaic (PV) module can affect the performance of the PV arrays, thus reducing the power generation efficiency. In serious cases, PV module failures may even jeopardize the safe operation of the power plant. Traditional methods cannot meet the current demand for fast and correct PV module fault detection. Therefore, this paper proposes a PV module fault identification method based on the improved EfficientNet algorithm. First, the collected infrared images of PV modules are utilized to establish a fault dataset, which is then preprocessed by using image segmentation and data enhancement technology. Second, a fault recognition model is constructed based on the EfficientNet network. Meanwhile, a dual-channel convolutional block attention module (CBAM) is introduced into the model, which can inhibit the recognition of unnecessary features and enhance the ability to extract spatial feature information, thus improving the recognition accuracy. Finally, comparative simulation experiments are conducted to validate the effectiveness and advancement of the proposed model. The experimental results show that the fault recognition accuracy of the model reaches 90. 83% , which is 2. 83% higher than that of the original EfficientNet model; in addition, the model size is only 20. 3 M, which shows good practicability and can meet the requirements of practical application of PV power plants.

Abstract:A cross structured light calibration method based on planar targets is proposed to achieve high-precision calibration of line structured light. According to the Blinn-Phong lighting model, it can be inferred that when the laser stripe is projected on the edge of the feature point on a planar target, the cross-section no longer follows a Gaussian distribution, the commonly used center extraction algorithm cannot accurately extract the target. Firstly, morphological operation is used to remove center points at the edges of feature points on flat targets, reducing the error of line fitting from 0. 060 8 pixels to 0. 035 8 pixels, effectively reducing the impact of algorithm extraction errors on calibration results. Secondly, based on Gaussian curve fitting method, correcting the intersection position of the extracted line structured light and auxiliary structured light, based on the Ransac algorithm for line fitting, the results also effectively eliminate points with significant errors caused by uneven lighting and material changes. Experimental results show that the combination of the two methods improves calibration accuracy by 25% , the difference between the three-dimensional coordinates of the center point and the plane equation of the calibration plate is only 4 μm, effectively improve the accuracy of line structure cursor positioning.

Abstract:In response to the challenging issue of indistinct surface rock contours and difficulties in detecting small-sized rocks in dim environments on small celestial bodies, a method and model for rock target detection in landing areas on small celestial body surfaces is proposed. This approach integrates a multi-head self-attention mechanism into the YOLOv8x framework to enhance the model′s capability to capture the global view of images, thereby improving its adaptability to different lighting conditions in deep space environments. Additionally, a small object detection layer is added to the model to increase its focus on small-sized rocks, enhancing its adaptability to rocks of varying sizes. Comparative experimental results demonstrate that compared to the original algorithm, the proposed method achieves improvements of 6. 4% in rock detection precision, 3% in recall rate, and 5% in mean average precision. Furthermore, compared with other mainstream object detection algorithms, the proposed method shows significant improvements in performance metrics. This method provides a theoretical and technical foundation for the autonomous identification of rocks in landing areas on small celestial body surfaces in dim environments. Keywords:rocks detection on small body surf

Abstract:In order to fully extract the spatial-spectral features of hyperspectral image with limited training samples and improve classification accuracy, a hyperspectral image classification method combining dilated convolution and dense network is proposed. Firstly, a multi-scale dilated feature extraction module is constructed by introducing different numbers of dilated convolutional layers and ordinary convolutional layers to increase the receptive field of model through cascading and extract multi-scale features. Then, the dense connections are established between multi-scale dilated feature extraction modules to achieve feature reuse while alleviating the problem of gradient vanishing. However, there are no dense connections within the modules to avoid the problem of building a deep network with excessive network parameters. Finally, the obtained features are sequentially classified through pooling layers, fully connected layers, and Softmax layers. In addition, this study adds the dropout regularization after the fully connected layer to prevent overfitting. Compared with classical classification methods on the Indian Pines and WHU-Hi-Longkou datasets, our method provides an OA of 98. 75% and 98. 82% , respectively. The experimental results show that the network model designed in this study provides the best classification performance at the limited sample conditions.

Abstract:The GNSS signal within the urban canyon areas suffers from the severe blockage and variable quality, which can lead to the inaccurate or even ineffective positioning of intelligent vehicles. To effectively utilize available satellite observations, a multi-sensor fusion method based on camera-aided GNSS / INS integration is proposed. Firstly, a sky-pointing camera is utilized to capture the sky view image and exclude the NLOS measurements, meanwhile the satellites distribution state is defined by the remaining LOS measurements with orthogonal linear regression method. Additionally a factor graph fusion framework based on GNSS / INS integration is proposed by considering the instability of observations, three factors consisting of pseudorange, Doppler frequency, and carrier phase are added for the optimization estimation when the corresponding observation conditions are met. Lastly, the dynamic window optimization rules are designed according to the satellites distribution state, and the length of optimization window is adjusted to follow the change of GNSS blockage. The road tests show that the proposed method enhances more than 40% of positioning accuracy in the blockage interval compared to the conventional fusion method and improves positioning accuracy in urban canyons effectively.

Abstract:This article establishes a position servo control model based on the basic electromagnetic relationship and dynamic mathematical model of Permanent Magnet Synchronous Motor ( PMSM). Following the principle of closed-loop feedback control, a position control system with three closed loops for current, speed, and position is designed. Firstly, the article, according to the characteristics of PMSM vector control, utilizes the coordinate transformation theory of 3S / 2S and 2S / 2R to establish the dynamic mathematical model of the position servo system in the synchronous coordinate system. It also provides the position control strategy and algorithm based on rotor field orientation. Secondly, this paper studies the digital implementation based on Space Vector Pulse Width Modulation (SVPWM) and constructs a three-loop model in the MATLAB/ Simulink environment according to the mathematical model of the motor. Additionally, a prototype based on DSP is built. In simulation and experimentation, the authors employ PI control for both the direct and quadrature axes as well as for speed and position. During the tuning process of control parameters, they adhere to the principle of incremental debugging starting with the inner loop ( current loop), then the speed loop, and finally the position loop, achieving closed-loop control of the synchronous motor′s position. Lastly, the article compares and analyzes the simulation results with the experimental results, demonstrating the correctness and effectiveness of the model.

Abstract:Aiming at the planning problem of the end-effector path and the trajectory in the joint space of the 6-DOF manipulator, an optimal motion planning method based on the hybrid honey badger algorithm and 3-5-3 polynomial interpolation is proposed to achieve the shortest movement path and the optimal joint movement time of the manipulator end-effector, and effectively reduce the difference between the optimized path and the planned path. Firstly, based on the standard honey badger algorithm framework, chaotic reverse learning, transfer operators, sine-cosine operators and adaptive perturbation coefficient strategies are used in initialization, global optimization and local exploration to improve the quality and optimization ability of the optimal solution. Based on this, a collision-free and shortest path planning method was designed for the end-effector of the manipulator to guide the joint trajectory planning process. Secondly, the hybrid honey badger algorithm is used to find the optimal motion time of each joint in the joint space. On this basis, two times 3-5-3 polynomial interpolation algorithm is used to complete the joint smoothing and time-optimal trajectory planning of the manipulator while satisfying the constraints of displacement, velocity and acceleration of each joint. Finally, through simulation comparison with other planning methods, it was verified that the method proposed in this article can shorten the length of the planned path, reduce the joint motion time, reduce the difference between the optimized path and the planned path, and the feasibility of this method was tested using the UR5 manipulator grasping experiment as an example. Keywords:honey badger algorithm; manipulator; motion planning; time optimizati

Abstract:The empirical modeling method based on modern control theory is complex to establish a standard thermal error solution for different production conditions of CNC machine tools. Explore the research on adaptive prediction of thermal errors in CNC machine tools using model-free driving under the digital twin framework. Firstly, a digital twin framework based on the “thermal sensing-mapping fusionoptimization-drive” structure of machine tools is established to achieve the storage and fusion of thermal feature information in the digital twin. Then, based on the assumptions of the MISO system and the dynamic linearization geometric interpretation, a thermal error model free adaptive control (MFAC) method is proposed that is not affected by any structural data of the controlled system. Furthermore, based on the dynamic discovery probability and adaptive step size of the DACS-MFAC algorithm, the system parameters are updated according to a certain period to achieve dynamic optimization of thermal error prediction values in the digital twin system. Experimental result shows that the DACS-MFAC method has advantages such as strong adaptability, high accuracy, and good convergence.

Abstract:The bionics of control and the stability of walking are two important aspects in the gait research of quadruped robots. In order to improve the stability of quadruped robot motion, this paper constructed a CPG model using a Hopf oscillator, which achieved various gaits and transitions between them. We compared the advantages and disadvantages of gait control methods based on CPG and gait planning methods based on trajectory planning in walking. In order to simultaneously utilize the advantages of CPG control and trajectory planning, a neural network is proposed to nonlinearly map the CPG control curve with the driving curve obtained from the inverse kinematics of the foot trajectory, so that the quadruped robot has biomimetic characteristics in control and zero impact characteristics in foot contact. The simulation and experimental results show that the theoretical walking speed of the quadruped robot using CPG gait generation method and trajectory planning method is similar to 80 mm/ s. However, the lateral displacement of the quadruped robot using CPG gait generation method is within ±10 mm and the pitch angle is between ±1. 5°, while the lateral displacement of the quadruped robot using trajectory planning control method is within ±35 mm and the pitch angle is between ±4°, It can be seen that the performance of the two control methods on lateral displacement and pitch motion is inconsistent. Through experimental measurements, it is known that the robot adopts a walk gait with a walking speed of 18. 57 mm/ s, which is close to the theoretical walking speed of 20 mm/ s. After gait conversion, it walks in a trot gait with a walking speed of 76. 15 mm/ s, which is close to the theoretical walking speed of 80 mm/ s. A small error may be caused by assembly and slipping during walking. By measuring its lateral displacement, it can be seen that the lateral displacement is within 15 mm on the left and 25 mm on the right, both of which are within an appropriate range, proving the effectiveness of the proposed algorithm.

Abstract:A force control compensation method for robot polishing system based on algorithm is proposed in order to solve the problems that currently polishing robots cannot achieve both accuracy and compliance in complex environment. First of all, the mechanical characteristics of the robot polishing system and the principle of force control optimization algorithm are explained. Then the experimental system is established to perform the allowable response range and active soft and constant force polishing experiment. Finally, the Extended Kalman filter algorithm, least squares fitting algorithm and particle filter algorithm are used to optimize the real-time compensation value of the polishing force and the compensation effects of each algorithm are compared. The experimental results show that 100% compensation for system structure errors can be achieved within 20 mm through the force control compensation function. Compared with the setting expectation, the average relative error is 5. 44% . After optimization using Extended Kalman filter algorithm, least squares fitting algorithm and particle filter algorithm, the average error is reduced to 1. 20% , 1. 24% and 1. 64% respectively. Expanding and optimizing the real-time / bit compensation function of the robot collaborative control system will help improve the accuracy and stability of the robot′ s polishing system, which provide theoretical basis and technical support for the development of robot technology.

Abstract:To address the issue of low tracking efficiency of robot fish under disturbance, this study proposes an anti-disturbance and adaptive error constraint control scheme. Firstly, by designing a virtual control input and updating the adaptive line-of-sight guidance law using integral link, the motion position deviation caused by side slip is eliminated and the robot′s anti-interference ability is enhanced. Secondly, by constructing an adaptive controller for yaw and surge of the robot fish, the neural network function fits uncertainties and flow disturbances in the model, compensating for system control input with an approximation value. This improves the body′s adaptability to environmental conditions. Finally, utilizing obstacle Lyapunov theory, consistent final boundedness of robot fish tracking position and angle is proven. Through simulation and experiment, compared with the classical guidance scheme, the proposed scheme improves the tracking efficiency and steady-state performance of the robot fish, and the position error convergence rate of the robot fish is increased by 14. 57% on average.

Abstract:Addressing the challenges of a long-tailed distribution of data and low detection accuracy caused by the difficulty in collecting defect samples for printed circuit boards (PCBs) in real-world environments, as well as the high computational complexity when using Vision Transformer (ViT) for detection, we propose an end-to-end PCB defect detection algorithm that incorporates multi-scale ViT feature extraction and attention feature fusion. Firstly, a multi-scale feature extraction network is constructed by combining ViT and partial convolution. Hierarchical multi-head attention is employed to perform adaptive attention operations on different scales of feature maps, enabling the network to better capture local and global information, thereby enhancing its feature extraction capabilities. Partial convolution is utilized to reduce computational costs. Secondly, a non-parametric attention mechanism based on the energy domain suppression effectively fuses multi-scale features, enhancing the expressive power of the network′s fused feature maps. Finally, a classification function sensitive to class imbalance is introduced to improve the loss function of the network, enhancing its fitting ability to imbalanced data and improving generalization. The experimental results on three different types of publicly available PCB datasets indicate that the proposed detection algorithm shows improvement in the mean Average Precision ( mAP) for PCB surface defect datasets, with respective values of 99. 13% , 98. 67% , and 99. 82% . In the case of class-imbalanced PCB defect detection tasks, the mAP is improved by 11. 94% compared to the previous method, and the network achieves a detection speed of 25 FPS, providing a fast and effective approach for PCB defect detection.

Abstract:At present, the multi-core power cable current reconstruction method is limited by the cable specifications or the sensor array and the core need to be placed according to specific rules, which makes it difficult in practical engineering applications. In this paper, a method based on genetic algorithm is proposed to solve the core distribution correlation of three-phase four-core power cable, then the coupling coefficient matrix between the output of each unit of the magnetic sensor array and the current of each core is obtained when the distribution radius of any core and the angle between any core and the sensor array are obtained, which is used to reconstruct the current of each core in the four-core cable. The feasibility of the algorithm is verified by simulation. In order to effectively reduce the error caused by external factors such as interference, a prior solution model is proposed to improve the quality of the solution. Then, the topology of the sensor unit array circuit is designed. The prototype is tested on the experimental platform. The experimental results show that the maximum error of the measured three-phase current is 2. 42% and the maximum phase error is 2. 77° in the case of three-phase balance. In the case of three-phase imbalance, the maximum error of the measured three-phase current is 2. 52% and the maximum phase error is 4. 17°. The experimental results verify the feasibility and effectiveness of the method.

Abstract:Nowadays, with the popularization of low current device in railway communications, sensors and other aspects, electromagnetic interference has attracted more and more attention, however, the work about the coupling analysis of electromagnetic interference mechanism is rare. This study focuses on a typical electromagnetic interference problem of rail pressure sensors. First a physical model of capacitive coupling between power and signal lines is established based on the investigation of interference sources. Then the distributed capacitance parameters between cables are extracted, and the value of coupling electric field is quantitatively analyzed. At the same time, the interference source is identified and the capacitive coupling voltage interference caused by power line crosstalk is determined. Corresponding measurements of suppressing interference are provided, which verifies the analysis. The results show that the grounding device has a great influence on the electromagnetic interference of signal voltage. When the buried depth of the grounding resistance increases from 0. 1 m to 0. 6 m, the electromagnetic interference voltage of the power line is suppressed. Furthermore it′s found that the connecting end of signal line is a weak link valuable to the electromagnetic interference, which is the key to improve the shielding efficiency. The mechanism analysis and calculation method of capacitive coupling proposed in this paper and the suppression measures of electromagnetic interference of sensors can provide a general reference for suppressing the electromagnetic interference of electrical equipment in engineering K .

Abstract:In response to the issue of the inability of SINS / GNSS integrated navigation system to continuously correct errors in the event of a global navigation satellite system signal interruption, a dual-channel Residual-LSTM based SINS / GNSS integrated navigation algorithm is proposed. First, considering the nonlinear correlation difference between the input and output information of the model caused by the different transmission characteristics of SINS longitude and latitude errors, a dual-channel long and short-term memory neural network model structure with different weight coefficients was constructed. A adaptive forgetting information sharing mechanism was introduced to effectively use historical navigation data to fit and predict the longitude and latitude information. Second, in view of the model degradation and gradient vanishing problems existing in deep neural networks, a Residual-LSTM model structure is formed by establishing a Residual-LSTM high-speed channel between multi-layer and dual-channel LSTM networks to increase the information propagation paths between different network layers. Finally, the effectiveness of the proposed algorithm is verified by the real ship data. The experimental results show that compared with the SINS / GNSS integrated navigation algorithm based on conventional intelligence method, the proposed integrated navigation algorithm reduces the longitude error by 51. 97% and latitude error by 31. 45% during the GNSS signal interruption period. Keywords:SINS / GNSS integrated