Search

Unmanned Aerial Vehicle (UAV) inspection has gradually replaced manual inspection of transmission tower, which produces many images. While it is laborious and time-consuming to manually analyze these images, there are also challenges in automatically detecting small-size defective components such as bolts in transmission tower images, due to problems including complex background, small size, and many similar objects of bolts. In this paper, by virtue of multi-scale features and context information, we propose a deep neural network named Camp-Net (Context Information and Multi-Scale Pyramid Network) to identify bolts defect in transmission tower images. First, multi-scale feature fusion combines deep features and shallow features in convolutional networks to detect small-size bolts. Second, context information fusion puts the information around bolts into the detection network to remove the disturbance of complex background and similar objects. An image dataset containing defective bolts and normal bolts is constructed for model training and testing. Experimental results show that bolts with loose pins and bolts without pins among fittings in transmission tower can be accurately identified with the proposed model. The Average Precision (AP) of defective bolts detection of this model can be 11.4% higher than that of the commonly used high performance model, Faster R-CNN. [ABSTRACT FROM AUTHOR]

Threaded fasteners are prevalent throughout modern manufacturing. Thus, as the demand for automation in manufacturing increases, so does the demand for automated threaded fastening systems. However, many fundamental issues and engineering challenges still hinder robustness in automation, particularly for smaller screws and critical product finishing requirements. This paper surveys the state of the art in threaded fastening automation and discusses open questions for further research. This survey covers the following areas: 1) fundamentals of threaded fastening, including basic concepts and definitions; 2) analysis of the entire assembly process (consisting of part feeding and orientation, pickup, alignment, and driving), including discussions of tools, control strategies, and other considerations; 3) failure modes and techniques to mitigate them; 4) threaded fastening systems and electromechanical approaches; and 5) open challenges and suggestions for future development. Understanding the current state of automation in threaded fastening will provide a foundation for researchers to advance this field. Note to Practitioners—This paper is motivated by the problem of an automated assembly of small screws, one of the most challenging problems in a smartphone assembly. It represents a rigorous review of robotic screwdriving literature to identify the state-of-the-art and open problems. The review material was targeted toward engineers working on related problems with the sponsor and has proved useful to them. To benefit researchers in the field of robotic and automated assembly, we have compiled the review material in the form of a survey paper. This paper covers theoretical fundamentals, tools, control and failure detection strategies, industrial applications, and open problems for robotic screwdriving. It provides a foundation for readers to familiarize themselves with the state of the art and conduct further research on this thread. [ABSTRACT FROM AUTHOR]

This paper investigates an approach that uses the cognitive architecture Soar to improve the performance of an automated robotic system, which uses a combination of vision and force sensing to remove screws from laptop cases. Soar’s long-term memory module, semantic memory, was used to remember pieces of information regarding laptop models and screw holes. The system was trained with multiple laptop models and the method in which Soar was used to facilitate the removal of screws was varied to determine the best performance of the system. In all the cases, Soar could determine the correct laptop model and in what orientation it was placed in the system. Soar was also used to remember what circle locations that were explored contained screws and what circles did not. Remembering the locations of the holes decreased a trial time by over 60%. The system performed the best when the number of training trials used to explore circle locations was limited, as this decreased the total trial time by over 10% for most of the laptop models and orientations. Note to Practitioners—Although the amount of discarded electronic waste in the world is rapidly increasing, efficient methods that can handle this in an automated non-destructive fashion have not been developed. Screws are a common fastener used on electronic products, such as laptops, and must be removed during nondestructive methods. In this paper, we focus on using the cognitive architecture Soar to facilitate the disassembly sequence of removing these screws from the back of laptops. Soar is able to differentiate between different models of laptops and store the locations of screws for these models leading to an improvement of the disassembly time when the same laptop model is used. Currently, this paper only uses one of Soar’s long-term memory modules (semantic memory) and a screwdriver tool. However, this paper can be extended to use multiple tools by using different features available in Soar such as other long-term memory modules and substates. [ABSTRACT FROM AUTHOR]

Modeling of the structural behavior of the housing of an electromagnetic railgun is usually carried out without considering the electromagnetic origin of the forces; the magnetic pressure acting on the rails is assumed to act at the rail surfaces only. This approximation proves to be a great challenge, as moving pressure profiles at speeds of a railgun projectile are very rare in purely mechanical investigations. In the past, we have published several papers on the structural mechanics of railgun housings characterized by discrete supports. We could show that the displacement of rail surfaces at the position of the armature can reach amplitudes that are not negligible. In fact, in the case of metal brush armatures, the increase in distance between the rail surfaces of only tenths of a millimeter can make the difference between the solid contact and arcing to increase. We also matched our findings to experimental data that were, however, so far only gained in static tests. In this paper, we analyze the foundations of our mechanical modeling using electromagnetic calculations. We calculate a 3-D distribution of the $J \times B$ volume force density in the rails and use them as input for our structural mechanics model. A second point to be addressed concerns the parts made from conducting materials, such as our discrete supports made from steel bolts, which are situated in rapidly changing magnetic fields during the experiments. Again, the influence on our structural mechanics modeling is discussed. [ABSTRACT FROM AUTHOR]

This paper investigates the performance of a magnetic lead screw (MLS) equipped with skewed arc magnets. Skewed arc magnets are adopted to approximate the helical ones, because the latter is hard to realize. Although similar method had been proposed in the previous literature, discussion about the parameters of the skewed arc magnets was insufficient and little attention was paid to the gear ratio of the resultant MLS. In this paper, the principle of approximation by skewed arc magnets is proposed and discussed, and the performance of the MLS with skewed arc magnets is evaluated and compared with the ideal MLS in terms of thrust force, torque, and gear ratio. It is highlighted that the gear ratio is the fundamental indicator of the MLS’s performance, besides the thrust force and torque. It is shown that the gear ratio of MLS is mainly affected by the reference angle and the arc angle of skewed arc magnets, and optimal values for the two angles are obtained. In addition, the effects of the magnet magnetization on thrust force and torque are analyzed. Finally, dovetail budge and groove are introduced between the adjacent magnet segments for simpler fabrication. [ABSTRACT FROM AUTHOR]

This paper introduces a batch fabrication method to manufacture Micro-Electro-Mechanical System (MEMS) platforms for 3D integration of sensors spatially-distributed on a 2D plane. In the heart of the concept is a foldable MEMS structure with polymer hinges which is used to support and guide the assembly of the discrete planar sensors by means of folding them into a 3D shape, like origami, providing controlled distribution of sensors in space. Flexible hinges carrying the electrical interconnects are a critical structural element of the platform and material selection study for those is the main focus of this paper. We analyzed different materials for flexible hinges fabrication, including photo-definable polyimide and parylene-C. Three approaches for sensor integration are presented: 1) co-fabrication; 2) sensor drop-in; and 3) transfer bonding. The prototypes of structures with different flexible hinge materials were fabricated and evaluated based on their mechanical flexibility, chemical compatibility, and material outgassing. Parylene-C exhibited similar or better performances compared to polyimide, demonstrating in each of the experiments a higher degree of thermal flexibility up to 350 °C, a superior chemical resistance against hydrofluoric acid, and 2.9 times lower outgassing. [2020-0301] [ABSTRACT FROM AUTHOR]

Hyperspectral imagery (HSI) has shown promising results in real-world applications. However, the technological evolution of optical sensors poses two main challenges in HSI classification: 1) the spectral band is usually redundant and noisy and 2) HSI with millions of pixels has become increasingly common in real-world applications. Motivated by the recent success of hybrid huberized support vector machines (HHSVMs), which inherit the benefits of both lasso and ridge regression, this paper first investigates the advantages of HHSVM for HSI applications. Unfortunately, the existing HHSVM solvers suffer from prohibitive computational costs on large-scale data sets. To solve this problem, this paper proposes simple and effective stochastic HHSVM algorithms for HSI classification. In the stochastic settings, we show that with a probability of at least $1-\varrho $ , our algorithms find an $\epsilon $ -accurate solution using $\tilde {O}({1}/{\lambda _{2}\epsilon })$ iterations. Since the convergence rate of our algorithms does not depend on the size of the training set, our algorithms are suitable for handling large-scale problems. We demonstrate the superiority of our algorithms by conducting experiments on large-scale binary and multiclass classification problems, comparing to the state-of-the-art HHSVM solvers. Finally, we apply our algorithms to real HSI classification and achieve promising results. [ABSTRACT FROM AUTHOR]

In power electronic applications, due to switching components, the inverter dc-bus generates current pulses at low frequencies and high frequencies (HFs). Usually, a dc-bus capacitor is sized to absorb the current peaks at the switching frequency and to stabilize the dc voltage. For decoupling the HF harmonics, an additional capacitor is required. This decoupling capacitor needs to have a low impedance over a wide range of frequencies. However, their performances for the HF range are highly limited by their stray inductance. In order to minimize the need for an additional decoupling capacitor at HFs, this paper proposes to improve the connections of the dc bus capacitor. Busbar configuration and printed circuit board allow reducing parasitics of the cabling. Several types of interconnections are proposed and characterized in this paper. Both two-terminal and four-terminal capacitors are evaluated and compared regarding operating frequency ranges. The measurement and modeling results show a significant improvement for HF decoupling by using quadrupole coaxial connections. [ABSTRACT FROM AUTHOR]

This paper deals with equivalent 2-D finite-element (FE) models of permanent magnet lead screw (PMLS), to accurately predict the thrust force. Due to the helical-shape permanent magnets (PMs), the magnetic fields of PMLS are typically 3-D and non-symmetric. However, the PMLS is approximately symmetric in some cases and an equivalent 2-D axis-symmetric FE model is developed instead of 3-D models. The problem lies in that no evaluation about the accuracy and applicability of the 2-D axis-symmetric FE model was offered in previous literature. In this paper, the 2-D axis-symmetric FE model is derived based on the concept of equivalent current sheets of PMs. The limitations of the existing 2-D axis-symmetric FE model are clarified, and a novel equivalent 2-D torque FE model is then proposed to predict thrust force with the gear ratio when the 2-D axis-symmetric FE model is not appropriate. It is shown that the choice of equivalent 2-D FE model depends on the lead angle of PMLS, 2-D axis-symmetric FE model is applicable in the range of (0°, 15°), and 2-D torque FE model is preferred in the range of (65°, 90°). [ABSTRACT FROM AUTHOR]

With the remarkable recent progress on learning deep generative models, it becomes increasingly interesting to develop models for controllable image synthesis from reconfigurable structured inputs. This paper focuses on a recently emerged task, layout-to-image, whose goal is to learn generative models for synthesizing photo-realistic images from a spatial layout (i.e., object bounding boxes configured in an image lattice) and its style codes (i.e., structural and appearance variations encoded by latent vectors). This paper first proposes an intuitive paradigm for the task, layout-to-mask-to-image, which learns to unfold object masks in a weakly-supervised way based on an input layout and object style codes. The layout-to-mask component deeply interacts with layers in the generator network to bridge the gap between an input layout and synthesized images. Then, this paper presents a method built on Generative Adversarial Networks (GANs) for the proposed layout-to-mask-to-image synthesis with layout and style control at both image and object levels. The controllability is realized by a proposed novel Instance-Sensitive and Layout-Aware Normalization (ISLA-Norm) scheme. A layout semi-supervised version of the proposed method is further developed without sacrificing performance. In experiments, the proposed method is tested in the COCO-Stuff dataset and the Visual Genome dataset with state-of-the-art performance obtained. [ABSTRACT FROM AUTHOR]

In order to break through low sensitivity limitation of the existing fiber Bragg grating (FBG) acceleration sensor, a new high sensitivity FBG acceleration sensor based on rigid hinge is proposed. The working principle innovation is that the sensor uses the rigid body rotation motion instead of the existing elastic deformation. The novel structure is that the pendulum rotates around the rigid hinge and the structure is symmetrically designed, which increases the elongation of the fiber grating by twice. The resonance frequency and sensitivity formula of FBG acceleration sensor is theoretically obtained by analysis. The influence of size parameters of the structure on sensor sensitivity and resonance frequency is further analyzed by MATLAB, and finally the structural parameters are optimized. On the basis, the first-order natural frequency of the sensor is simulated by ANSYS Workbench17.0. Experimental studies show that the natural frequency of the acceleration sensor designed in this paper is 39 Hz, the sensitivity of the sensor is 949.2 pm/g at 15 Hz, and the cross-interference is less than 6.49%, which is suitable for precision measurement under low frequency micro-vibration. [ABSTRACT FROM AUTHOR]

Fasteners are an important part of railway security. The application of computer vision and image processing on fastener detection has become a research hot spot. However, most of the researches have focused on the missing and damage detection of fasteners up to now. In this paper, a method based on image processing is proposed to locate the center of the hexagon nut in the railway fastener. The precise position of the hexagon nut can be used for the automatic replacement of the nut, which is of great significance for the automatic railway construction and maintenance. The locating method proposed in this paper can be divided into two steps. First, the improved Canny detector which combines surrounding suppression and iterative excitation is used to extract the contours of the fastener. Then, a multifeature hierarchical locating (MHL) algorithm is proposed, which integrates the circle contour of the gasket and the hexagon features of the nut. The experimental results show that the improved Canny detector suppresses noises effectively and highlights the contours of fastener. The MHL algorithm in the next step improves the locating accuracy remarkably while guaranteeing the robustness meantime. [ABSTRACT FROM AUTHOR]

The objective of this paper is to advance the state of the art of aggressive space mapping (ASM) by demonstrating how, for resonant structures, and microwave filters in particular, an ASM-based optimization procedure may converge to the desired target performance in just one step. This behavior is first justified in physical terms, using the electromagnetic cavity perturbation theory, and is then investigated numerically. Several practical filter design demonstrations are also presented. This paper is concluded with a conceptual definition of Space Alignment in the context of ASM. [ABSTRACT FROM AUTHOR]

This paper presents a novel structure-controlled variable stiffness robotic gripper that enables adaptive gripping of soft and rigid objects with a wide range of compliance. With the structure-controllable principle, the stiffness is controlled by the mechanical structure configurations rather than by material properties or electronic means. The principle is realized by changing the effective second moment of area of the gripper finger through rotating a built-in flexure hinge shaft. Based on this principle, the states of the stiffness can be continuously, instead of discretely, studied and assessed over the intermediate states from compliant to almost completely rigid. A variable stiffness mechanism has been developed to demonstrate the validity of the proposed principle. It enables that the finger stiffness and gripping position are independently controlled. With the introduction of flexure hinges, the undesired lateral buckling resulted from the rotation of a normal leaf spring is eliminated. In addition, a two-finger parallel gripper with this variable stiffness mechanism is developed which can provide the grasping stiffness according to the grasping task requirements. The effectiveness of the gripper has been demonstrated to handle the objects range from light, fragile to heavy, rigid without using any feedback loop or soft pads. Note to Practitioners—This work was inspired by the fact that the capability of variable stiffness in the robot actuator can reduce the undesired impacts to the robot arms. It can also allow a safer interaction between robot and human. Gripping of objects with uncertainties in shapes and position, as well as large variations in fragility and weight can also benefit from the concept of the variable stiffness. However, existing designs of variable stiffness grippers either have limited stiffness range or bulky configurations. They compromised the practical applications. This paper introduces a design in that a rotating flexure hinge shaft is embedded inside the robotic gripper finger. The mechanical stiffness of such fingers can be varied by changing the rotation angle of the flexure hinges. We present the working principle supported by mathematical models in the design and development. We also show an example design of a two-finger parallel gripper equipped with the VSFs. Extensive experiments demonstrated that the gripper is effective in gripping objects with wide range of uncertainties. Such gripper design avoids the use of soft pads as well as closed-loop control and high-precision sensors. In the future work, we shall implement such grippers for actual industrial applications. [ABSTRACT FROM AUTHOR]

In this paper, a real-time railway fastener detection system using a high-speed laser range finder camera is presented. First, an extensive analysis of various methods based on pixel-wise and histogram similarities are conducted on a specific railway route. Then, a fusing stage is introduced which combines least correlated approaches also considering the performance upgrade after fusing. Then, the resulting method is tested on a larger database collected from a different railway route. After observing repeated successes, the method is implemented on NI LabVIEW and run real-time with a high-speed 3-D camera placed under a railway carriage designed for railway quality inspection. [ABSTRACT FROM PUBLISHER]

Emerging applications of robots with direct physical interactions with humans have led to the development of a variety of series elastic actuators (SEAs) which are compliant, force controllable, and back drivable. The performance of current SEAs is mainly dependent on the specific stiffness of the spring. In our previous work, a compliant actuator with two-stage stiffness has been designed to overcome the performance limitations in current SEAs. The key novelty is that a low-stiffness spring and a high-stiffness spring are employed instead of a single spring in current SEAs, which has the advantages of high fidelity, low output impedance, and also large force range and bandwidth. In this paper, a tracking control scheme is proposed for the compliant actuator with two-stage stiffness. Although the overall stiffness is discontinuous, the proposed controller is continuous by integrating different control modes for two springs into a single one. The transition between control modes is smooth and embedded inside the controller, and it is also automatically realized by monitoring the output force of the actuator. The stability and convergence of the closed-loop system are analyzed, and experimental results are presented to demonstrate the effectiveness of the proposed control scheme.

Note to Practitioners—An SEA is developed by placing an elastic element into the actuator; this elasticity gives SEAs several unique properties including low mechanical output impedance, tolerance to impact loads, and passive mechanical energy storage, which makes it suitable for human–robot interaction. The performance of existing SEAs is highly dependent on the stiffness of a single spring. To overcome the limitations, a novel SEA with two-stage stiffness was proposed in our previous work. This paper suggests a continuous tracking control method for the proposed compliant actuator. Although the overall stiffness is discontinuous, the transition between different control modes for two springs is smooth and automatically realized. Experimental results show that the output force of the actuator is bounded. In future research, uncertainties in actuator dynamics will be considered, such that system identification or calibration is not required. [ABSTRACT FROM AUTHOR]

This paper is an investigation in the performance of two types of magnetic screws (MSs). The first is the magnet to magnet and consists of permanent magnet (PM) nut and PM screw and the second is a magnet to reluctance and consists of PM nut and a double-start reluctance screw. It is argued that for cost sensitive and long stoke applications, a magnet-to-reluctance MS may be the preferred option. It is also shown that capacitor discharge magnetization techniques can be employed for imprinting helical magnetization distributions on cylindrical PMs, significantly reducing the complexity and facilitating the practical realization of MS systems. [ABSTRACT FROM AUTHOR]

This paper presents a gain-enhanced circularly polarized (CP) patch antenna, where 3-D printing and computer numerical control machining technologies are employed for the fabrication of its dielectric and metal parts, respectively. In order to assemble the whole structure effectively and accurately, a hybrid strategy and pin-loaded example are here proposed to deal with the detached dielectric and metal. On the one hand, asymmetrically 3-D printed substrate with the mesh-grid architecture is formed with resin chunks to introduce the perturbation and produce the CP radiation. On the other hand, its metal part likewise possesses the 3-D structure with four metal screws to be symmetrically placed in the two diagonals of square patch, thus making the patch and ground as a whole. More importantly, the screws have dual functions, which can not only integrate the dielectric and metal parts, but also work as shorting pins. Therefore, due to the shunt inductive effect of shorting pins, the dominant mode of patch is excited at higher frequency, resulting in the enlarged antenna area and enhanced gain. As a result, the gain-enhanced CP patch antenna is realized, and an antenna prototype is then fabricated and tested, exhibiting a high CP gain of about 10 dBic. [ABSTRACT FROM AUTHOR]

This paper reports on mechanical design, optimization, and experimental testing of a novel piezo-actuated elliptical vibration-assisted cutting (EVC) system constructed by flexure hinges. The stroke and natural frequency were analyzed based on the theoretical modeling. An enhanced central composite design was chosen as the design of experiments methodology to reduce the modeling error, and a nondominated sorted genetic algorithm-II (NSGA-II) was adopted for structure optimization. The optimized EVC generator was manufactured and experimentally tested to investigate practical properties of the proposed EVC system. It shows that the stroke of input end can reach to 30 μ m with a motion resolution of 10 nm, and the first natural frequency can reach to 2600 Hz without considering the manufacturing error. Besides, a relatively small cross-axis coupling ratio (within 0.21%) can be effectively obtained. The developed EVC system is advantageous not only to being equipped with machine tools with various configurations, but also to easily achieving arbitrary vibrations in three-dimensional space through two actuators, which is especially important for the generation of complex structured surfaces. With this paper, it is of great significance to promote industrial application of EVC techniques. [ABSTRACT FROM AUTHOR]

The Hessian matrix, one of the most pivotal parts in a robotic system's acceleration model, is a prerequisite for dynamic modeling and control of parallel robots. Aimed at providing a straightforward and effective way for modeling, design, and control of parallel robots, a new approach to formulate the Hessian matrix using the finite and instantaneous screw theory is presented in this paper. The new approach is generalizable in that it can be used to formulate the Hessian matrix of any parallel robot in a simple and concise manner. It also provides a theoretical foundation for the dynamic control of industrial robotic systems with parallel structures. The new approach is verified through comparison with the traditional exponential matrix method. The Exechon parallel robot and a hybrid serial-parallel robot are used as examples to further detail the procedures of this approach. The acceleration model formulated in this paper is closely related with the topological and parametric models, resulting in a unified mathematical framework for modeling and control of parallel robots. [ABSTRACT FROM AUTHOR]

This paper presents a new type electromagnetic lead screw (EMLS) intended for wave energy application. Similar to the mechanical lead screw, this electromagnetic version can transfer slow linear motion to high-rotational motion, offering gearing effects. Compared with the existing pure magnetic lead screw (MLS) employing permanent magnets only, the new EMLS proposed uses dc current to provide the required helical-shape magnetic field, offering a much simpler, robust structure compared with the MLS. The working principle and the performances of this EMLS are analyzed in this paper. Comparison with PM-type lead screw is given. In addition to this, the theoretical analysis to support the use of a 2-D finite element model to analyze the 3-D magnetic-type lead screw is given, which has not been well covered in the existing literature. [ABSTRACT FROM AUTHOR]

This paper proposes a mechatronic suspension system with capabilities of energy regeneration and semi-active control [i.e., skyhook (SK)], incorporated into the mechanism. The system overcomes the trade-off between energy consumption and ride handling/comfort in a semi-active/active suspension system. Central to the concept is the development of a switched-mode rectifier (SMR) capable of providing either a positive or negative damping ratio by controlling the electric current to alternate between regenerative and motoring modes. Using the implemented circuit, a regenerative SK control strategy is utilized to provide a continuous variable damping force that significantly improves the vibration isolation. Simulation results are presented to demonstrate power flow in both regenerating and motoring modes by utilizing a current controller. Furthermore, experimental results are presented highlighting performance of the regenerative suspension prototype in terms of power regeneration, sprung mass absolute acceleration, and relative displacement. In addition, both theoretical and experimental power conversion efficiencies in various segments of the mechatronic suspension are analyzed and reported. [ABSTRACT FROM PUBLISHER]

A novel substrateless power amplifier module realized by ridge gap waveguide (RGW) technology is proposed in Ka-band. Up to our knowledge, this is the first time that an active device is directly integrated with an RGW structure. No substrate is used in the RF section of the implemented power amplifier and the whole structure is made by machining a piece of metal. The amplifier was made of two pieces and was easily assembled without needing a precise contact between the two sections. The cavity resonance of the amplifier is suppressed using a pining structure. The fabricated amplifier module is self-packaged and does not need any packaging. It is shown that the simulated and measured results are in good agreement. This paper shows the capability of RGW technology for implementing the whole RF frontend in microwave- and millimeter-wave systems. [ABSTRACT FROM PUBLISHER]

For the safety of railways, well-trained workers are required to check the fastener constantly, which shows the disadvantage of large time cost, huge labor cost and might being dangerous to workers. To address this and achieve automatic detection, an inspection model based on deep convolutional neural network (DCNN) is adopted in this paper. However, the inspection model suffering from the unbalanced training samples of defective vs normal due to defective fasteners are far less than normal fasteners in real railways. To tackle this problem, a novel sample generation method is proposed to generate defective fastener samples using the normal fasteners to realize sample augmentation. The comprehensive experiments are conducted on the collected real fastener samples and generated samples. The experimental results show that our method has good performance for fastener inspection on unbalanced samples and outperforms other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

This paper proposes a weighted large margin nearest center (WLMNC) distance-based human depth recovery method for tele-immersive video interaction systems with limited bandwidth consumption. In the remote stage, the proposed method highly compresses the depth data of the remote human into skeletal block structures by learning the WLMNC distance, which is equivalent to downsampling the human depth map at $64{\times}$ the sampling rate. In the local stage, the method first recovers a rough human depth map based on a WLMNC distance augmented clustering approach and then obtains a fine depth map based on a rough depth-guided autoregressive model to preserve the depth discontinuities and suppress texture copy artifacts. The proposed WLMNC distance is learned by the large margin clustering problem with a weighted hinge loss to balance the clustering accuracy and depth recovery accuracy and is verified to be able to preserve depth discontinuities between skeletal block structures with occlusion. A theoretical analysis is conducted to verify the effectiveness of using the weighted hinge loss. Furthermore, a novel data set containing various types of human postures with self-occlusion is built to benchmark the human depth recovery methods. The quantitative comparison with the state-of-the-art depth recovery methods on the introduced benchmark data set demonstrates the effectiveness of the proposed method for human depth recovery with such a high upsampling rate. [ABSTRACT FROM AUTHOR]

In this paper, we develop a stress wave-based active-sensing approach using piezoceramic transducers to detect the bolt looseness-induced damage in steel truss arch structures. A specimen of a steel tress arch structure was designed and fabricated. The structure consisted of top chords, bottom chords, Web members, and gusset plates, which were all connected by bolted connections. To implement the active sensing approach, the lead zirconate titanate (PZT) transducer bonded on the gusset plate was used as an actuator to generate stress waves, and the PZT transducers mounted on the other parts, such as top chords, bottom chords, and Web members, were used as sensors to detect the propagated stress waves. Based on the tightness of the bolts, the specimen had three different states: the healthy state, damage state I, and damage state II. The signals received by the PZT sensors were analyzed using the wavelet packet analysis. In addition, the structure stiffness was also considered as a comparative approach in this paper. The experimental results illustrate that when the initial looseness-induced damage occurs, the structure stiffness is not significantly reduced while the wavelet packet energy changes significantly, revealing the advantage of the proposed approach over the stiffness-based method. These research results demonstrated that the developed piezoceramic-based active-sensing approach has potentials to identify the initial bolt looseness occurrence for steel truss arch structures. [ABSTRACT FROM AUTHOR]

Controlled force–torque exertion by a small-scaled free-flying space robot will be potentially required for future on-orbit assembly missions. This paper presents a hybrid controller to achieve the desired end-effector motion and contact forces required for driving a screw into a floating target. Attitude of the spacecraft base is controlled simultaneously. Postcapture control to stabilize the entire system is also addressed in this paper. Simulation results have demonstrated the effectiveness of the proposed control laws. [ABSTRACT FROM AUTHOR]

The objective of this paper is to use electromagnetic-based computer-aided design (CAD) tools to investigate the maximum tuning range of channel filters, typically used in satellite payloads. Both circular and rectangular waveguide technologies are investigated. The results of the investigation show that single-mode rectangular waveguide implementations offer substantially wider tuning range, as opposed to classical dual-mode circular waveguide implementations. In addition to simulations, measurements are also presented indicating very good agreement with theory, thereby fully validating the CAD procedure. [ABSTRACT FROM PUBLISHER]

Aging aircraft is susceptible to cyclic fatigue cracks around ferrous fasteners in multilayer aluminum wing lap joints, as found on Canadian Military aircraft CP-140 Aurora and CC-130 Hercules. Current methods of crack detection utilize bolt hole eddy current, requiring fastener removal, a time consuming process that also has the potential to result in collateral damage. In this paper, pulsed eddy currents (PECs) are induced in the aluminum wing structure by a probe centered over a ferrous fastener. The response signals are analyzed using a modified principal components analysis (PCA) and cluster analysis to show separation of groups of PCA scores from fasteners with and without notches in the bore hole. The Mahalanobis distance (MD), a cluster analysis method, is used to obtain a relative distance measurement between scores associated with cracks from those without cracks. The MD evaluation provides confidence intervals for detection of cracks, which range in extent from fastener holes from 0.9 to 5.5 mm, and a means of potentially sizing larger (>3 mm) second layer cracks. In this paper, 100% of the cracks in the sample were detectable. The results demonstrate PEC as a viable technique for second layer crack detection with ferrous fasteners present. [ABSTRACT FROM PUBLISHER]

Ocean waves provide a promising and abundant renewable energy resource. One reason wave energy technology is still not mature enough for commercialization is the high unit cost of generated electricity. This needs to be improved by a combination of device and associated controller design. A multi-float and multi-mode-motion wave energy converter (M-WEC) enables much higher energy conversion compared with a single-float, single-mode WEC (S-WEC); however, the added complexity in dynamics of a M-WEC makes the corresponding controller design more challenging. While the majority of current WEC control research has been based on the control of S-WECs it has shown that control can significantly improve energy conversion. This paper aims to design a linear non-causal optimal controller for a M-WEC to demonstrate that this improvement also applies to more complex WEC systems. We choose a multi-body attenuator type M-WEC called M4 as a case study for which the desirable feature of predominantly linear dynamics has been demonstrated. This means that a linear controller can be designed based on a linear hydrodynamic model without introducing an intractable computational burden for real-time controller implementation. Numerical results show that the linear non-causal optimal controller can significantly improve the power capture of M4 over a broad range of peak spectral wave periods by $\text{40}{\%}$ to $\text{100}{\%}$. [ABSTRACT FROM AUTHOR]

The electro-mechanical brake booster (EMBB) is a kind of mechatronic actuator, which is developed to suit the brake assist requirement of electric vehicles. In this paper, we report on the design of an EMBB system consisting of a dc motor, a two-state reduction of a gear and ball screw, a servo body, and a reaction disk. Considering the inconvenience of installation and high price of the pedal force sensor, we translate the control problem of brake power assist control to position tracking control. Meanwhile, a nonlinear control method for position tracking is presented to solve the problem of power assist braking, which is formalized as three parts: the steady-state control, feed-forward control based on reference dynamics, and state-dependent feedback control. The benefit of the nonlinear control method is that it offers a concise control law and performs well in engineering implementations. In addition, a second-order filter was designed to do the signal processing and obtain a higher-order derivative. Finally, the bench tests based on rapid control prototyping environment were designed and implemented to verify the performance of the controller. Test results show that both the position tracking performance and response time of the EMBB system performed well. [ABSTRACT FROM AUTHOR]

The performance of distance-based classifiers heavily depends on the underlying distance metric, so it is valuable to learn a suitable metric from the data. To address the problem of multimodality, it is desirable to learn local metrics. In this short paper, we define a new intuitive distance with local metrics and influential regions, and subsequently propose a novel local metric learning algorithm called LMLIR for distance-based classification. Our key intuition is to partition the metric space into influential regions and a background region, and then regulate the effectiveness of each local metric to be within the related influential regions. We learn multiple local metrics and influential regions to reduce the empirical hinge loss, and regularize the parameters on the basis of a resultant learning bound. Encouraging experimental results are obtained from various public and popular data sets. [ABSTRACT FROM AUTHOR]

Piezo-driven precision positioning platforms with large stroke are greatly demanded in both scientific research and industrial applications. Although various principles have been proposed to design piezo-driven positioning platforms, the problem of backward motion commonly exists, greatly hindering their applications. To suppress the backward motion, in this paper, an idea by generating a forward frictional force to balance the reversed frictional force was proposed. Correspondingly, a specific arc-shape flexure hinge was designed to generate the forward frictional force via its elastic recovery. To verify the feasibility of the proposed idea, a positioning platform with alterable initial gap between contact surfaces is designed by the parasitic motion principle (PMP). By measuring the output displacement under various initial gaps, a critical initial gap was identified, at which the positioning platform could output stepping displacement without backward motion. Under this critical initial gap, effects of driving voltage and frequency on the backward motion are further investigated. In addition, the natural frequency and loading capacity of the positioning platform are tested. This paper confirms that using PMP, it is feasible to design piezo-driven positioning platforms with completely suppressed backward motion, which would enhance the applications of the PMP and PMP positioning platforms. [ABSTRACT FROM AUTHOR]

The diagnosis of freight train defects is essential for railway traffic safety. Recently, many freight train defect detection systems based on image processing were developed. However, most of them used hand-engineered features, whose performance was easily degraded by the low image quality and dynamic background. The deep learning method has the merit of extracting robust multi-scale information from images. In this paper, a hierarchical features-based instance detection (HID) model was developed for instance-level defect detections. A deep residual neural network model was firstly used to extract invariant and discriminative features. Then, acquired multi-scale features were processed by a region proposal neural network to generate a coarse defect region and make defect classification. Finally, instance-level predictions were made on the coarse defect region using acquired features. Experiments on the collected dataset that has six types of defects demonstrated the effectiveness of the HID. Additionally, comparisons with the state-of-the-art approaches further demonstrated its promising performance. We also verified its high performance in transfer learning. [ABSTRACT FROM AUTHOR]

Homodyne laser vibrometer plays an essential role in the precision vibration measurement, but its application is restricted when measuring subfringe vibration and targets with various reflectivities. In order to achieve the nanoscale vibration measurement and to adapt to various reflectivities, an enhanced homodyne laser vibrometer is presented in this paper. First, an active vibration mechanism (AVM) based on a flexure hinge stimulated by a piezoelectric transducer (PZT) is introduced into the reference arm, with which the parameters for nonlinear error correction can be predetermined, and thereby the measurement of vibration less than quarter wavelength is feasible. Second, an automatic gain module is employed to optimize the intensity of the interference signals, which significantly improves the adaptability to target reflectivity. The experimental results show that the laser vibrometer proposed is capable of measuring vibration with amplitude down to 2 nm and vibration with surface reflectivity down to 0.08%. [ABSTRACT FROM AUTHOR]

This paper addresses two fundamental problems: 1) learning discriminative model parameters and 2) avoiding over-fitting, which often occurs in regression-based classification tasks. We formulate these two problems in terms of relaxing both the strict binary label matrix and graph regularization term into more flexible forms so that the margins between different classes are enlarged as much as possible and the problem of over-fitting is avoided to some extent. This task is accomplished by the proposed double relaxed regression (DRR) method. The convex problem of DRR is solved efficiently with an iterative procedure. Extensive experiments on synthetic and real world image data sets demonstrate the effectiveness of the proposed method in terms of both classification accuracy and running time. [ABSTRACT FROM AUTHOR]

Support vector machines (SVM) have drawn wide attention for the last two decades due to its extensive applications, so a vast body of work has developed optimization algorithms to solve SVM with various soft-margin losses. To distinguish all, in this paper, we aim at solving an ideal soft-margin loss SVM: $L_{0/1}$ L 0 / 1 soft-margin loss SVM (dubbed as $L_{0/1}$ L 0 / 1 -SVM). Many of the existing (non)convex soft-margin losses can be viewed as one of the surrogates of the $L_{0/1}$ L 0 / 1 soft-margin loss. Despite its discrete nature, we manage to establish the optimality theory for the $L_{0/1}$ L 0 / 1 -SVM including the existence of the optimal solutions, the relationship between them and P-stationary points. These not only enable us to deliver a rigorous definition of $L_{0/1}$ L 0 / 1 support vectors but also allow us to define a working set. Integrating such a working set, a fast alternating direction method of multipliers is then proposed with its limit point being a locally optimal solution to the $L_{0/1}$ L 0 / 1 -SVM. Finally, numerical experiments demonstrate that our proposed method outperforms some leading classification solvers from SVM communities, in terms of faster computational speed and a fewer number of support vectors. The bigger the data size is, the more evident its advantage appears. [ABSTRACT FROM AUTHOR]

Kernel-based methods for support vector machines (SVM) have shown highly advantageous performance in various applications. However, they may incur prohibitive computational costs for large-scale sample datasets. Therefore, data reduction (reducing the number of support vectors) appears to be necessary, which gives rise to the topic of the sparse SVM. Motivated by this problem, the sparsity constrained kernel SVM optimization has been considered in this paper in order to control the number of support vectors. Based on the established optimality conditions associated with the stationary equations, a Newton-type method is developed to handle the sparsity constrained optimization. This method is found to enjoy the one-step convergence property if the starting point is chosen to be close to a local region of a stationary point, thereby leading to a super-high computational speed. Numerical comparisons with several powerful solvers demonstrate that the proposed method performs exceptionally well, particularly for large-scale datasets in terms of a much lower number of support vectors and shorter computational time. [ABSTRACT FROM AUTHOR]

This paper presents a miniature 50 mm3 inertial measurement unit (IMU) implemented using a folded microelectromechanical systems (MEMS) process. The approach is based on wafer-level fabrication of high aspect-ratio single-axis sensors interconnected by flexible hinges and folded into a 3-D configuration, like a silicon Origami [1]. Two different materials for flexible hinges have been explored, including photo-definable polyimide and parylene-C. We report, for the first time, an IMU prototype with seven operational sensors: three accelerometers, three gyroscopes, and a prototype of a reference clock. This paper concludes with the results of experimental characterization of inertial sensors demonstrating the feasibility of the proposed approach for a compact IMU. [2016-0267] [ABSTRACT FROM AUTHOR]

A new fiber Bragg grating (FBG)-based 2-D vibration sensor with adjustable sensitivity has been presented in this paper. The sensor uses screw as elastic body and arranges FBGs in two vertical directions, which can make the sensitivity of the two measurement directions equal to each other and ensure measurement of plane isometric. Four FBGs are used to measure the deflection of the screw tip caused by the combination of fiber and mass rings, which can increase the sensitivity and achieve temperature compensation. The position of the mass rings can be adjusted to achieve adjustable sensitivity. The principle of the sensor as well as modal simulation analyses is introduced. The dynamic and static property experimental analyses of the sensor are proposed. Experimental results show that the sensitivities of the sensor in the $X$ - and $Y$ -directions resides in 16.096 pm/g ~ 152.290 pm/g and 16.398 pm/g ~ 150.680 pm/g, respectively. The flat frequency range is 0–180 Hz. [ABSTRACT FROM PUBLISHER]

The virtual track train (VTT) is a new type of urban transit system with all-wheel active steering technology utilized to make the long combination vehicle run along the planned route (virtual track) like railway vehicles along physical track. This paper proposed a scalable cascade modular path following control strategy for the newly designed gantry virtual track train (G-VTT) based on preview & tracking controller, focusing on the lateral control for low-speed turning manoeuvre. The mathematical characteristics of the 2-norm was utilized to establish a local tracking objective function (LTOF) to find the optimal lateral acceleration (OLA) of the tracking point in the preview window and calculate the optimal wheel steering angle (WSA) to achieve the optimal local trajectory. Considering the lag time Td, the yaw stability and path following ability of the vehicle are contradictory in some cases, so the damping adjustment factor λ was proposed to optimize the damping characteristics and the time-delay stability factor ξ was proposed to evaluate the time-delay stability. The influence of the key parameters, preview time Tp, Td, ξ and λ, on the stability and following ability were analyzed by frequency domain analysis method, and the optimal region of ξ and λ was obtained to instruct the configuration of λ and Tp. Finally, steady-state and dynamic forward predictive models were proposed to compensate for the oscillation caused by the lag time. The effectiveness of the proposed control strategy, forward predictive models and the configuration method of λ and Tp were verified by theoretical analysis and numerical simulation. [ABSTRACT FROM AUTHOR]

In this paper, we propose a general framework termed centroid estimation with guaranteed efficiency (CEGE) for weakly supervised learning (WSL) with incomplete, inexact, and inaccurate supervision. The core of our framework is to devise an unbiased and statistically efficient risk estimator that is applicable to various weak supervision. Specifically, by decomposing the loss function (e.g., the squared loss and hinge loss) into a label-independent term and a label-dependent term, we discover that only the latter is influenced by the weak supervision and is related to the centroid of the entire dataset. Therefore, by constructing two auxiliary pseudo-labeled datasets with synthesized labels, we derive unbiased estimates of centroid based on the two auxiliary datasets, respectively. These two estimates are further linearly combined with a properly decided coefficient which makes the final combined estimate not only unbiased but also statistically efficient. This is better than some existing methods that only care about the unbiasedness of estimation but ignore the statistical efficiency. The good statistical efficiency of the derived estimator is guaranteed as we theoretically prove that it acquires the minimum variance when estimating the centroid. As a result, intensive experimental results on a large number of benchmark datasets demonstrate that our CEGE generally obtains better performance than the existing approaches related to typical WSL problems including semi-supervised learning, positive-unlabeled learning, multiple instance learning, and label noise learning. [ABSTRACT FROM AUTHOR]

This paper introduces the concept of the trans-rotary magnetic gear (TROMAG). The TROMAG is a magnetic device consisting of a rotor and a translator with an air gap in between, capable of converting linear motion to rotation, and vice versa, and doing the gearing action at the same time. Although its structure suggests that the TROMAG is simply the magnetic dual to the mechanical lead-screw, the authors show that the concept could also be derived from the elementary magnet arrays. The paper tries to establish an analogy between the TROMAG and the ordinary electrical machines with respect to the number of poles and the stack length, thus providing a deeper insight into the device operation. Three-dimensional (3-D) finite element analysis (FEA) is deployed to obtain the torque and force characteristics. Furthermore, it is shown that in addition to the permanent magnet (PM) TROMAG, developing the reluctance and the induction TROMAG is possible as well, but at a lower force density. [ABSTRACT FROM AUTHOR]

The most important components of the internal nut thread inspection process are measuring the pitch and pitch diameter value and comparing these with their respective specifications, as a test nut is considered to be abnormal when its pitch or pitch diameter does not meet specifications. This paper focuses on defects arising when the pitch diameter does not meet its specification. Because conventional contact strategies are not suited to the real-time inspection of internal thread defects, there have been a number of recent attempts to develop noncontact methods for real-time measurement. Some approaches have applied laser triangulation techniques in which reflected light is used to measure the z-axis depth on an object’s surface. Although point lasers have been shown to be highly precise in measuring pitch diameter, the optical architecture involved in these approaches is complicated and, as the laser must be scanned in a pointwise manner to derive cross-sectional measurements, the measurement process is too long to be compatible with real-time defect inspection. In this papere, an image-based method employing a line laser projector was used to develop a real-time image-based nut pitch diameter defect inspection system. The proposed method, which improves upon the conventional, widely used template-matching method for rapid defect detection, was shown to be effective through experimental validation. [ABSTRACT FROM AUTHOR]

This paper introduces a new type of reluctance-based magnetic lead screw (RBMLS) that is attractive for many potential applications. The proposed topology presents the implementation of discretized permanent magnets (PMs) and ferromagnetic semi-rings. The proposed RBMLS has PMs in only one of its movable parts, making it suitable for long-stroke applications. The proposed RBMLS is analyzed through the 2-D and 3-D finite-element method in order to test its topology and obtain the force produced per PM volume. Detailed design features are presented for the realization of a prototype. [ABSTRACT FROM AUTHOR]

Based on the unique advantages of fiber Bragg (FBG) in the field of vibration monitoring, a 2-D FBG accelerometer based on multi-axis flexure hinge is proposed. With the adoption of symmetrical multi-axis flexure hinges, the sensing probe has the same sensitivity in two measuring directions. The differential arrangement of the two couple FBGs can eliminate the influence of temperature, and in addition, its sensitivity is twice that of single FBG. The structure and principle of the sensor are explained, and its acceleration sensitivity and resonance frequency are analyzed theoretically in this paper. On this basis, the simulation analysis is carried out and two kinds of sensors for different occasions are made finally. Experimental studies show that the sensors can pick up the vibration signal accurately and have preferably temperature self-compensation effect under the condition form 20 °C to 90 °C. The stable frequency band of the max-sensor is 5 to 170 Hz, the acceleration sensitivity is up to 220 pm/g, the cross-interference in two measuring directions is not exceeding 4%, and the resonant frequency is 279.2 Hz. The min-sensor has a wide flat range from 20 to 600 Hz with an acceleration sensitivity of 40 pm/g. Its cross-interference is less than 3.5%, and the resonant frequency is 814.3 Hz. [ABSTRACT FROM AUTHOR]

Continuous gesture recognition aims at recognizing the ongoing gestures from continuous gesture sequences and is more meaningful for the scenarios, where the start and end frames of each gesture instance are generally unknown in practical applications. This paper presents an effective deep architecture for continuous gesture recognition. First, continuous gesture sequences are segmented into isolated gesture instances using the proposed temporal dilated Res3D network. A balanced squared hinge loss function is proposed to deal with the imbalance between boundaries and nonboundaries. Temporal dilation can preserve the temporal information for the dense detection of the boundaries at fine granularity, and the large temporal receptive field makes the segmentation results more reasonable and effective. Then, the recognition network is constructed based on the 3-D convolutional neural network (3DCNN), the convolutional long-short-term-memory network (ConvLSTM), and the 2-D convolutional neural network (2DCNN) for isolated gesture recognition. The “3DCNN-ConvLSTM-2DCNN” architecture is more effective to learn long-term and deep spatiotemporal features. The proposed segmentation and recognition networks obtain the Jaccard index of 0.7163 on the Chalearn LAP ConGD dataset, which is 0.106 higher than the winner of 2017 ChaLearn LAP Large-Scale Continuous Gesture Recognition Challenge. [ABSTRACT FROM AUTHOR]