The flexible strain sensor has found widespread application due to its excellent flexibility, extensibility,  and adaptability to various scenarios.  This type of sensors face challenges in direction identification owing to  strong coupling between the principal strain and transverse resistance.  In this study, a silver nanowires (AgNWs)/polydimethylsiloxane (PDMS) strain sensor was developed, using a filtration method for preparing the AgNWs film which was then combined with PDMS to create a unidirectional, highly sensitive, fast-responsive,  and linear flexible strain sensor.  When the grid width is 0.25 mm, the AgNWs/PDMS strain sensor demonstrates  an outstanding unidirectional sensitivity, with a strain response solely along the parallel direction of the grid  lines (noise ratio α ≈ 8%), and a fast reaction time of roughly 106.99 ms.  In the end, this sensor’s ability to  detect curvature was also demonstrated through LEDs, demonstrating its potential applications in various fields,  including automotive, medical, and wearable devices.

A Ka-band wideband microstrip-to-microstrip (MS-to-MS) vialess vertical transition on slotline multimode  resonator (MMR) is presented. The proposed transition mainly consists of a slotline MMR on the common  ground plane, and two microstrip (MS) lines facing each other at the top and third layers in the four-layered liquid  crystal polymer (LCP) substrate. In order to improve the bandwidth of the proposed transition, a U-shaped  branch is added to the top- and third-layer MS lines, separately. The slotline MMR can be properly excited by  setting the position of the U-shaped branch line. As such, a three-pole wideband vertical transition is obtained,  which shows a good transmission performance over a wide frequency range of 29.27—39.95 GHz. The three-pole  wideband vertical transition based on multilayer LCP substrate is designed, fabricated, and measured. Test results  indicate that a wide frequency range of 26.84—36.26 GHz can be obtained with return loss better than −10 dB  and insertion loss less than −3dB.

Design of a miniaturized lumped-element bandpass filter in multilayer liquid crystal polymer technology  is proposed. Fractional bandwidth of the bandpass filter is 20%, operating at a center frequency of 500MHz. In  order to further reduce the size and improve the performance of the proposed filter, defected ground structure (DGS) has been implemented in the filter. Based on this structure, the volume of the inductor is reduced by 60% efficiently compared with the inductor without DGS, and the Q-factor is increased up to 257% compared  with the traditional multilayer spiral inductor. The measured results indicate that the designed filter has a very  sharp stopband, an insertion loss of 2.3 dB, and a return loss of 18.6 dB in the passband. The whole volume of  the fabricated filter is 0.032λg × 0.05λg × 0.000 75λg, where λg is the guided wavelength of the center frequency. The proposed filter is easily integrated into radio-frequency/microwave circuitry at a low manufacturing cost,  especially wireless communication.

In this paper, a compact defected ground structure loaded ultra high frequency dual-band bandpass filter is designed and implemented based on multilayer liquid crystal polymer technology. This novel filter is simply composed with several lumped and semi-lumped elements, to create a dual-passband response. In order to enhance the out-of-band rejection, a feedback capacitor Cz at the in/out ports of the filter is introduced, and four transmission zeros (TZs) are obtained outside the pass band. Furthermore, the position of TZs can be determined by adjusting the value of Cz. The schematic and design process of the filter are given in this paper. The center frequencies of dual-band bandpass filter are 0.9GHz and 2.45GHz, and the 3-dB bandwidths are 13.7% and 14.3%, respectively. The circuit size is 11mm × 9.5mm × 0.193mm. The proposed filter has been fabricated and tested, and the measured result is in good agreement with the simulation result.

The stratospheric airship is affected by harsh conditions in the stratosphere environment. To ensure the safety of the airship, it is necessary to detect the material state of the airship envelope. Since digital image correlation possesses non-contact strain measurement ability, this paper explores the influence of different shapes of the subset on measurement accuracy. Through the results, it is found that increasing the aspect ratio of subsets can improve the strain accuracy measured in the x-direction, and reducing the aspect ratio can improve the strain accuracy measured in the y-direction. This trend becomes more obvious as the strain increases. Based on this discovery, a subset adaptive algorithm is proposed. The feasibility of the algorithm is verified by experiments, and the precision of strain measurement can be effectively improved by adjusting the threshold value. Therefore, the algorithm can be utilized to increase the measurement accuracy in the larger strain direction without changing the size of the subset.

To ensure a long-term safety and reliability of electric vehicle and energy storage system, an accurate estimation of the state of health (SOH) for lithium-ion battery is important. In this study, a method for estimating the lithium-ion battery SOH was proposed based on an improved extreme learning machine (ELM). Input weights and hidden layer biases were generated randomly in traditional ELM. To improve the estimation accuracy of ELM, the differential evolution algorithm was used to optimize these parameters in feasible solution spaces. First, incremental capacity curves were obtained by incremental capacity analysis and smoothed by Gaussian filter to extract health interests. Then, the ELM based on differential evolution algorithm (DE-ELM model) was used for a lithium-ion battery SOH estimation. At last, four battery historical aging data sets and one random walk data set were employed to validate the prediction performance of DE-ELM model. Results show that the DE-ELM has a better performance than other studied algorithms in terms of generalization ability.

A novel characterization method for full-matrix constants of PZT-8 piezoceramics based on local electrodes excitation using one sample is proposed to avoid resonant peaks missing and overlapping in the inversion process of resonant ultrasound spectroscopy technology. Elastic matrix, which is sensitive to the resonance spectrum, is obtained by resonant ultrasound spectroscopy. Piezoelectric and dielectric matrices, which are sensitive to the capacitance of driving electrodes, are determined by capacitance inversion. The initial values of elastic constants are deviated by 30% to validate the reliability of this method. The relative errors between measured and inversed values of resonant frequencies are less than 1% and the relative errors of the capacitance are mostly less than 5%. The work has extensive applications in piezoelectric materials characterization.

Permanent magnet synchronous motor (PMSM) is widely used in alternating current servo systems as it provides high efficiency, high power density, and a wide speed regulation range. The servo system is placing higher demands on its control performance. The model predictive control (MPC) algorithm is emerging as a potential high-performance motor control algorithm due to its capability of handling multiple-input and multipleoutput variables and imposed constraints. For the MPC used in the PMSM control process, there is a nonlinear disturbance caused by the change of electromagnetic parameters or load disturbance that may lead to a mismatch between the nominal model and the controlled object, which causes the prediction error and thus affects the dynamic stability of the control system. This paper proposes a data-driven MPC strategy in which the historical data in an appropriate range are utilized to eliminate the impact of parameter mismatch and further improve the control performance. The stability of the proposed algorithm is proved as the simulation demonstrates the feasibility. Compared with the classical MPC strategy, the superiority of the algorithm has also been verified.

Traditional calibration method for the digital inclinometer relies on manual inspection, and results in its disadvantages of complicated process, low-efficiency and human errors easy to be introduced. To improve both the calibration accuracy and efficiency of digital inclinometer, an automatic digital inclinometer calibration system was developed in this study, and a new display tube recognition algorithm was proposed. First, a high-precision automatic turntable was taken as the reference to calculate the indication error of the inclinometer. Then, the automatic inclinometer calibration control process and the digital inclinometer zero-setting function were formulated. For display tube recognition, a new display tube recognition algorithm combining threading method and feature extraction method was proposed. Finally, the calibration system was calibrated by photoelectric autocollimator and regular polygon mirror, and the calibration system error and repeatability were calculated via a series of experiments. The experimental results showed that the indication error of the proposed calibration system was less than 4 , and the repeatability was 3.9 . A digital inclinometer with the resolution of 0.1 ◦ was taken as a testing example, within the calibration points’ range of [ − 90 ◦ , 90 ◦ ], the repeatability of the testing was 0.085◦, and the whole testing process was less than 90 s. The digital inclinometer indication error is mainly introduced by the digital inclinometer resolution according to the uncertainty evaluation.

Zn vapour is easily generated on the surface by fusion welding galvanized steel sheet, resulting in the formation of defects. Rapidly developing computer vision sensing technology collects weld images in the welding process, then obtains laser fringe information through digital image processing, identifies welding defects, and finally realizes online control of weld defects. The performance of a convolutional neural network is related to its structure and the quality of the input image. The acquired original images are labeled with LabelMe, and repeated attempts are made to determine the appropriate filtering and edge detection image preprocessing methods. Two-stage convolutional neural networks with different structures are built on the Tensorflow deep learning framework, different thresholds of intersection over union are set, and deep learning methods are used to evaluate the collected original images and the preprocessed images separately. Compared with the test results, the comprehensive performance of the improved feature pyramid networks algorithm based on the basic network VGG16 is lower than that of the basic network Resnet101. Edge detection of the image will significantly improve the accuracy of the model. Adding blur will reduce the accuracy of the model slightly; however, the overall performance of the improved algorithm is still relatively good, which proves the stability of the algorithm. The self-developed software inspection system can be used for image preprocessing and defect recognition, which can be used to record the number and location of typical defects in continuous welds.

With increasing the number of wind power generators, the consumption time of electromagnetic simulation of the wind farm explodes. To reduce the simulation time while meeting the accuracy requirement, a genetic clustering-based equivalent model is proposed for the wind farm with numerous doubly fed induction generators. In the proposed model, active power together with the reactive power and the wind speed are selected to form the set of clustering indicators. A normalization technique is utilized to cope with the multiple orders of magnitude in these factors. An exponential fitness value is formulated as a function of the sorting number of the primary fitness value, and the fitness-based selection probability is constructed to overcome the property of premature and slow convergence of the genetic clustering algorithm. The sum of squares due to error is used to determine the optimal clustering number. In addition, a decoupled parameter equivalence method is adopted to obtain the equivalent parameters of the collection network. Simulation results and comparisons with various methods under different voltage scenarios show the feasibility and effectiveness of the proposed model.

Aimed at the long and narrow geometric features and poor generalization ability of the damage detection in conveyor belts with steel rope cores using the X-ray image, a detection method of damage X-ray image is proposed based on the improved fully convolutional one-stage object detection (FCOS) algorithm. The regression performance of bounding boxes was optimized by introducing the complete intersection over union loss function into the improved algorithm. The feature fusion network structure is modified by adding adaptive fusion paths to the feature fusion network structure, which makes full use of the features of accurate localization and semantics of multi-scale feature fusion networks. Finally, the network structure was trained and validated by using the X-ray image dataset of damages in conveyor belts with steel rope cores provided by a flaw detection equipment manufacturer. In addition, the data enhancement methods such as rotating, mirroring, and scaling, were employed to enrich the image dataset so that the model is adequately trained. Experimental results showed that the improved FCOS algorithm promoted the precision rate and the recall rate by 20.9% and 14.8% respectively, compared with the original algorithm. Meanwhile, compared with Fast R-CNN, Faster R-CNN, SSD, and YOLOv3, the improved FCOS algorithm has obvious advantages; detection precision rate and recall rate of the modified network reached 95.8% and 97.0% respectively. Furthermore, it demonstrated a higher detection accuracy without affecting the speed. The results of this work have some reference significance for the automatic identification and detection of steel core conveyor belt damage.

For the process of multi-robot collaboration to lift the same lifted object by flexible cables, the existing collision detection algorithm of cables between the environmental obstacles has the problem of misjudgment and omission. In this work, the collision detection of cable vector was studied, and the purpose of collision detection was realized by algorithm. Considering the characteristics of cables themselves, based on oriented bounding box theory, the cable optimization model and environmental obstacle model were established, and a new basic geometric collision detection model was proposed. Then a fast cable vector collision detection algorithm and an optimization principle were proposed. Finally, the rationality of the cable collision detection model and the effectiveness of the proposed algorithm were verified by simulation. Simulation results show that the proposed method can meet the requirements of the fast detection and the accuracy in complex virtual environment. The results lay a foundation for obstacle avoidance motion planning of system.

Representation learning from unlabeled skeleton data is a challenging task. Prior unsupervised learning algorithms mainly rely on the modeling ability of recurrent neural networks to extract the action representations. However, the structural information of the skeleton data, which also plays a critical role in action recognition, is rarely explored in existing unsupervised methods. To deal with this limitation, we propose a novel twostream autoencoder network to combine the topological information with temporal information of skeleton data. Specifically, we encode the graph structure by graph convolutional network (GCN) and integrate the extracted GCN-based representations into the gate recurrent unit stream. Then we design a transfer module to merge the representations of the two streams adaptively. According to the characteristics of the two-stream autoencoder, a unified loss function composed of multiple tasks is proposed to update the learnable parameters of our model. Comprehensive experiments on NW-UCLA, UWA3D, and NTU-RGBD 60 datasets demonstrate that our proposed method can achieve an excellent performance among the unsupervised skeleton-based methods and even perform a similar or superior performance over numerous supervised skeleton-based methods.

In response to the shortcomings of the common encoders in the industry, of which the photoelectric encoders have a poor anti-interference ability in harsh industrial environments with water, oil, dust, or strong vibrations and the magnetic encoders are too sensitive to magnetic field density, this paper designs a new differential encoder based on the grating eddy-current measurement principle, abbreviated as differential grating eddy-current encoder (DGECE). The grating eddy-current of DGECE consists of a circular array of trapezoidal reflection conductors and 16 trapezoidal coils with a special structure to form a differential relationship, which are respectively located on the code plate and the readout plate designed by a printed circuit board. The differential structure of DGECE corrects the common mode interference and the amplitude distortion due to the assembly to some extent, possesses a certain anti-interference capability, and greatly simplifies the regularization algorithm of the original data. By means of the corresponding readout circuit and demodulation algorithm, the DGECE can convert the periodic impedance variation of 16 coils into an angular output within the 360◦ cycle. Due to its simple manufacturing process and certain interference immunity, DGECE is easy to be integrated and mass-produced as well as applicable in the industrial spindles, especially in robot joints. This paper presents the measurement principle, implementation methods, and results of the experiment of the DGECE. The experimental results show that the accuracy of the DGECE can reach 0.237% and the measurement standard deviation can reach ±0.14 ◦ within 360 ◦ cycle.

Passive optical motion capture technology is an effective mean to conduct high-precision pose estimation of small scenes of mobile robots; nevertheless, in the case of complex background and stray light interference in the scene, due to the influence of target adhesion and environmental reflection, this technology cannot estimate the pose accurately. A passive binocular optical motion capture technology under complex illumination based on binocular camera and fixed retroreflective marker balls has been proposed. By fixing multiple hemispherical retroreflective marker balls on a rigid base, it uses binocular camera for depth estimation to obtain the fixed position relationship between the feature points. After performing unsupervised state estimation without manual operation, it overcomes the influence of reflection spots in the background. Meanwhile, contour extraction and ellipse least square fitting are used to extract the marker balls with incomplete shape as the feature points, so as to solve the problem of target adhesion in the scene. A FANUC m10i-a robot moving with 6-DOF is used for verification using the above methods in a complex lighting environment of a welding laboratory. The result shows that the average of absolute position errors is 5.793mm, the average of absolute rotation errors is 1.997 ◦ , the average of relative position errors is 0.972mm, and the average of relative rotation errors is 0.002 ◦ . Therefore, this technology meets the requirements of high-precision measurement in a complex lighting environment when estimating the 6-DOF-motion mobile robot and has very significant application prospects in complex scenes.

In view of the weak ability of the convolutional neural networks to explicitly learn spatial invariance and the probabilistic loss of discriminative features caused by occlusion and background interference in pedestrian re-identification tasks, a person re-identification method combining spatial feature learning and multi-granularity feature fusion was proposed. First, an attention spatial transformation network (A-STN) is proposed to learn spatial features and solve the problem of misalignment of pedestrian spatial features. Then the network was divided into a global branch, a local coarse-grained fusion branch, and a local fine-grained fusion branch to extract pedestrian global features, coarse-grained fusion features, and fine-grained fusion features, respectively. Among them, the global branch enriches the global features by fusing different pooling features. The local coarse-grained fusion branch uses an overlay pooling to enhance each local feature while learning the correlation relationship between multi-granularity features. The local fine-grained fusion branch uses a differential pooling to obtain the differential features that were fused with global features to learn the relationship between pedestrian local features and pedestrian global features. Finally, the proposed method was compared on three public datasets: Market1501, DukeMTMC-ReID and CUHK03. The experimental results were better than those of the comparative methods, which verifies the effectiveness of the proposed method.

Industrial data mining usually deals with data from different sources. These heterogeneous datasets describe the same object in different views. However, samples from some of the datasets may be lost. Then the remaining samples do not correspond one-to-one correctly. Mismatched datasets caused by missing samples make the industrial data unavailable for further machine learning. In order to align the mismatched samples, this article presents a cooperative iteration matching method (CIMM) based on the modified dynamic time warping (DTW). The proposed method regards the sequentially accumulated industrial data as the time series. Mismatched samples are aligned by the DTW. In addition, dynamic constraints are applied to the warping distance of the DTW process to make the alignment more efficient. Then a series of models are trained with the cumulated samples iteratively. Several groups of numerical experiments on different missing patterns and missing locations are designed and analyzed to prove the effectiveness and the applicability of the proposed method.

In repeated zero-sum games, instead of constantly playing an equilibrium strategy of the stage game, learning to exploit the opponent given historical interactions could typically obtain a higher utility. However, when playing against a fully adaptive opponent, one would have difficulty identifying the opponent’s adaptive dynamics and further exploiting its potential weakness. In this paper, we study the problem of optimizing against the adaptive opponent who uses no-regret learning. No-regret learning is a classic and widely-used branch of adaptive learning algorithms. We propose a general framework for online modeling no-regret opponents and exploiting their weakness. With this framework, one could approximate the opponent’s no-regret learning dynamics and then develop a response plan to obtain a significant profit based on the inferences of the opponent’s strategies. We employ two system identification architectures, including the recurrent neural network (RNN) and the nonlinear autoregressive exogenous model, and adopt an efficient greedy response plan within the framework. Theoretically, we prove the approximation capability of our RNN architecture at approximating specific no-regret dynamics. Empirically, we demonstrate that during interactions at a low level of non-stationarity, our architectures could approximate the dynamics with a low error, and the derived policies could exploit the no-regret opponent to obtain a decent utility.

Robot grasp detection is a fundamental vision task for robots. Deep learning-based methods have shown excellent results in enhancing the grasp detection capabilities for model-free objects in unstructured scenes. Most popular approaches explore deep network models and exploit RGB-D images combining colour and depth data to acquire enriched feature expressions. However, current work struggles to achieve a satisfactory balance between the accuracy and real-time performance; the variability of RGB and depth feature distributions receives inadequate attention. The treatment of predicted failure cases is also lacking. We propose an efficient fully convolutional network to predict the pixel-level antipodal grasp parameters in RGB-D images. A structure with hierarchical feature fusion is established using multiple lightweight feature extraction blocks. The feature fusion module with 3D global attention is used to select the complementary information in RGB and depth images sufficiently. Additionally, a grasp configuration optimization method based on local grasp path is proposed to cope with the possible failures predicted by the model. Extensive experiments on two public grasping datasets, Cornell and Jacquard, demonstrate that the approach can improve the performance of grasping unknown objects.