Search

Quantile regression, which estimates various conditional quantiles of a response variable, including the median (0.5th quantile), is particularly useful when the conditional distribution is asymmetric or heterogeneous or fat-tailed or truncated. Bayesian methods for the inference of quantile regression have been receiving increasing attention from both theoretical and empirical viewpoints but facing the challenge of scaling up when the data are too large to be processed by a single machine under many big data environments nowadays. In this paper, we develop a structure link between Bayesian scale mixtures of normals linear regression and Bayesian quantile regression (B Q R) via normal-inverse-gamma (N I G) distribution type of likelihood function, prior distribution and posterior distribution. We further explore the detailed methods of B Q R for big data, variable selection and posterior prediction. The performance of the proposed techniques is evaluated via simulation studies and a real data analysis. • Algorithms for Bayesian quantile regression in big data scenario are developed. • The proposed approach is based on ALD-based working likelihood functions. • Big data based algorithms for for variable selection is also provided. [ABSTRACT FROM AUTHOR]

3D velocity field estimation is the key to high spatial resolution and high-accuracy measurements in tomographic particle image velocimetry (Tomo-PIV), especially when characterizing flow fields with delicate vortex structures. However, the cross-correlation velocity estimation method has limited spatial resolution due to the limitation of the interrogation sub-volume size. The 3D optical flow method can improve the spatial resolution of the velocity field, but its accuracy needs to be improved because it does not take into account the physical properties of the fluids. In this study, we propose a novel velocity decomposition-based 3D variational optical flow (VD-3DVOF) method to achieve high spatial resolution and high-accuracy measurement of 3D fluids. First, we present a novel regularization term based on the velocity decomposition theorem to constrain the different physical quantities, which can prevent the physical quantities from being over-smoothed. Second, we present a novel data term based on particle volume reconstruction feature weighting to reduce the influence of reconstruction errors on the velocity field estimation accuracy. Third, we present a multiscale technique and a volume warping operation to prevent the solution from falling into local optimal solutions. The newly proposed method considers both the physical properties of the fluid and the errors of reconstructed particle volumes. Velocity fields are estimated over synthetic and experimental particle volumes, and the results and comparisons show that the newly proposed VD-3DVOF method successfully achieves better performance and greater measurement accuracy than existing 3D motion estimation methods. [ABSTRACT FROM AUTHOR]

Recent attempts to employ deep learning technology for the super-resolution (SR) reconstruction of turbulence have focused chiefly on reconstructing two-dimensional (2D) slices of the three-dimensional (3D) flow fields. However, tomographic particle image velocimetry (Tomo-PIV) measurements yield 3D velocity fields. The resolutions of 3D velocity field data along three directions cannot be simultaneously enhanced using a 2D SR reconstruction model. Additionally, a 3D velocity flow field structure cannot be accurately reconstructed using a 2D model. Therefore, we present a 3D flow field SR reconstruction method (FSR-3D) based on a back-projection network. The FSR-3D model incorporates a multiscale convolutional residual block and an up-and-down projection module and establishes the mapping relationship between low-resolution features and high-resolution (HR) features through an iterative mechanism. The proposed model is validated using two experiments. First, the original velocity field is reconstructed from downsampled forced isotropic turbulence data. Second, the downsampled direct numerical simulation (DNS) data of a turbulent channel flow are used to reconstruct the HR flow field to verify the reconstruction capability of the FSR-3D model with respect to the flow field of the boundary layer. Third, Tomo-PIV measurements of the wake flow behind a circular cylinder are used to further verify the generalization ability of the model and the reconstruction effect of the measured data. The experimental results demonstrate that the proposed FSR-3D model can accurately reconstruct the HR 3D velocity field. At the same time, the properties of the reconstruction results, such as their vortex structure, kinetic energy spectrum, and Reynolds stress, are closer to the DNS results than those of the 2D model. The reconstruction results of the Tomo-PIV measurement data show that the FSR-3D model has good generalization ability. [ABSTRACT FROM AUTHOR]

Inkjet printing technology is expected to enhance printed display mass production technology in the future. Nozzle-array printheads form the basis for printed display mass production applications. However, jet instability caused by air bubble entrapment and nozzle wettability changes during the printing process is a major challenge in the application of this technology. To adapt to possible nozzle abnormalities, a high-adaptability nozzle-array printing system based on a set covering printing planning (SCPP) model for printed display manufacturing is designed in this study. The study consists of two parts. First, a printing system based on multistep visual inspection and closed-loop feedback is proposed to accurately detect and screen abnormal nozzle positions. Notably, the inkjet printing system can identify nozzles with abnormal ejection characteristics and ensure that the remaining nozzles work accurately and stably. Then, an SCPP model is established for display pixel printing planning by using the remaining normal nozzles on the nozzle-array printhead. This model can output the most efficient printing path and nozzle printing action and can adapt to any pixel pattern, nozzle type, and abnormal nozzle distribution. The system and technology are highly adaptable and scalable for fabricating large-area printed display devices. [ABSTRACT FROM AUTHOR]

Impulse noise detection is a critical issue when removing impulse noise and impulse/Gaussian mixed noise. In this paper, we propose a new detection mechanism for universal noise and a universal noise-filtering framework based on the nonlocal means (NL-means). The operation is carried out in two stages, i.e., detection followed by filtering. For detection, first, we propose the robust outlyingness ratio (ROR) for measuring how impulselike each pixel is, and then all the pixels are divided into four clusters according to the ROR values. Second, different decision rules are used to detect the impulse noise based on the absolute deviation to the median in each cluster. In order to make the detection results more accurate and more robust, the from-coarse-to-fine strategy and the iterative framework are used. In addition, the detection procedure consists of two stages, i.e., the coarse and fine detection stages. For filtering, the NL-means are extended to the impulse noise by introducing a reference image. Then, a universal denoising framework is proposed by combining the new detection mechanism with the NL-means (ROR-NLM). Finally, extensive simulation results show that the proposed noise detector is superior to most existing detectors, and the ROR-NLM produces excellent results and outperforms most existing filters for different noise models. Unlike most of the other impulse noise filters, the proposed ROR-NLM also achieves high peak signal-to-noise ratio and great image quality by efficiently removing impulse/Gaussian mixed noise. [ABSTRACT FROM AUTHOR]

Inkjet printing, known as digital writing technique, can directly deposit functional materials to form pattern onto substrate. This paper provides an overview of inkjet printing technologies for flexible electronics. Firstly, we highlight materials challenges in implementing flexible devices into practical application, especially for inkjet printing process. Then the micro/nano-patterning technologies of inkjet printing are discussed, including conventional inkjet printing techniques and electrohydrodynamic printing techniques. Thirdly, the related equipments on inkjet printing are shown. Finally, challenges for its future development are also discussed. The main purpose of the work is to condense the basic knowledge and highlight the challenges associated with the burgeoning and exciting field of inkjet printing for flexible electronics. [ABSTRACT FROM AUTHOR]

Abstract: Thermomechanical analysis is presented to study the basic temperature effects on elastomeric substrate of flexible electronics. Strains of a films-on-substrate structure related with three key temperatures are given based on the interfacial continuum model. An improved strain model is given and compared with other two models. The role of the temperature-dependent effects is highlighted and adopted to design a flexible inorganic/organic heterogeneous structure subject to little thermal action. The sensitivity analysis of three key temperatures is investigated, by which proper selection of technological parameter for poly(dimethylsiloxane) fabrication may be determined to eliminate the variation of stress of the interface in circumstances with temperature varying severely. This work contributes to systemic reliability and compatibility, structural design and thermal management of flexible electronics. [Copyright &y& Elsevier]

Recently, it has been shown that a novel parallel manipulator called H4 can be employed to perform high speed motion for semiconductor applications. In this new family of parallel manipulator, singularities are associated with either loss or gain of degrees of freedom (DOF). This paper deals with singularity analysis of H4 utilizing line geometry tools and screw theory. Firstly, the static equilibrium condition of the end effector is derived to obtain the full 6×6 matrix, which is set of governing lines of the manipulator. Based on linear dependency of these lines, the singular configurations of the manipulator can be identified. Moreover, in order to deal with singularities associated with loss of DOFs (serial singularity), the static equilibrium of the actuators is also defined. Secondly, architecture and constraint singularities associated with gain of DOFs (parallel singularity) are defined and analyzed using linear complex approximation algorithm (LCAA), which is employed to obtain the closest linear complex, presented by its screw coordinates, to the set of governing lines. The linear complex axis and pitch provide additional information and a better physical understanding of the manipulator’s self-motion when in or closed to a singular configuration. Lastly, various singularities of an example H4 manipulator are presented and analyzed using the proposed methods. [ABSTRACT FROM AUTHOR]

This paper presents an approach to find feasible and practical plans for mechanical assemblies based on a connector-based structure (CBS) hierarchy. The basic idea of the approach is to construct plans for an assembly (i.e. the root node of the CBS hierarchy) by systematically merging plans for primitive structures in the CBS hierarchy, while the plans for primitive structures are built using one of three methods: by reusing the existing plans, by retrieving the stored plans, and by geometric reasoning. The input to the approach consists of the assembly solid model, the connector-based relational model (CBRM) graph, the spatial constraint graphs, and the selected base part for the assembly, all of which are assumed to be generated previously in an interactive and/or automated way. Then, a CBS hierarchy for the assembly is automatically derived from the input CBRM by firstly establishing its functional model, which is constructed from the CBRM by decomposing the assembly or resulted part sets with respect to their connectors. Based on the CBS hierarchy, a set of assembly precedence diagrams representing good plans for the assembly are generated by merging plans for primitive structures systematically in a bottom-up manner. It can be proved that the proposed approach is both correct and complete. To verify the validness and efficiency of the approach, a variety of assemblies including some complicated products from industry are tested in the experimental CBHAP planner. [Copyright &y& Elsevier]

Electrohydrodynamic (EHD) printing has emerged as a promising technique for high-resolution additive manufacturing, but the frequency limits and optimal operating ranges for drop-on-demand EHD printing remain elusive. This lack of clarity has significant implications for printing efficiency and accuracy. Here, we present a theoretical model based on the Lagrange equation to calculate the eigenfrequency in EHD printing. Our study elucidates the dynamic relationships between the eigenfrequency and key process parameters (e.g., voltage, meniscus diameter, and center angle), and reveals the intrinsic relationship between the jetting frequency and the oscillation eigenfrequency, demonstrating that the jetting frequency is limited by the eigenfrequency. This conclusion has significant implications for the design of high frequency EHD printing systems; it highlights opportunities to increase the eigenfrequency by optimizing parameters such as meniscus radius/center angle. In addition, it is interesting to note that the high-order-mode jetting phenomenon can be utilized to break through the traditional frequency limit of the low-order mode. This work can serve as an instructive benchmark of the EHD printing frequency, which provides a basis for designing a high-frequency EHD printing system. [ABSTRACT FROM AUTHOR]

Flexible sensors have been developed for the perception of various stimuli. However, complex deformation, usually resulting from forces or strains from multi-axes, can be challenging to measure due to the lack of independent perception of multiaxial stimuli. Herein, flexible sensors based on the metamaterial membrane with zero Poisson's ratio (ZPR) are proposed to achieve independent detection of biaxial stimuli. By deliberately designing the geometric dimensions and arrangement parameters of elements, the Poisson's ratio of an elastomer membrane can be modulated from negative to positive, and the ZPR membrane can maintain a constant transverse dimension under longitudinal stimuli. Due to the accurate monitoring of grasping force by ZPR sensors that are insensitive to curvatures of contact surfaces, rigid robotic manipulators can be guided to safely grasp deformable objects. Meanwhile, the ZPR sensor can also precisely distinguish different states of manipulators. When ZPR sensors are attached to a thermal-actuation soft robot, they can accurately detect the moving distance and direction. This work presents a new strategy for independent biaxial stimuli perception through the design of mechanical metamaterials, and may inspire the future development of advanced flexible sensors for healthcare, human–machine interfaces and robotic tactile sensing. [ABSTRACT FROM AUTHOR]

On‐skin electronics have been widely used in fields, such as wearable healthcare and human–machine interfaces due to their excellent performances and wearable comfort. Among them, on‐skin sensors for collecting signals have been greatly developed, while the stimulators that provide tactile feedback are difficult to achieve in terms of flexibility and stretchability due to their rather complicated structures. Achieving on‐skin stimulators with superior performances and outstanding wearability requires major advances in materials, fabrication, and structure design. The on‐skin stimulator needs to be soft with a lightweight and thin structure so that it can be laminated onto the skin with conformal contact for accurate feedback. Three stimulation modes, namely electrotactile, thermal, mechanotactile, and vibrotactile stimulation, are investigated. Among them, electrotactile stimulation is the characteristics of high resolution and quick response. Thermal stimulation can provide a unique heat sensation without crosstalk with other stimuli. Mechanotactile and vibrotactile stimulation can bring comfortable feedback to skin, showing great potential in wearable applications. This paper summarizes the working principles, materials, structures, and applications of these on‐skin stimulators, and compares their advantages with conventional rigid devices. At the end of this paper, the challenging issues and future prospects in this field are discussed. [ABSTRACT FROM AUTHOR]

Texture recognition remains a challenging visual task due to the complex appearance variations caused by scale changes in the real world. In most existing texture recognition methods, textures are represented at a single scale; thus, multi-scale texture information is not fully utilized, resulting in insufficient representation and inaccurate recognition. In this study, with the goal of addressing the challenge of scale changes, we propose a novel multi-scale boosting feature encoding network (MSBFEN) for accurate texture recognition. MSBFEN first extracts multi-scale features with multi-scale texture structure information under the guidance of texture priors using a novel prior-guided feature extraction (PFE) method. Then, a multi-scale texture encoding (MSTE) method is devised to capture discriminative multi-scale texture representations by encoding the extracted features. Finally, to fully utilize the multi-scale texture representations for accurate texture recognition, a novel multi-scale boosting learning (MSBL) method is proposed. In MSBL, the learning procedure for multi-scale texture recognition is boosted in a hierarchical, progressively reinforced manner, significantly addressing the challenge of scale changes and greatly enhancing the recognition accuracy. In addition, a novel outlier-aware texture encoding (OTE) method is proposed for robust texture encoding at each scale of MSTE. OTE can resist the influence of background interference and can further enhance the robustness of MSBFEN. In extensive experiments conducted on six challenging texture recognition datasets, namely, KTH-TIPS2b, FMD, DTD, MINC, GTOS and GTOS-mobile, MSBFEN achieves accuracies of 86.2%, 86.4%, 77.8%, 85.3%, 86.4% and 87.57%, respectively, representing state-of-the-art texture recognition performance. [ABSTRACT FROM AUTHOR]

Fast-moving soft robotics usually suffer large-strain deformation with a relatively high velocity. To improve their performance, it is necessary to study the underly dynamical mechanism. However, the inherent strong material nonlinearity brings considerable difficulties in modeling and calculating the soft structure’s nonlinear dynamics. This work focuses on the large-strain vibrations of the hyperelastic rod under an axial tensile force. Several typical hyperelastic models, including the neo-Hookean, Mooney, and Yeoh models, are used to derive the equations governing the rod’s dynamics. The equations based on these models are relatively complex, and their corresponding numerical calculation diverges in some cases. Therefore, a reference stretch ratio(s) polynomial fitting (RSRPF) model is proposed to derive the concise governing equation. Such a governing equation is effectively solved by the combination of the Galerkin discretization and the fourth-order Runge–Kutta integration algorithm. The correctness of the developed model and the employed algorithm is verified by comparing with the previous data of a hyperelastic membrane. In most cases, the rod periodically vibrates under a periodic tensile force. However, when the structural damping is small and the load frequency approaches the resonant frequency, the hyperelastic rod may quasi-periodically vibrate. The revealed large-strain vibration mechanism is supposed to be useful for guiding the dynamical design of soft structures. [ABSTRACT FROM AUTHOR]

Despite their remarkable effect on printing accuracy and uniformity, charge migrations that dominate the deformation of ink droplets during electrohydrodynamic jet printing have not been widely investigated. In this work, the large deformation mechanisms of a conductive nanodroplet under a strong electric field are examined from the point of view of charge migrations. It is found that the charge migrations include the charge relaxation in the bulk of the droplet and surface charge convection at the fluid interface. A conductive nanodroplet first evolves into an ellipsoid through charge relaxation. Then, the ellipsoid is deformed by the convection of the surface charges in four modes, namely, tip streaming (mode 1), lobe formation (mode 2), finger stretching (mode 3), and dumbbell stretching (mode 4). Finally, the stretched nanodroplet is broken into secondary droplets. Modes 1, 2, and 4 are in agreement with the experimental observations. Furthermore, it is found that over 20% of the charges are distributed inside the bulk nanodroplet and the other charges are distributed at the surface, causing the four deformation modes. Analysis based on the electric Reynolds number (the ratio of electric field force to viscous force) and the Coulombic capillary number (the ratio of surface tension to Coulombic force) shows that the nanodroplet is prolate if the electric field force is dominant. When the Coulombic force plays a crucial role, the nanodroplet deforms into an ellipsoid with wide cones. By contrast, the nanodroplet will generate hemispherical ends if the deformation is dominated by the effect of surface tension. [ABSTRACT FROM AUTHOR]

Object detection remains a challenging task in computer vision due to the tremendous extent of changes in the appearances of objects caused by clustered backgrounds, occlusion, truncation, and scale change. Current deep neural network (DNN)-based object detection methods cannot simultaneously achieve a high accuracy and a high efficiency. To overcome this limitation, in this paper, we propose a novel multi-scale attention (MSA) DNN for accurate object detection with high efficiency. The proposed MSA-DNN method utilizes a novel multi-scale feature fusion module (MSFFM) to construct high-level semantic features. Subsequently, a novel MSA module (MSAM) based on the fused layers of the MSFFM is introduced to exploit the global semantic information of image-level labels to guide detection. On the one hand, MSAM can capture global semantic information to further enhance the semantic feature representation of the fused layers constructed by the MSFFM, thereby improving the detection accuracy. On the other hand, the MSA maps generated by MSAM can be employed to rapidly and coarsely locate objects at different scales. In addition, an attention-based hard negative mining strategy is introduced to filter out negative samples to reduce the search space, dramatically alleviating the severe class imbalance problem. Extensive experimental results on the challenging PASCAL VOC 2007, PASCAL VOC 2012, and MS COCO datasets demonstrate that MSA-DNN achieves a state-of-the-art detection accuracy while maintaining a high efficiency. Furthermore, MSA-DNN significantly improves the small-object detection accuracy. [ABSTRACT FROM AUTHOR]

The IC chip-ejecting and pick-up process plays a critical role in advanced packages since the success ratio and productivity are determined by the delamination of the chip-on-substrate structure. The paper investigates analytically the interfacial peeling mechanism of a chip-on-substrate structure subjected to a transverse concentrated load resulting from ejecting needle from the fracture mechanics point of view. The effects of key factors, including chip size, initial crack length, and substrate material, are uncovered. Finite element calculations are performed to obtain the interfacial peeling energy-release rate by using virtual crack-closure technique with dummy nodes. Analytical formulas and numerical results match fairly well for the entire range of the chip length and the crack length. It is shown that the greater the ratio of length to thickness of the chip is, the smaller the energy-release rate is, and length is the more important factor than thickness to affect the peeling. It implies the interfacial peeling gets tougher for thin or big chips during the pick-up process. Second, as the crack grows, the peeling energy-release rate increases. Third, the softer the substrate is, the greater the peeling energy-release rate is under the action of a constant transverse load. For the pick-up of thin or big chips, in order to achieve high success rate and suppress damage to the chip, it is suggested that more compliant and thinner substrate be adopted, weaker adhesive strength be chosen, and multiple needles be used. [ABSTRACT FROM AUTHOR]

The recently developed hard-magnetic soft (HMS) materials manufactured by embedding high-coercivity micro-particles into soft matrices have received considerable attention from researchers in diverse fields, e.g., soft robotics, flexible electronics, and biomedicine. Theoretical investigations on large deformations of HMS structures are significant foundations of their applications. This work is devoted to developing a powerful theoretical tool for modeling and computing the complicated nonplanar deformations of flexible beams. A so-called quaternion beam model is proposed to break the singularity limitation of the existing geometrically exact (GE) beam model. The singularity-free governing equations for the three-dimensional (3D) large deformations of an HMS beam are first derived, and then solved with the Galerkin discretization method and the trust-region-dogleg iterative algorithm. The correctness of this new model and the utilized algorithms is verified by comparing the present results with the previous ones. The superiority of a quaternion beam model in calculating the complicated large deformations of a flexible beam is shown through several benchmark examples. It is found that the purpose of the HMS beam deformation is to eliminate the direction deviation between the residual magnetization and the applied magnetic field. The proposed new model and the revealed mechanism are supposed to be useful for guiding the engineering applications of flexible structures. [ABSTRACT FROM AUTHOR]

Automated defect inspection has long been a challenging task especially in industrial applications, where collecting and labeling large amounts of defective samples are usually harsh and impracticable. In this paper, we propose an approach to detect and localize defects with only defect-free samples for model training. This approach is carried out by reconstructing image patches with convolutional denoising autoencoder networks at different Gaussian pyramid levels, and synthesizing detection results from these different resolution channels. Reconstruction residuals of the training patches are used as the indicator for direct pixelwise defect prediction, and the reconstruction residual map generated in each channel is combined to generate the final inspection result. This novel method has two prominent characteristics, which benefit the implementation of automatic defect inspection in practice. First, it is absolutely unsupervised that no human intervention is needed throughout the inspection process. Second, multimodal strategy is utilized in this method to synthesize results from multiple pyramid levels. This strategy is capable of improving the robustness and accuracy of the method. To evaluate this approach, experiments on convergence, noise immunity, and defect inspection accuracy are conducted. Furthermore, comparative tests with some excellent algorithms on actual and simulated data sets are performed. Experimental results demonstrated the effectiveness and superiority of the proposed method on homogeneous and nonregular textured surfaces. [ABSTRACT FROM AUTHOR]

Emerging flexible electronics industries, such as solar cells and RFID tags manufactures, adopt the roll-to-roll (R2R) technology to improve their production efficiency. However, the flexible webs used in these fields wrinkle easily due to their tiny thickness and complex structure. Moreover, the webs are easily buckled if the rollers are misaligned. Therefore, it is desired to introduce an efficient method which can help to determine the critical buckling conditions of the webs when there is misalignment of rolls during their production. The current research presents a model based on the Kirchhoff plates buckling theory to determine the critical buckling misalignment and tension for webs transported through misaligned rollers. Webs are modeled as tensioned Timoshenko beams to obtain their stress distributions. The critical buckling conditions are calculated using the Galerkin method in the current research. Results indicate that the calculated critical conditions are close to the practical ones considering the parabolic variation of shear stress and the linear variation of bending stress. The results also indicate that the critical misalignment increases with the increase of the span ratio and tension, and critical misalignment decreases with the increase of the elastic modulus and web width. Moreover, the critical tension increases as the thickness and elastic modulus increases. The presented model provides a useful method to design and maintain R2R equipment. [ABSTRACT FROM AUTHOR]

The motion planning module is the core module of the automated vehicle software system, which plays a key role in connecting its preceding element, i.e., the sensing module, and its following element, i.e., the control module. The design of an adaptive polar lattice-based local obstacle avoidance (APOLLO) algorithm proposed in this paper takes full account of the characteristics of the vehicle's sensing and control systems. The core of our approach mainly consists of three phases, i.e., the adaptive polar lattice-based local search space design, the collision-free path generation and the path smoothing. By adjusting a few parameters, the algorithm can be adapted to different driving environments and different kinds of vehicle chassis. Simulations show that the proposed method owns strong environmental adaptability and low computation complexity. [ABSTRACT FROM AUTHOR]

A new model of the radio frequency identification antenna using the solenoid coil, surrounded by a coaxial ferrite toroid, is proposed in this paper. The truncated region eigenfunction expansion method is adopted to derive the expressions of the magnetic vector potential. Besides, due to the low conductivity and high permeability of the ferrite, a convenient solution region division is introduced to avoid solving the transcendent eigenvalue equation caused by the end effect as usual. Numerical calculations are performed, and the results are in excellent agreement with those obtained from the finite-element method simulations. [ABSTRACT FROM AUTHOR]

Precision tomographic reconstruction is critical for obtaining high-accuracy velocity measurements in tomographic particle image velocimetry. Traditional tomographic reconstruction methods, such as the multiplicative algebraic reconstruction technique (MART), can only be applied at low particle concentrations, limiting the spatial resolution of velocity measurements. In addition, the actual shape of the particles is not reconstructed well due to the limited views. In this study, we propose a novel method named particle field deconvolution MART (Deconv-MART) to repair the shape of actual particles while suppressing ghost particles reconstructed by MART iterations. This method first uses the Gaussian particle shape prior to estimate the convolution kernel obtained by MART reconstruction. Then, the estimated kernel is utilized to deconvolute the particle field and suppress ghost particles based on the prior information of the lower intensity of ghost particles as well as the sparsity of the particle field. Reconstruction fields are estimated with numerical and real experiments, and the results are compared with the results of advanced reconstruction methods. Comparisons of reconstruction demonstrate that the proposed method is effective at suppressing ghost particles and restoring the shape of actual particles. Comparisons of velocity measurements reveal that Deconv-MART has good performance and high measurement accuracy. [ABSTRACT FROM AUTHOR]

In complex robot applications, such as human−robot interaction and robot machining, robots should interact with an unknown environment. To learn the interactive skill, a model-based actor−critic learning algorithm and a safety-learning strategy are proposed in this article to find the optimal impedance control, in which the learning process is safe and fully automatic and does not know the system parameter. In the learning algorithm, a critic is defined as a quadratic form of the system states and the external force. A modified deterministic policy gradient algorithm is presented to improve the learning efficiency. The proposed approach utilizes a model-based constraint and a highly efficient learning algorithm. In the safety-learning strategy, the robot is trained under a constant force, and the learned impedance control can transfer to different interaction situations by choosing the suitable impedance index. The effectiveness of the learning algorithm and the performance of the learned impedance control are validated in a UR5 robot. The robot can perform human−robot interaction and robot machining tasks after the training process with 100 s training time. [ABSTRACT FROM AUTHOR]

The arrival of surgical robots in high-end medical equipment is a landmark, and the realization of tactile sensation a major challenge in this important cutting-edge research field. Aiming to address this issue, we present ultra-sensitive ionic electronic skin in the form of flexible capacitive pressure sensors, which incorporate multistage bionic microstructures in ion gels for the purpose of monitoring the delicate operations of surgical robots. Significantly, the ionic skin exhibits an ultra-high sensitivity of 9484.3 kPa−1 (<15 kPa), and the sensitivity remains higher than 235 kPa−1 in the wide range of 15–155 kPa. The device has also achieved a detection limit as low as 0.12 Pa or, equivalently, 0.31 mg, fast response within 24 ms, and high robustness (loading/unloading for 5000 cycles without fatigue). The sensor facilitates the challenging task of tele-operated robotic threading, which exceeds the human tactile perception limit when threading a needle. We have also confirmed that ionic skin can be used in robot-assisted invasive surgery, such as incision/resection of tissues and suturing of wounds, providing tactile information to surgeons to improve operation success rates. The flexible ionic skin is capable of conforming to the various shapes of robotic manipulators, thus has great promise for applications in robotic dexterous manipulation, prosthetics and human–machine interfaces. [ABSTRACT FROM AUTHOR]

Electrohydrodynamic (EHD) printing is a promising micro/nanofabrication technique, due to its ultra-high resolution and wide material applicability. However, it suffers from low printing efficiency which urgently calls for a high density and addressable nozzle array. This paper presents a nozzle array chip made of a silicon plate and polymer nozzle structure, where the large silicon plate is conducive to a uniform spatial electric field distribution, and the polymer SU8 nozzle can inhibit tip discharge due to its insulating character and liquid flooding as SU8 is hydrophobic. By carefully designing the nozzle array structure via simulation, and fabricating it through MEMS technology, a high-density nozzle array chip has been achieved which can generate very uniform dots without crosstalk. Meanwhile, by adding extractors underneath the nozzle array, and utilizing a digital switch array to tune their on/off state, addressable printing has been realized. This novel printhead design has solved the discharge, liquid flooding, and crosstalk behavior in EHD nozzle arrays, and is compatible with traditional silicon-based MEMS technology, which will promote the practical applications of EHD printing in micro/nanoelectronics, biomedical/energy devices, etc. [ABSTRACT FROM AUTHOR]

Electrohydrodynamic (EHD) jetting is a promising technique with ultra-high resolution, broad compatibility with various inks, and tunable printing modes. However, it suffers from poor efficiency, which urgently calls for an addressable multi-nozzle printhead. Unlike traditional EHD printing which treats the liquid inside the printhead as an equal potential body, this work has proposed to realize addressable jetting by tuning the potential drop between the non-triggered nozzles (connected to a resistor and then grounded) and the triggered nozzles (connected to a high-voltage) through voltage division principle. We present an equivalent circuit model to predict the potential drop within the printhead, which largely depends on the liquid conductivity, channel character, and external auxiliary resistors, and these results match well with experiments and simulations. Besides, we further investigate the jet deflection behavior of the printhead, finding that 1.5 < H / d n < 3 (the ratio of printing height to the nozzle diameter) and 0.6 < k < 0.8 (the ratio of the non-triggered nozzles to the triggered nozzles) are preferred working area for addressable and precise EHD printing. This addressable design does not need a complex extractor underneath the nozzle, which is promising for future high-density and large-scale EHD printheads. [ABSTRACT FROM AUTHOR]

In many industrial processes, sessile droplets are well separated in an ordered array and evaporate to form various patterns. For an evaporating droplet in arrays, the presence of surrounding droplets causes a shielding effect that, in turn, leads to a decrease in the evaporation rate compared to the same droplet in isolation. Here, we demonstrate that, theoretically and experimentally, the shielding effect results in a significant increase in the lifetime of an evaporating droplet. Based on a recent theory, we determine the lifetimes of evaporating droplets in the ordered array. The theoretical prediction shows excellent agreement with our experimental data and even performs well outside its range of validity. These findings strengthen our fundamental understanding of interactions between evaporating droplets in arrays and provide new strategies for controlling the droplet evaporation. [ABSTRACT FROM AUTHOR]

With the development of integrated circuit (IC) chips, the semiconductor industry has begun using many ultrathin chips. To prevent cracking when they are picked up by the machinery, the precise pose of every chip must be provided using the image matching approach. Traditional feature point matching methods are robust to the pose changes of a single chip, but they are not effective for positioning multiple IC instances in one image. To solve this problem, a novel spatial-constraint-based feature point matching algorithm that uses both local information and the spatial topological relationship between feature points to divide the feature points among different IC instances is proposed. Various experiments using images captured of RFID chip pick-up equipment demonstrate that our algorithm results in correct positioning for multiple IC chips; is robust to translation, rotation, scaling, shearing and perspective transformation; and is not sensitive to occlusion and clutter. This algorithm may have many applications in the field of IC manufacturing and packaging. [ABSTRACT FROM AUTHOR]

Outlier (spurious vector) is a common problem in practical velocity field measurement using particle image velocimetry technology (PIV), and it should be validated and replaced by a reliable value. One of the most challenging problems is to correctly label the outliers under the circumstance that measurement noise exists or the flow becomes turbulent. Moreover, the outlier's cluster occurrence makes it difficult to pick out all the outliers. Most of current methods validate and correct the outliers using local statistical models in a single pass. In this work, a vector field correction (VFC) method is proposed directly from a mixture statistical model of PIV signal. Actually, this problem is formulated as a maximum a posteriori (MAP) estimation of a Bayesian model with hidden/latent variables, labeling the outliers in the original field. The solution of this MAP estimation, i.e., the outlier set and the restored flow field, is optimized iteratively using an expectation-maximization algorithm. We illustrated this VFC method on two kinds of synthetic velocity fields and two kinds of experimental data and demonstrated that it is robust to a very large number of outliers (even up to 60 %). Besides, the proposed VFC method has high accuracy and excellent compatibility for clustered outliers, compared with the state-of-the-art methods. Our VFC algorithm is computationally efficient, and corresponding Matlab code is provided for others to use it. In addition, our approach is general and can be seamlessly extended to three-dimensional-three-component (3D3C) PIV data. [ABSTRACT FROM AUTHOR]

With the development of integrated circuit (IC) chips, many elongated chips with small amounts of textures have begun to be used in IC packaging and manufacturing. Traditional template-matching methods are not effective for positioning these chips; this limitation results in IC equipment that cannot accurately pick up an elongated IC chip or may package a nonworking IC chip onto an electronic device. In this paper, a novel region-based normalized cross correlation algorithm that uses a signal function to separate the influences of the IC chip and the background pixel information is proposed. To obtain an increased speed, the bounding box of the IC chip is employed to rapidly estimate the region of interest of an IC chip in a target image. Various experiments using LED sorting equipment demonstrate that our algorithm results in correct positioning for elongated IC chips and can distinguish good chips from nonworking chips with reasonable accuracy. When the size of the target image is 512\times 512 pixels and the size of the template image is smaller than 140\times 140 pixels, the matching time of our algorithm is less than 14 ms. This algorithm may have many applications in the field of IC packaging. [ABSTRACT FROM AUTHOR]

Traditional human emotion recognition is based on electroencephalogram (EEG) data collection technologies which rely on plenty of rigid electrodes and lack anti‐interference, wearing comfort, and portability. Moreover, a significant distribution difference in EEG data also results in low classification accuracy. Here, on‐skin biosensors with adhesive and hydrophobic bilayer hydrogel (AHBH) as interfaces for high accuracy emotion classification are proposed. The AHBH achieves remarkable adhesion (59.7 N m−1) by combining the adhesion mechanism of catechol groups and electrostatic attraction. Meanwhile, based on the synergistic effects of hydrophobic group rearrangements and surface energy reduction, the AHB‐hydrophobic layer exhibits 133.87° water contact angles through hydrophobic treatment of only 0.5 h. Hydrogen and electrostatic bonds are also introduced to form a seamless adhesive‐hydrophobic hydrogel interface and inhibit adhesion attenuation, respectively. With the AHBH as an ideal device/skin interface, the biosensor can reliably collect high‐quality electrophysiological signals even under vibration, sweating, and long‐lasting monitoring condition. Furthermore, the on‐skin electrodes, data processing, and wireless modules are integrated into a portable headband for EEG‐based emotion classification. A domain adaptive neural network based on the transfer learning technique is introduced to alleviate the effect of domain shift and achieve high classification accuracy. [ABSTRACT FROM AUTHOR]

Electronic‐skin (E‐skin) has been investigated extensively for robotic tactile sensing. However, E‐skin sensors based on flexible metamaterials are still challenging to achieve. Moreover, the implementation of E‐skin sensor arrays in the actual monitoring of robotic grasping and manipulation conditions are rather limited due to the difficulty in data processing. Herein, high‐performance E‐skin strain sensors based on flexible auxetic metamaterials are reported, which endow the sensors with the capability of measuring both compressive (40%) and tensile (>80%) strain in a wide range and superior sensitivity, as compared with sensors without the structure. With perception data collected by the sensors, a generic method for real‐time detection of unstable robotic grasping is established. Through this method, the complicated problem of processing large‐scale arrayed sensor signals is simplified into the calculation of two indices, which extract both time and frequency domain characteristics of the signals. The total detection time (including sensor measurement response and data processing) can be as short as 100 ms, in line with human skin response in slippage perception. Accurate detections in real‐time during various grasping and manipulation tasks are presented, demonstrating the great value of the sensors and the detection approach in robotic perception and dexterous manipulation. [ABSTRACT FROM AUTHOR]

This study describes the transformation of molecular orientation which has a significant effect on the charge carrier mobility of organic thin film transistors. Theoretical analyses and molecular dynamics simulations are performed to uncover the mechanism of the orientation transformation of pentacene molecules on the amorphous SiO 2 surface during the early stage of the physical vapor deposition process, where the initially grown lateral-oriented cluster will transfer to the normal-oriented one at a critical size n c . The reorientation behavior is related to the competition between the molecule-molecule interactions and the molecule-substrate interactions. For small size clusters, the pentacene molecules are found to prefer to form a lateral-oriented cluster as driven by the molecule-substrate interactions, which can be increased by the existence of the sylanol groups on the amorphous SiO 2 surface. Furthermore, when the molecular number n is greater than n c , the normal orientation in the herringbone structure is gradually formed and becomes the stable configuration because the molecule-molecule interactions is dominant. [ABSTRACT FROM AUTHOR]

To better understand the coalescence of droplets, which play critical roles in diverse natural processes and industrial applications, we give attention to the non-Newtonian rheology of liquid drops—in particular, studying the coalescence of two non-Newtonian droplets on a solid surface, with special attention to the effect of the shear thinning behavior. Based on a theoretical power-law model, we show that the height h 0 of the liquid bridge connecting two adjacent droplets grows with a power function of time as h 0 ∼ t n , where n indicates the power-law exponent. Through numerical simulations, we reveal a self-similar regime during the initial stage of coalescence and propose an accurate prediction for capturing the spatial structure of the flow. Our results also update the effect of the contact angle, which significantly alters the coalescence dynamics. [ABSTRACT FROM AUTHOR]

Due to variations in light transmission and wear on the contact head, existing visual-tactile dataset building methods typically require a large amount of real-world data, making the dataset building process time-consuming and labor-intensive. Sim-to-Real learning has been proposed to realize Multi-Source Visual-Tactile Information Understanding (MSVTIU) in simulate and real environment, which can efficiently promote visual-tactile dataset building using simulation method for emerged robotic applications. However, the existing Sim-to-Real learning also requires more than 10,000 real data, while the corresponding data need to be re-collected when the sensor version changes. To address this challenge, we propose a powerful Sim-to-Real transfer for MSVTIU which requires only one single real-world tactile sample. To effectively extract features from the single real tactile sample, a multi-scale vision transformers-based Generative Adversarial Network (GAN) is proposed to address MSVTIU task under extremely limited data. We introduce a novel scale-dependent self-attention mechanism that allows attention layers to adapt their behavior at different stages of the generating process. In addition, we introduced a residual block for capturing contextual information between adjacent scales, which utilizes shortcut connections to fully preserve texture and structure information. We subsequently enhanced the model understanding of visual-tactile information using elastic transform and adaptive adversarial training strategy, both of which are designed specifically for MSVTIU. Experiments on two public datasets with diverse objects indicate that our Sim-to-Real transfer approach utilizing only one single real-world visual-tactile sample outperforms the state-of-the-art methods that requires tens of thousands of samples. • Single sample-driven and new state-of-the-art method. • Our method outperforms the others requiring 10000+ samples. • A scale-dependent self-attention mechanism to extract features efficiently. • An adaptive residual block to better capturing contextual features. [ABSTRACT FROM AUTHOR]

• The revelation of the evaporation and solidification mechanism of microdroplets. • Modeling of evaporative microdroplets across scales. • Propose a modified set of boundary conditions to control the evaporation rate. • Optimize the cured morphology using vacuum drying system. Inkjet printing is already regarded as a transformative technique for producing organic light-emitting diodes (OLED) due to its low cost, high material utilization, and short procedure. However, the application of printed OLED has been severely limited by the appearance fault of cured microdroplets. The process of microdroplets evaporation is complicated and hard to monitor experimentally since it spans the nano and micro scales. In this work, a hybrid 3D lattice Boltzmann (LB) model coupled with the Langevin equation model is used to study the evaporation and solidification mechanism of microdroplets in the pit. During the microdroplets evaporation process, the nanoparticles (NPs) are primarily affected by the internal field of microdroplets and the falling liquid level. The internal field will bring the NPs inside the microdroplets to the edge of pit, while the falling liquid level causes the NPs to fall vertically. Based on this law, we propose a modified set of boundary conditions to control the evaporation rate. The slower the evaporation rate, the more NPs gather around the edge. The maximum particle uniformity of 68.27 % is achieved with an evaporation coefficient of 0.85. Ultimately, we successfully achieved a uniform and flat morphology of the cured microdroplets by making the microdroplets evaporate at varying speeds. The mechanism of the evaporation and the solidification of microdroplets in pits are revealed, which is of great significance to the application of printed OLED technology. [ABSTRACT FROM AUTHOR]

Abstract: The effect of the different hygrothermal conditions on the shear strength degradation of ACA joints and the difference among them are studied in this paper using experiments and numerical simulations. First of all, several hygrothermal aging tests with different accelerating levels are carried out to explore the hygrothermal environment induced degradation paths of interfacial strength of ACA joints. It is found that the shear strength degradation paths differ from each other when the ACA joints are tested under different hygrothermal conditions, while they are similar to each other. For each case, the shear strength firstly decreases rapidly and then slowly approximates to a fixed value, and the shear strength difference between each other goes to an approximate constant after certain early test hours. After that, to understand the degradation mechanism better, finite element modeling is achieved to study the relationship of moisture concentration in ACA layer and shear strength of the ACA joints, and it is found that moisture saturation and diffusion rate are considered to be the main causes of the shear strength degradation. [Copyright &y& Elsevier]

Abstract: Electrospinning is an effective method for nanofiber production, but seldom used in the fabrication of patterned structures directly due to the whipping instability of the electrospinning jet. The whipping instability of electrospinning is adopted to fabricate stretchable patterned film by combination with an improved thermal transfer printing. The electrospun film is composed of small-scale wavy/coiled fibers, which make the patterned film highly stretchable. The optimal process parameters of whipping-based electrospinning are investigated to fabricate electrospun film with uniform and compact wavy/coiled fiber. Then the transfer printing and thermal detachment lithography are studied to generate patterned film, including the pressure, temperature, and peeling-off speed. Finally, the stretchability of the patterned electrospun film is studied through experiment and finite element analysis. It may open a cost-effective and high-throughput way for flexible/stretchable electronics fabrication. [Copyright &y& Elsevier]

The accurate measurement of fluid flows with large velocity gradients and a high velocity range, such as vortex flow and jet flow, is still paramount and challenging in fluid dynamics research. Particle image velocimetry (PIV) is a well-established experimental technique for determining fluid velocities; however, it cannot obtain an accurate result when the displacements of tracer particles are much larger than their sizes (Liu in J Fluids Eng 142(5):054051, 2020). In this study, we propose a novel hybrid optical flow formulation that consists of two parts, namely a circulant-matrix-based correlation and an adaptive weight optical flow, to solve this problem. First, a novel correlation method based on a circulant matrix is proposed, where the interrogation window is cyclically sampled to improve the measurable velocity range and measurement accuracy. Subsequently, we design a specific optical flow whose regularization parameters can be adjusted with changes in the velocity gradient and introduce a non-quadratic penalty function to achieve better stability and convergence. The solution provided by the circulant-matrix-based correlation is used as an initialization to account for large displacements with small particles and corrects the following adaptive weight optical flow in real time to accurately capture small vortex and turbulent structures. Velocity fields are estimated over synthetic and experimental particle images, and the velocity results are compared with advanced PIV methods. The results and comparisons show that the proposed method successfully achieves good performance in capturing small-scale vortices and turbulent structures when measuring fluid flows with large velocity gradients and a high velocity range. [ABSTRACT FROM AUTHOR]

The IC chip ejecting and pick-up process in electronic assembly plays a critical role in advanced packaging since the success ratio and productivity are determined by the delamination of thin chips from the underlying substrate. For a chip-on-substrate structure subjected to a transverse concentrated load resulting from the ejecting needle, a competing relationship between chip peeling-off and chip cracking is discovered from the perspective of integral structure. A competing index, \Gamma, is defined to characterize the competing fracture behavior between cracking and peeling-off, and the competing fracture mechanism is then uncovered fundamentally. Based on the index, the effects of chip geometry on competing behavior are investigated using the virtual crack-closure technique with dummy nodes. Critical thickness and length, which separate chip peeling-off and cracking modes, are found and they represent the extremity of successful pick-up process. The critical values are determined by critical competing index, \Gamma=\Gammac. The competing mechanism uncovered applies to general pick-up process with various chip sizes and materials. The critical sizes and the method of evaluating critical values are meaningful for chip profile design, optimization of pick-up process, and propelling new designs of electronic assembly process, especially for thin chips. [ABSTRACT FROM PUBLISHER]

Abstract: This paper presents a new approach to predict the loads acting on the shield periphery during excavation. This approach not only represents the dynamic variation of loads, but also can be used to calculate the rectification loads for moving shield backward to the planned alignment. Firstly, the ground reaction curves are adopted to calculate the earth pressure. Particularly, the ground displacement is considered to be induced by two kinds of motion: tunneling-induced ground loss and snake-like motion of shield movement. Then, optimization criteria for load balance, minimum potential energy, shield behavior, and ground settlement have been considered. Finally, basic example is carried out in a straight alignment for the single layer of homogeneous ground. The validity is discussed by comparing the optimization model with numerical simulation, and the results confirm that the proposed approach is reasonable enough to predict the distributed loads acting on the shield periphery. [Copyright &y& Elsevier]

The roll-to-roll (R2R) processing plays an increasingly important role in the high-throughput fabrication of flexible electronics. This paper highlights the dramatic influence of material properties, geometrical dimensions of film-on-substrate structure, process temperatures, and web tension of the R2R processing. A co-optimization approach has been presented for the R2R processing, and all the design and process parameters are to be simultaneously optimized to reduce the interfacial residual stress. First, the misfit strain between the thin film and the substrate is established based on the process parameters. The temperature-dependent properties are involved in polymer/rubber substrate. Nonlinear phenomena are discovered when the process temperatures and the web tension are involved in substrate together. Then, the interfacial shear and peeling stresses resulted from R2R processing are derived in analytical model based on the generalized plane strain theory. The interfacial stresses are related with process temperatures, material properties, and structural dimensions. Finally, a multivariable optimization model is established to compensate the misfit strain. The results imply that device structures and process parameters are correlated phenomena and therefore should be simultaneously optimized in the R2R processing of flexible electronics. [ABSTRACT FROM PUBLISHER]

This article proposes a standalone radio frequency identification (RFID)-based mobile robot navigation method, in which a mobile robot equipped with reader antennas can be continuously guided to a static object marked with a single passive UHF RFID tag. An observation model based on the RFID phase difference is built and integrated into a particle filter, by which the instantaneous relative position between the mobile robot and the tagged object can be detected in real time. Based on the position information extracted from the RFID system, the mobile robot adjusts its pose to move toward the RFID-tagged object. Compared with the existing RFID-based mobile robot navigation methods, the proposed method requires no external sensors other than the RFID and requires only a single passive tag. Experiments using commercial off-the-shelf (COTS) RFID devices are performed, and the results indicate that the mobile robot can satisfactorily realize navigation task with a distance accuracy of 4.04 cm and a bearing accuracy of 2.23°. The proposed method is well applicable for the navigation scenes in which the absolute position of the tagged target object is not known beforehand. Note to Practitioners—UHF radio frequency identification (RFID) has been widely applied as an asset management ID sensor in many fields. RFID-based mobile robot navigation technology can further increase its application value as a location sensor. This article proposes a standalone RFID-based mobile robot navigation method, in which the reference tag and the external sensors other than RFID are both not required. In the proposed method, only a single passive tag is attached to the static target object. Experimental results indicate that the proposed method can enable navigation task with good performance in situations in which the absolute navigation goal position is not known in advance. [ABSTRACT FROM AUTHOR]

We investigate multi-mode oscillation and ejection behavior in pulsed electrohydrodynamic (EHD) printing in this paper. The results show that multi-mode oscillation is determined by the damping ratio of the meniscus, while the phase difference between meniscus oscillation and the electric field governs whether jetting can occur. High-order-mode jetting can occur around the corresponding resonant frequency of a low damping ratio system. This enhances the frequency limit to approximately 2.62 times that of traditional pulsed EHD printing, providing an efficient approach to high-frequency EHD printing. [ABSTRACT FROM AUTHOR]

Direct full‐color image photodetectors without dichroic prisms or sophisticated color filters have considerable advantages in target recognition and information acquisition for electronic eyes and wearable sensors. However, the ability to combine various multispectral semiconductors in a high‐resolution and cost‐effective manner is still challenging. Here, high‐resolution electrohydrodynamic (EHD) printing, together with ionic liquid methylammonium acetate (MAAc) as the solvent, is first introduced to directly integrate various spectral‐response perovskite films into a pixelized full‐color photodetector. EHD printing enables micro/nanopatterning by using high electrical force to induce jetting, and MAAc improves film quality with scant pinholes and large‐size grains by decreasing the perovskite growth rate. By optimizing the printing process and crystallization condition, 1 µm perovskite dot arrays are EHD printed; this is, to the best of knowledge, the smallest printed feature size of perovskite application. And the photodetector still achieves high R and D* values of 14.97 A W‐1 and 1.41 × 1012 Jones, respectively. Finally, an integrated flexible full‐color image photodetector is constructed, which successfully realizes light signal detection and color recognition, paving a versatile and competitive approach for future full‐color image sensors and artificial vision systems. [ABSTRACT FROM AUTHOR]

Most face recognition methods employ single-bit binary descriptors for face representation. The information from these methods is lost in the process of quantization from real-valued descriptors to binary descriptors, which greatly limits their robustness for face recognition. In this study, we propose a novel weighted feature histogram (WFH) method of multi-scale local patches using multi-bit binary descriptors for face recognition. First, to obtain multi-scale information of the face image, the local patches are extracted using a multi-scale local patch generation (MSLPG) method. Second, with the goal of reducing the quantization information loss of binary descriptors, a novel multi-bit local binary descriptor learning (MBLBDL) method is proposed to extract multi-bit local binary descriptors (MBLBDs). In MBLBDL, a learned mapping matrix and novel multi-bit coding rules are employed to project pixel difference vectors (PDVs) into the MBLBDs in each local patch. Finally, a novel robust weight learning (RWL) method is proposed to learn a set of robust weights for each patch to integrate the MBLBDs into the final face representation. In RWL, a codebook is first constructed by clustering MBLBDs on each local patch to extract a feature histogram. Then, considering that different parts of the face have different degrees of robustness to local changes, a set of weights is learned to concatenate the feature histograms of all local patches into the final representation of a face image. In addition, to further improve the performance for heterogeneous face recognition, a coupled WFH (C-WFH) method is proposed. C-WFH maintains the similarity of the corresponding MBLBDs and feature histograms for a pair of heterogeneous face images by means of a novel coupled feature learning (CFL) method to reduce the modality gap. A series of experiments are conducted on widely used face datasets to analyze the performance of WFH and C-WFH. Extensive experimental results show that WFH and C-WFH outperform state-of-the-art face recognition methods. [ABSTRACT FROM AUTHOR]

Large-scale complex surfaces are the core components of large aircraft, ships, and wind turbine generators. High-accuracy manufacturing of the components is very crucial for its performance. Automated and high-accuracy large-scale 3-D measurement is an effective means for the manufacturing accuracy control of these components. This article proposes an automated mobile robotic profile measurement method for large-scale complex surfaces, and it also proposes a high-accuracy point clouds alignment algorithm to improve the measurement accuracy. A fringe projection scanner, integrated on the end of a mobile robot, is utilized to scan the surface and obtain the local scanning point clouds. The local scanning point clouds’ coarse alignment is implemented using the robot’s joint data and the laser sensor’s positioning data. After that, a fine alignment algorithm combining the fruit fly optimization algorithm (FOA) with the improved iterative closest point (IICP) algorithm is proposed for point clouds’ fine alignment. Practical measurement was performed on a sectional wind turbine blade, and a broad region of $4.6 \text { m} \times 2.6$ m was measured. An accuracy evaluation method based on point-to-plane distance was adopted to evaluate the measurement error. The result that the mean distance error is less than 0.4 mm shows the proposed mobile robotic measurement system and the aligning method are of high accuracy. [ABSTRACT FROM AUTHOR]

Dynamic object following is an essential ability for mobile robots in various application scenarios such as auxiliary cargo transportation, disabled assistant caring, and multirobot formation. A dynamic object following method for a mobile robot based on ultrahigh-frequency (UHF) radio frequency identification (RFID), which uses a phase difference information-based dynamic following model, is presented in this article. Through this method, a mobile robot equipped with an RFID reader can follow a dynamic target object attached with a passive RFID tag. A cost function is constructed to indicate the changes in relative distance and bearing between the target object and the following robot, allowing the real-time calculation of the translational and rotational velocities of the following robot using an algorithm base on proportional integral differential (PID)-based controllers. Compared with the existing RFID-based following methods, the proposed method can achieve a satisfactory distance and bearing following accuracy by using RFID phase difference information without any tedious and time-consuming phase drift or antenna installation calibration. Real experiments were conducted using commercial off-the-shelf (COTS) UHF RFID, and results show that the proposed method is effective and robust, the respective average distance and bearing following errors were approximately 1.65 cm and 4.94° during the robot-following tests and 4.45 cm and 6.03° during the human-following tests. The proposed method has promising applications in mobile robot following of dynamic object to provide service or assistance in large storage and manufacturing industries. [ABSTRACT FROM AUTHOR]

Active learning can reduce the human effort required for labeling training samples while preserving the performance of visual classifiers. However, existing active learning frameworks cannot be used to perform visual classification of industrial product surface defects because they still require intensive manual annotation efforts. In this study, we propose a cost-efficient autolabeling-enhanced active learning (ALEAL) framework to reduce the human annotation effort required for surface defect visual classification. The proposed ALEAL framework employs a deep convolutional neural network (CNN) as a visual classifier trained from an initial set of human-labeled training samples. Then, the collected unlabeled training samples are input into the classifier for category confidence estimation. Next, a novel diverse cost-effective query strategy (DCEQS) is proposed to select some high-confidence samples for autolabeling and some informative samples for sample proposals that need labeling. Subsequently, to further reduce the human annotation effort, a novel autolabeling module is proposed and introduced in ALEAL that can automatically label a portion of the informative unlabeled training samples selected by the DCEQS. In this study, a novel attention-based similarity measurement network (ASMN) is proposed as an implementation of this autolabeling module by measuring the similarity between unlabeled and labeled samples. Finally, the remaining unlabeled samples are annotated by human experts, and all the newly labeled samples are used to retrain the classifier. Through the autolabeling process from the DCEQS and ASMN, ALEAL can automatically label additional training samples and achieve a competitive performance while requiring few human-labeled training samples, which is highly important in industrial applications. Extensive experimental results show that, compared with popular active learning methods, ALEAL can dramatically reduce the effort involved in human annotation and achieve state-of-the-art cost efficiency for the visual classification of industrial product surface defects. [ABSTRACT FROM AUTHOR]