Search

It is a significant capability for multi-industrial robots to plan an optimal collision-free path for both end-effectors and robotic arms. However, the path planning methods for co-worked multi-industrial robots, especially for closely co-worked industrial robots is still very limited. In this paper, to tackle the planning problem that has a specified distance constraint of end-effector in the context of complex collision avoidance, a two stage path planning method is proposed for the co-worked double industrial robot. In this two stage path planning method, the dual path planning with the distance constraint and the joint space planning of double robots are integrated sequentially. For the first stage, an algorithm named random sampling particle swarm optimization (RSPSO) is developed to plan the path for each end-effector, which can plan an optimal collision-free path with the specified distance constraint. For the second stage, the joint space planning that combines the inverse kinematics, D-H method and collision detection is performed to find the angular displacements with collision avoidance for dual robotic arms. Two simulation examples and an experiment are used to verify the proposed method.

Facial beauty prediction (FBP) has become an emerging area in the field of artificial intelligence. However, the lacks of data and accurate face representation hinder the development of FBP. Multi-task transfer learning can effectively avoid over-fitting, and utilize auxiliary information of related tasks to optimize the main task. In this paper, we present a network named Multi-input Multi-task Beauty Network (2M BeautyNet) and use transfer learning to predict facial beauty. In the experiment, beauty prediction is the main task, and gender recognition is the auxiliary. For multi-task training, we employ multi-task loss weights automatic learning strategy to improve the performance of FBP. Finally, we replace the softmax classifier with a random forest. We conduct experiments on the Large Scale Facial Beauty Database (LSFBD) and SCUT-FBP5500 database. Results show that our method has achieved good results on LSFBD, the accuracy of FBP is up to 68.23%. Our 2M BeautyNet structure is suitable for multiple inputs of different databases.

Artificial bee colony (ABC) algorithm is widely known for its distinguished exploration ability. However, its exploitation ability is relatively poor. To solve the problem, we propose a novel combinatorial search strategy, whose guided vector can be freely switched between a random vector and the global best vector. It can help improve the exploitation ability of ABC. At the same time, a random vector is beneficial to regulate the enhanced exploitation ability. In addition, both of them can pass information on to a current vector instead of only perturbing a current vector itself. The two guided vectors are chosen with a probability depending on the ratio of fitness information of a current vector to that of the global-best vector. Thus, one of the two guided vectors can be adaptively selected to direct the search. In addition, a mechanism of frequency of perturbation is employed to enhance the scale of information sharing between a current vector and a guided vector for each onlooker bee. Moreover, a modified greedy selection mechanism is designed to choose a child vector inspired by simulated annealing. Furthermore, the search strategy of multiple scouts is also employed in the last stage. Based on all these modifications, an improved ABC (IABC) is proposed. Finally, a few experiments are carried on 58 benchmark problems, including CEC2014 benchmark problems. The computational results exhibit the merit of IABC.

Gait impairment in Parkinson’s Disease (PD) is quantitatively assessed using the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS), a well-established clinical tool. Objective and efficient PD gait assessment is crucial for developing interventions to slow or halt its advancement. Skeleton-based PD gait MDS-UPDRS score estimation has attracted increasing interest in improving diagnostic efficiency and objectivity. However, previous works ignore the important cross-spacetime dependencies between joints in PD gait. Moreover, existing PD gait skeleton datasets are very small, which is a big issue in deep learning-based gait studies. In this work, we collect a sizable PD gait skeleton dataset by multi-view Azure Kinect sensors. The collected dataset contains 102 PD patients and 30 healthy older adults. In addition, gait data from 16 young adults (aged 24–50 years) are collected to further examine the effect of age on PD gait assessment. For skeleton-based automatic PD gait analysis, we propose a novel cross-spatiotemporal graph convolution network (CST-GCN) to learn complex features of gait patterns. Specifically, a gait graph labeling strategy is designed to assemble and group cross-spacetime neighbors of the root node according to the spatiotemporal semantics of the gait skeleton. Based on this strategy, the CST-GCN module explicitly models the cross-spacetime dependencies among joints. Finally, a dual-path model is presented to realize the modeling and fusion of spatial, temporal, and cross-spacetime gait features. Extensive experiments validate the effectiveness of our method on the collected dataset. [ABSTRACT FROM AUTHOR]

Photoacoustic imaging (PAI) has become a popular technique for biomedical diagnosis. However, current systems are overly complex or use bulky piezoelectric materials, necessitating options other than these methods for future PAI systems. In this study, a wearable transparent ultrasonic transducer was developed and applied in a PAI system with excellent performance. The acoustic sensor was fabricated with a sandwich structure consisting of 150 nm thick amorphous indium tin oxide on both sides of a $110~\mu \text{m}$ commercial polyvinylidene fluoride film. This single-element sensor exhibited the performance of a 6.7 MHz center frequency, an 86.3% −3 dB fractional bandwidth, and a $20\times20\,\,\times $ 36 mm3 volumetric imaging field. Modified imaging phantoms were made to evaluate the lateral resolution, imaging depth, ability to make a convex photoacoustic image, quality of 3D imaging, and capability to perform PAI directly in the ambient environment with a customized phantom. In summary, the study has demonstrated the sensor’s potential for large-field PAI application in wearable health devices at a low cost. [ABSTRACT FROM AUTHOR]

This article presents a high-efficient thermal analysis methodology and outlines a set of useful design guidelines for high-power linear ultrasonic motors. A thermal network modeling technique is introduced to rapidly calculate the temperature rise for a typical high-power linear ultrasonic motor with a V-shaped stator. First, a six-terminal equivalent circuit for Langevin piezoelectric transducer considering the prestress bolt loss and all three losses in piezoelectrics is proposed to calculate the stator loss. Then, the calculated losses along with the friction loss are used as inputs into the developed thermal network model to calculate the temperature rise of critical parts of the motor. Furthermore, an analytical mechanical model combined with the proposed thermal network is established to calculate the temporal–spatial distribution of the stator thermal stress for evaluating its fatigue under the heating-up process. Finally, some experimental measurements and finite-element simulations are conducted to validate the theoretical analyses. [ABSTRACT FROM AUTHOR]

Many previous works of soft wearable exoskeletons (exosuit) target at improving the human locomotion assistance, without considering the impedance adaption to interact with the unpredictable dynamics and external environment, preferably outside the laboratory environments. This article proposes a novel hierarchical human-in-the-loop paradigm that aims to produce suitable assistance powers for cable-driven lower limb exosuits to aid the ankle joint in pushing off the ground. It includes two primary loop layers: impedance learning in the external loop and human-in-the-loop adaptive management in the inner loop. Considering unknown terrains, its impedance model can be transferred to a quadratic programming problem with specified constraints, which a designed primal-dual optimization prototype then solves. Then, the presented impedance learning strategy is introduced to regulate the impedance model with the adaptive assistant powers for humans on different terrains. An adaptive controller is designed in the inner loop to balance the nonlinearities and compliance existing in the human-exosuit coexistence, while the robust mechanism compensates for disturbances to facilitate trajectory management without employing the general regressor. The advantage of the proposed technique over conventional solutions with fixed impedance parameters is that it can improve human walking performance over different terrains. Experiments demonstrate the significance of the approach. [ABSTRACT FROM AUTHOR]

Objects in aerial images show greater variations in scale and orientation than in other images, making them harder to detect using vanilla deep convolutional neural networks. Networks with sampling equivariance can adapt sampling from input feature maps to object transformation, allowing a convolutional kernel to extract effective object features under different transformations. However, methods such as deformable convolutional networks can only provide sampling equivariance under certain circumstances, as they sample by location. We propose sampling equivariant self-attention networks, which treat self-attention restricted to a local image patch as convolution sampling by masks instead of locations, and a transformation embedding module to improve the equivariant sampling further. We further propose a novel randomized normalization module to enhance network generalization and a quantitative evaluation metric to fairly evaluate the ability of sampling equivariance of different models. Experiments show that our model provides significantly better sampling equivariance than existing methods without additional supervision and can thus extract more effective image features. Our model achieves state-of-the-art results on the DOTA-v1.0, DOTA-v1.5, and HRSC2016 datasets without additional computations or parameters. [ABSTRACT FROM AUTHOR]

Objective: The tradeoff between calibration effort and model performance still hinders the user experience for steady-state visual evoked brain-computer interfaces (SSVEP-BCI). To address this issue and improve model generalizability, this work investigated the adaptation from the cross-dataset model to avoid the training process, while maintaining high prediction ability. Methods: When a new subject enrolls, a group of user-independent (UI) models is recommended as the representative from a multi-source data pool. The representative model is then augmented with online adaptation and transfer learning techniques based on user-dependent (UD) data. The proposed method is validated on both offline (N=55) and online (N=12) experiments. Results: Compared with the UD adaptation, the recommended representative model relieved approximately 160 trials of calibration efforts for a new user. In the online experiment, the time window decreased from 2 s to 0.56±0.2 s, while maintaining high prediction accuracy of 0.89-0.96. Finally, the proposed method achieved the average information transfer rate (ITR) of 243.49 bits/min, which is the highest ITR ever reported in a complete calibration-free setting. The results of the offline result were consistent with the online experiment. Conclusion: Representatives can be recommended even in a cross-subject/device/session situation. With the help of represented UI data, the proposed method can achieve sustained high performance without a training process. Significance: This work provides an adaptive approach to the transferable model for SSVEP-BCIs, enabling a more generalized, plug-and-play and high-performance BCI free of calibrations. [ABSTRACT FROM AUTHOR]

A multistatic inverse synthetic aperture radar (ISAR) system can observe a target from multiple observation angles. Compared with the monostatic ISAR system, the multistatic ISAR system can obtain more spatial sampling data, which provides the ability for high-resolution ISAR imaging. In some cases, the locations of radars are close. There are overlaps among observing angles, which brings little cross-range resolution improvement. However, such scenes are less considered indicates the literature. In the scene with overlaps, the image obtained by each radar may be similar due to the similar observation angles. In this letter, a novel multistatic ISAR imaging model is proposed by applying the similarity prior as a constraint. And an efficient image reconstruction algorithm is derived based on the orthogonality of observation matrix. Compared with existing CS-based methods, the proposed method can be directly applied on multistatic ISAR echoes without rearranging, which is more convenient in practical applications. Experiment results of simulated and measured data show that the proposed method achieves better performance especially under low signal-to-noise ratio (SNR) conditions. [ABSTRACT FROM AUTHOR]

This work considers the long baseline (LBL) acoustic localization problem. In the classical LBL localization approach, the time of arrival (TOA) of direct sound from each buoy is estimated independently, and the estimated TOAs are then utilized to solve for the target’s position. This approach ignores the fact that all of the estimated TOAs from different buoys correspond to the same target and thus are intrinsically correlated. To address this shortcoming, a novel localization framework based on track-before-detect (TBD) theory is established in this work, which allows the target location to be directly determined. Since it is completely freed from the traditional localization paradigm, some of the persistent and challenging problems faced by classical localization methods, such as direct sound selection and unacceptable false alarm/detection tradeoffs caused by threshold setting, can be circumvented. The effectiveness of the TBD approach is validated on both simulated and real datasets. The experimental results indicate that the proposed approach can track targets continuously and accurately and can be effectively implemented even under challenging conditions. [ABSTRACT FROM AUTHOR]

With the maturity of deep learning algorithm in synthetic aperture radar (SAR) target recognition field, convolutional neural network (CNN) has become the most effective model. However, the interpretability and the separability of feature maps extracted from convolution layers have not been specially analyzed neither qualitatively nor quantitatively, which makes the traditional model work like a “black box.” To alleviate the problem, a novel model based on separability measure (SM)-CNN is proposed in this letter, which introduces the principle of maximal coding rate reduction (MCR $^{2}$) to the backbone module. SM-CNN quantitatively analyzes the separability of the feature maps and takes the value as a vital part of the loss function to guide the training process of the model. The calculation process of the SM values can be strictly derived mathematically, so it is more interpretable, turning the black box into a “gray box.” In addition, the proposed model can achieve comparable recognition performance of the backbone networks with reduced computational complexity. Comparative experiments based on moving and stationary target acquisition and recognition (MSTAR) and OpenSARShip datasets verify the effectiveness and practicability of the method proposed in this letter. [ABSTRACT FROM AUTHOR]

In a high-power semiconductor power module where multiple chips are populated in parallel, the thermal imbalance problem arises. In this article, the next-generation standard high-voltage insulated-gate bipolar transistor module for high-speed railcar traction is investigated to characterize the internal thermal imbalance among paralleling chips and evaluate the impact on temperature distribution in field application. The former is accomplished by a new test algorithm with a purposely designed module sample, while the latter is achieved via an advanced electro-thermal simulation from chip level to system level. The experimentally verified thermal network among paralleling chips reveals the multiple reasons for the thermal imbalance. A comparison between the simulation and experimental results regarding the paralleling chips’ power-cycling scenario helps to confirm the effectiveness of the chip model, the thermal network, and the electro-thermal simulation approach. A further electro-thermal simulation shows that the module internal thermal imbalance can lead to significant junction temperature discrepancy among paralleling chips in traction inverter applications, which should be taken into consideration for the long-term reliability assessment. [ABSTRACT FROM AUTHOR]

LLC resonant converter is widely used due to its soft switching, high power density, and high efficiency. To incorporate the merits of the LLC resonant tank into a three-phase single-stage ac/dc converter, this article proposes a three-phase single-stage ac/dc converter based on Swiss rectifier and three-level LLC topology. It takes advantage of the high-efficiency Swiss rectifier to convert three-phase ac voltages into three variable dc voltages and interface a three-level LLC resonant circuit. With help of the LLC resonant tank, all the high-frequency three-level switches achieve soft switching. To both regulate the output voltage and achieve three-phase power factor correction, the frequency modulation in combination with the duty cycle control strategy is employed by splitting the resonant current into three parts and then forming three phase sinusoidal currents. The proposed converter's detailed operation principles, steady-state characteristics, and design considerations are given. Finally, a 10 kW experimental prototype with three-phase 220 Vac inputs and 500 Vdc output is built to verify the effectiveness and feasibility of the proposed topology and its control strategy. [ABSTRACT FROM AUTHOR]

A method to derive the junction-to-case level coupling thermal resistances among paralleling chips in the multichip semiconductor power module has been proposed in this article. It is revealed that the traditional structure function methodology (SFM), which is developed for analyzing the self-heating effect of the driving-point chip, cannot be applied directly in the thermal coupling analysis, where the influence of the driving-point chip on the neighboring acceptor chip is concerned. This article shows that by combining the traditional SFM with the frequency-domain analysis (FDA), the effective separation points between the module case and the external cooling setup can be identified in both time and frequency domains, which correspond to the characteristic frequency points in the complex loci of the thermal impedance curves. By extracting the amplitudes at the characteristic frequencies from the complex loci in the FDA, the junction-to-case level coupling thermal resistances can be derived. Both simulation and experimental study have been carried out to verify the concept. The agreement between the two justifies the proposed method. [ABSTRACT FROM AUTHOR]

A rising challenge in manufacturing data analysis is training robust regression models using limited labeled data. In this work, we investigate a semi-supervised regression scenario, where a manufacturing process operates on multiple mutually correlated states. We exploit this inter-state correlation to improve regression accuracy by developing a novel co-training method, namely Correlated Bayesian Co-training (CBCT). CBCT adopts a block Sparse Bayesian Learning framework to enhance multiple individual regression models which share the same support. Additionally, CBCT casts a unified prior distribution on both the coefficient magnitude and the inter-state correlation. The model parameters are estimated using maximum-a-posteriori estimation (MAP), while hyper-parameters are estimated using the expectation-maximization (EM) algorithm. Experimental results from two industrial examples shows that CBCT successfully leverages inter-state correlation to reduce the modeling error by up to 79.40%, compared to other conventional approaches. This suggests that CBCT is of great value to multi-state manufacturing applications. [ABSTRACT FROM AUTHOR]

Acoustic levitation is considered one of the most effective noncontact particle manipulation methods, along with aerodynamic, ferromagnetic, and optical levitation techniques. It is not restricted by the material properties of the target. However, existing acoustic levitation techniques have some drawbacks that limit their potential applications. Therefore, in this article, an innovative approach is proposed to manipulate objects more intuitively and freely. By taking advantage of the transition periods between the acoustic pulse trains and electrical driving signals, acoustic traps can be created by switching the acoustic focal spots rapidly. Since the high-energy-density points are not formed simultaneously, the computation of the acoustic field distribution with complicated mutual interference can be eliminated. Therefore, compared to the existing approaches that created acoustic traps by solving pressure distributions using iterative methods, the proposed method simplifies the computation of time delay and makes it possible to be solved even with a microcontroller. In this work, three experiments have been demonstrated successfully to prove the capability of the proposed method including lifting a Styrofoam sphere, transportation of a single target, and suspending two objects. Besides, simulations of the distributions of acoustic pressure, radiation force, and Gor’kov potential were conducted to confirm the presence of acoustic traps in the scenarios of lifting one and two objects. The proposed tactic should be considered effective since the results of the practical experiments and simulations support each other. [ABSTRACT FROM AUTHOR]

The star-connected Cascaded H-Bridge (CHB) converter is one of the most attractive multilevel topologies for medium/high-voltage STATCOM. However, due to the fact that CHB has separated dc-links, three-phase unbalanced active powers cannot exchange automatically, which, thus, requires extra zero sequence voltage (ZSV) to eliminate them. Especially when negative sequence current is compensated, the negative sequence current compensation range (NCCR) is seriously limited. To extend the NCCR, this article combines the CHB and the converter with a common dc-link, which is utilized to exchange and eliminate the unbalanced active power generated from CHB. Based on this new active power exchange way and conventional ZSV injection, a novel cluster voltage balancing control method is proposed to eliminate the unbalanced cluster active powers. Moreover, based on the proposed control, the NCCR is identified, which is larger than that based on the conventional control. Finally, the proposed control and the derived NCCR are verified by simulation and experimental results on a 400 V/7.5 kvar Hybrid Cascaded Converter STATCOM. [ABSTRACT FROM AUTHOR]

With the widespread deployment of data centers, Internet service providers are expecting more efficient design strategies to build long-haul data center interconnect (DCI) networks. In this paper, we propose a heuristic-based optimization framework to design these networks. Through this framework, network designers can obtain a site-type design scheme that arranges customized site types such as in-line amplifiers, dynamic gain equalizers, optical terminal multiplexers, and electrical regenerators, and three strategies are provided for reference. Taking the quality of transmission as the main metric, and the overall cost of the network as the ancillary measurement, we compare the schemes obtained by the proposed framework against the baseline scheme obtained by a traditional periodic design strategy. Simulations are conducted on a topology of the Tencent DCI network. Under the condition that all schemes ensure that the minimum general signal-to-noise ratio (GSNR) remains above the given GSNR threshold, the schemes designed by our framework can achieve overall cost savings up to 25.73%. [ABSTRACT FROM AUTHOR]

Squeeze film damper (SFD) is one of the key components in the sophisticated rotating machines in the high-end aerospace industry. Rolling bearings in the aeroengine spindle system are often used in combination with elastic supports and SFDs to reduce rotor vibrations. Accurate coefficient identification of oil films in SFDs is critical to ensure safe and reliable operations of the complex equipment. However, this task remains a challenging issue in the past decades due to the difficulties in real-time measurement and computation during operations. The highly nonlinear relationship between the measurement and oil film coefficient cannot be effectively captured using traditional methods. This article proposes a novel deep-learning-based information fusion methodology for oil film coefficient identification of SFDs. An end-to-end deep neural network model structure is established, which explores heterogeneous SFD data simultaneously with fusion architecture and directly outputs the identification results. The high-dimensional data can be automatically processed, and high identification accuracy can be achieved. Experiments on the SFD test rig are carried out for validation. The experimental results demonstrate that the proposed method can well identify the oil film coefficients in an intelligent manner and is robust against environmental noise and different parameter settings. The proposed methodology is, thus, promising for applications in real industries. [ABSTRACT FROM AUTHOR]

Active three-dimensional(3-D) thermography combines 3-D shape and active thermography. Thus, it enables the provision of the intuitive thermographic inspection results for composite parts with a complex geometry. However, conventional 3-D thermography acquires thermographic information and 3-D shape through at least two independent sensors and requires complex cross-modal image registration algorithm based on the keypoint detection and matching. In this article, an active 3-D thermography system is proposed using only one thermal camera for moving objects. This system does not require an independent 3-D sensor while the thermal camera acts as the 3-D sensor. Nature behind that is that a mathematical model is proposed to unify the line scanning thermography and laser triangulation in the dynamic scanning process, taking a line laser as both heat excitation for active thermography and spatial coding for 3-D reconstruction. Furthermore, the model enables the feature-free registration of thermogram sequence and 3-D shape, so that the registration is fast and robust without keypoint features. The experiments on standard height specimens, 3-D printing glass fiber composites, and carbon fiber intake tube have shown the error is calibrated within 0.25 mm in the range of 1 to 150 mm and evidenced the capability for subsurface defects detection. The ease and robustness of the proposed active 3-D thermography have a bright future for 3-D measurement, defects detection, and quality control in the production line. [ABSTRACT FROM AUTHOR]

Exoskeletons are increasingly used to assist humans in military, industry, and healthcare applications, thereby enabling individuals to gain increased strength and endurance. This article proposes a novel human-in-the-loop control framework for a fully actuated lower limb exoskeleton with high degree-of-freedoms (DoFs), allowing users to walk without crutches or other external stabilization tools. To imitate the natural lower limb motion of users, a novel barrier energy function is utilized for the design of the control strategy, where the human-robot manipulation space is reformulated as a human-voluntary and a robot-constrained region. The variations in the barrier energy function are based on the distance between the center of mass and zero moment point of the walking exoskeleton, thereby constraining the lower limb motion of the user to a compliant region around various desired trajectories. Based on varying regional functions, the proposed strategy is designed to control the exoskeleton to follow appropriate ergonomic trajectories. For such a purpose, an adaptive controller is exploited considering the functions of the human effort and the robot's capabilities simultaneously, and a smooth motion transition can be achieved between the human and robot regions. Finally, physical experiments are conducted on a ten-DoFs walking exoskeleton to validate the stability and robustness of the proposed control framework with subjects performing flat walking, turning, and obstacle avoidance movements. [ABSTRACT FROM AUTHOR]

In reality, learning from multi-view multi-label data inevitably confronts three challenges: missing labels, incomplete views, and non-aligned views. Existing methods mainly concern the first two and commonly need multiple assumptions to attack them, making even state-of-the-arts involve at least two explicit hyper-parameters such that model selection is quite difficult. More toughly, they will fail in handling the third challenge, let alone addressing the three jointly. In this paper, we aim at meeting these under the least assumption by building a concise yet effective model with just one hyper-parameter. To ease insufficiency of available labels, we exploit not only the consensus of multiple views but also the global and local structures hidden among multiple labels. Specifically, we introduce an indicator matrix to tackle the first two challenges in a regression form while aligning the same individual labels and all labels of different views in a common label space to battle the third challenge. In aligning, we characterize the global and local structures of multiple labels to be high-rank and low-rank, respectively. Subsequently, an efficient algorithm with linear time complexity in the number of samples is established. Finally, even without view-alignment, our method substantially outperforms state-of-the-arts with view-alignment on five real datasets. [ABSTRACT FROM AUTHOR]

Sanding is a common yet important task in the manufacturing of many wooden objects (e.g. furniture, decoration box), where the coated layer attached on objects is removed after the interaction with sanding belts. Existing sanding operation is heavily dependent on manual works, which is highly labor-intensive and with the low consistency of quality, and the issues of safety and health also arise after continuous working in the noisy and dusty environment. To deal with the aforementioned, this paper presents the development of a new autonomous sanding robot. The autonomous capability of the developed robot is reflected in the whole procedure of sanding. In particular, the CAD model of the target object is automatically constructed with the structured-light technology, and the sanding behavior on the target surface is self-regulated under the desired impedance model. Such feature makes the robot capable of working towards uncertain objects with minimum human involvement. The proposed impedance controller has the model-free advantage, by using the adaptive neural networks (NNs) to compensate the uncertain dynamics and the unknown disturbances online. The stability of the closed-loop system is rigorously proved with Lyapunov methods, and experimental results on different objects are presented to validate the performance of the developed robot. The implementation of the developed robot can systematically address the problems associated with manual works. Note to Practitioners—The current working environment of sanding is not healthy or safe to humans, due to the nature of noise, dust, high-speed sanding belt. By controlling the robot to autonomously perform sanding tasks can keep humans away from such environment and hence systematically address the issues of health and safety. This paper presents a new impedance control method for sanding robot. In impedance control, the control goal is specified as a dynamic relationship between the contact position and the interaction force. In this paper, the contact position is determined by using the 3D vision sensor, while the interaction force is regulated by referring to the human experience. In addition, the influence caused by unmodeled factors (e.g. unknown dynamics of the sanded object) is also dealt with by using the techniques of NNs. Such setting can effectively guarantee the sanding quality and also avoid the physical damage to the sanded object. Therefore, it lays the foundation for the autonomous robotic sanding. [ABSTRACT FROM AUTHOR]

Scene text detection and recognition have been well explored in the past few years. Despite the progress, efficient and accurate end-to-end spotting of arbitrarily-shaped text remains challenging. In this work, we propose an end-to-end text spotting framework, termed PAN++, which can efficiently detect and recognize text of arbitrary shapes in natural scenes. PAN++ is based on the kernel representation that reformulates a text line as a text kernel (central region) surrounded by peripheral pixels. By systematically comparing with existing scene text representations, we show that our kernel representation can not only describe arbitrarily-shaped text but also well distinguish adjacent text. Moreover, as a pixel-based representation, the kernel representation can be predicted by a single fully convolutional network, which is very friendly to real-time applications. Taking the advantages of the kernel representation, we design a series of components as follows: 1) a computationally efficient feature enhancement network composed of stacked Feature Pyramid Enhancement Modules (FPEMs); 2) a lightweight detection head cooperating with Pixel Aggregation (PA); and 3) an efficient attention-based recognition head with Masked RoI. Benefiting from the kernel representation and the tailored components, our method achieves high inference speed while maintaining competitive accuracy. Extensive experiments show the superiority of our method. For example, the proposed PAN++ achieves an end-to-end text spotting F-measure of 64.9 at 29.2 FPS on the Total-Text dataset, which significantly outperforms the previous best method. Code will be available at: git.io/PAN. [ABSTRACT FROM AUTHOR]

Deformable object manipulation (DOM) is an emerging research problem in robotics. The ability to manipulate deformable objects endows robots with higher autonomy and promises new applications in the industrial, services, and health-care sectors. However, compared to rigid object manipulation, the manipulation of deformable objects is considerably more complex and is still an open research problem. Addressing DOM challenges demands breakthroughs in almost all aspects of robotics: hardware design, sensing, (deformation) modeling, planning, and control. In this article, we review recent advances and highlight the main challenges when considering deformation in each subfield. A particular focus of our article lies in the discussions of these challenges and proposing future directions of research. [ABSTRACT FROM AUTHOR]

One typical application of a robotic exoskeleton is to automate rehabilitation, where the robot is worn by a stroke patient and provides assistance to help perform repetitive motions and regain motor functions. The deployment of exoskeletons can alleviate the shortage of experienced therapists and would also play a vital role in countries with aging populations. However, the intelligence level of existing exoskeletons is relatively low, wherein a robot cannot adapt to either the online change of a subject’s motion (e.g., the gait pattern) or the variation of his/her body parameters (i.e., a new subject who is going to wear the robot), potentially resulting in conflict between human and robot and possibly even leading to physical damage. As such, the application of a robotic exoskeleton in clinical studies is limited. This article introduces a new bilateral exoskeletal assistive robot for rehabilitation (BEAR-H) in which the main novelty is the integration of multiple intelligent features, such as gait recognition and synchronization, cloud-computing diagnosis, and individualized gait generation. Such an integration helps the robot to better understand the patient’s condition and hence provide effective assistance, by the end achieving smart rehabilitation. BEAR-H is a successfully commercialized product, and its performance has been validated in actual clinical studies with 30 patients, producing experimental results from different aspects that are analyzed and presented. [ABSTRACT FROM AUTHOR]

In this letter, a bidirectional symmetric 100 Gb/s/ $\lambda $ coherent passive optical network (PON) scheme with a simplified optical network unit (ONU) transceiver is proposed and experimentally demonstrated. In the proposed scheme, the downstream signal is generated in a dual-polarization coherent transmitter with Alamouti-coding at the optical line terminal (OLT) side and then received by a single-polarization heterodyne receiver at the ONU side. The upstream signal is generated in a low-cost C-band electro-absorption modulated laser (EML) at the ONU side and then received by a standard dual-polarization integrated coherent receiver at the OLT side. The experimental results show that more than 30 dB power budget is achieved after 25 km standard single mode fibre transmission for both downstream and upstream signal at 100 Gb/s/ $\lambda $. The cost at the ONU side is maintained low as the required bandwidth of the optoelectrical device is >25 GHz and the local oscillator power at the ONU side can be as low as 4 dBm. [ABSTRACT FROM AUTHOR]

This article extends the theory of Class-X power amplifiers (PAs) with maximally flat waveforms by considering an arbitrary number of harmonics in the analysis. The proposed theory shows that the maximally flat voltage and current waveforms can be established with a proper fundamental load corresponding to a set of randomly and independently distributed purely reactive harmonic load impedances. It also shows that the efficiency increases with the number of harmonics regardless of the values of the harmonics. The proposed theory of Class-X PA allows for the design of a high-efficiency PA considering the effects of high-order harmonics in the presence of significant parasitic effects. To demonstrate this ability, a 3.5-GHz Class-X PA considering the effects of nine harmonics is designed and fabricated with a packaged 10-W device. The simulated and measured results show a drain efficiency of 84% for an output power of 40.5 dBm. [ABSTRACT FROM AUTHOR]

Dynamic topology, fast-changing channels and the time sensitivity of safety-related services present challenges to the status quo of resource allocation for cellular-underlaying vehicle-to-vehicle (V2V) communications. In this paper, we investigate a novel federated multi-agent deep reinforcement learning (FedMARL) approach for the decentralized joint optimization of channel selection and power control for V2V communication. The approach takes advantage of both deep reinforcement learning (DRL) and federated learning (FL), satisfying the reliability and delay requirements of V2V communication while maximizing the transmit rates of cellular links. Specifically, we elaborately construct individual V2V agent implement by the dueling double deep Q-network (D3QN), and design the reward function to train V2V agents collaboratively. As a result, each agent individually optimizes channel selection and power level based on its local observations, including the instantaneous channel state information (CSI) of corresponding V2V link, the instantaneous co-channel interference from the cellular link, the previous channels selections of nearby V2V pairs, and the queue backlog at the V2V transmitter. Another important aspect is that we incorporate FL to alleviate the training instability problem induced by cooperative multi-agent environment. The local DRL models of different V2V agents are federated periodically, addressing the limitations of partial observability on the entire network status for individual agent, and accelerating the training process of multi-agent learning. Validated via simulations, the proposed FedMARL scheme shows superiority to the baselines in terms of the cellular sum-rate and the V2V packet delivery rate. [ABSTRACT FROM AUTHOR]

Product failure resulting from sourcing supplier’s defective component has compelled a brand owner to enhance quality management, especially when the supplier has informational advantage. In this article, we examine a brand owner’s problem of screening a certain supplier’s inherent quality level with an attempt to induce supply chain partners’ quality efforts using the warranty contract based on information acquired from inspection technology. A supplier’s inherent quality level is herein determined by the private information held by the supplier and is typically characterized as an uncertain variable. The optimal warranty contracts and the expected profits of the brand owner and the supplier are derived from four different scenarios under the framework of uncertainty theory and principal–agent theory. We find that under the condition of pure double moral hazard or pure adverse selection, the first-best outcomes can be achieved without incurring agency cost under the designed contract. However, double moral hazard combined with adverse selection often leads to underinvestment in quality efforts as the supplier can shirk by misreporting her type. Consequently, we present the menu of warranty contracts to screen the supplier’s private information. Finally, we provide empirical managerial recommendations on mitigating potential adverse impacts caused by information asymmetry, supported with numerical investigations. [ABSTRACT FROM AUTHOR]

In the fully mechanized mining face, proper chain tension of chain transmission system is of great significance in reducing the failure rate and improving the service life of scraper conveyor. However, the disturbance caused by random load and variable stiffness makes it very difficult to control chain tension. An adaptive tension control scheme combining a neural command filtering backstepping algorithm and online identification is proposed in this article based on a novel tension model. First of all, the proposed tension model is represented by a linear regression part and a nonlinear mapping part, which effectively reflects the chain transmission characteristics and avoids the full measurement of each chain ring. Then, Levant's differentiator and state observer are used to expand input variables for identification algorithm and neural networks estimation. To avoid the adverse effect of random load on parameter adaption, stiffness online identification algorithm is employed in improving the tension tracking performance. To compensate for nonlinear mapping part and uncertain nonlinearity caused by random load, neural adaptive estimator is effectively integrated with the robust integral term in backstepping design. Moreover, Lyapunov stability of the proposed controller is guaranteed. Finally, comparative experimental results of six controllers are completed to verify the effectiveness of the proposed controller in the static and random load mode. [ABSTRACT FROM AUTHOR]

We propose a direct detection scheme for pulse amplitude modulation (PAM) signals over long-reach fiber transmission with phase noise mitigation. In this scheme, an optical carrier is transmitted as pilot together with the optical signal, whose frequency is located at the edge of the optical signal. The electrical field of the optical signal can be retrieved via the beating between the pilot and optical signal in one photodetector at the receiver side, which facilitates the digital compensation of chromatic dispersion effect. To enable real value signal processing, digital carrier regeneration is proposed to mitigate the phase noise fluctuation in the proposed scheme. To further improve the performance, the digital signal processing including Kramers–Kronig field reconstruction and Volterra equalizer are applied to reduce the signal-to-signal beat noise and nonlinear distortions after fiber transmission. The experimental results for both 56- and 64-Gb/s PAM-4 signal transmission cases show that the performances can be significantly improvement based on the proposed scheme and DSP methods. Both 56- and 64-Gb/s PAM-4 signal transmissions over 720- and 1440-km standard single mode fiber have been successfully achieved under 7% and 20% forward error correction codes, respectively. [ABSTRACT FROM AUTHOR]

A key step in remanufacturing is disassembly of the “core” or the returned product to be remanufactured. Disassembly sequence planning is challenging due to uncertainties in the conditions of the cores. Rust, corrosion, deformation, and missing parts may require disassembly plans to be changed and adapted frequently. Conventional industrial automation that usually serves in repetitive and structured activities may fail when it is applied to disassembly. This research investigates the flexible sequencing of robotic disassembly in the presence of failed automation operations and develops online recovery by incorporating backup actions. It starts with modeling the time and success rate of a backup action. The expected disassembly time and completion rate of a disassembly plan are deduced according to the failure probability of both the operations and their backup actions. A biobjective optimization model for robotic disassembly sequence planning is established using a dual-selection multiobjective evolutionary algorithm. Two solution selection criteria are combined to produce potential offspring candidates in each evolutionary generation. Experimental results show that the backup actions allow efficient recovery from automation and can potentially improve the robustness of robotic disassembly. Note to Practitioners—This research was motivated by the development of automated disassembly techniques. Industrial automation techniques usually use predetermined operation motions. Robotic disassembly using such an approach may fail due to uncertainties in the condition of the products (e.g., positioning and geometry). This article introduces backup actions for disassembly sequence planning and describes the logic and reasoning of their implementation. Our proposed method can theoretically increase the completion rate of automated robotic disassembly. Experimental studies suggested that backup actions are efficient in providing a reliable disassembly sequence and, thus, can improve the robustness of robotic disassembly. In future research, we will implement typical backup actions and establish an automated disassembly process with a replanning module. [ABSTRACT FROM AUTHOR]

The communication-based train control (CBTC) system is a typical cyber physical system in urban rail transit. The train-ground communication system is a very important subsystem of the CBTC system and uses the wireless communication protocols to transmit control commands. However, it faces some potential information security risks. To ensure information security of the train-ground communication system, an intrusion detection method based on machine learning and state observer is proposed to detect and recognize various attacks in this paper. The detection system not only detects the anomalies of the wireless network data, but also detects the anomalies of the train physical states. This method includes two layers. The first layer is used to detect and identify wireless network attacks based on machine learning algorithms, such as the random forest algorithm and the gradient boosted decision tree algorithm. The second layer is used to detect the abnormal physical state of train operation based on a state observer. By combining the results of the above two layers, a comprehensive intrusion detection result is given. The simulation results show that the proposed method is effective and practical. [ABSTRACT FROM AUTHOR]

Recently, deep unfolding networks with interpretable parameters have been widely utilized in direction of arrival (DOA) estimation due to the faster convergence speed and better generalization ability. However, few consider the nested array for gridless DOA estimation. In this letter, we propose a deep alternating projection network to address the problem. We first convert the covariance matrix into a measurement vector in the form of atomic norm, which can reduce the matrix dimension during projection. We then train the proposed network to alternately obtain the positive semi-definite matrix and the corresponding irregular Hermitian Toeplitz matrix, where the loss function is derived by employing the trace of network output. Finally, we apply the irregular root Multiple Signal Classification (MUSIC) method to obtain gridless DOA via nested array. We demonstrate that the proposed networks can accelerate the convergence rate and reduce computational cost. Simulations verify the performance of proposed networks in comparison with the existing methods. [ABSTRACT FROM AUTHOR]

This brief studies a bearing-based formation tracking control problem that the leaders move at a constant speed for a second-order multi-agent system. We propose a novel control law to achieve formation tracking control for multiple motion agents in a prescribed time. The convergence time can be selected by the user because the control input is based on a scaling gain and the differentiation of the control input is also continuous. Collision avoidance, as an essential problem of multi-agent in actual formation control, is also studied in this brief. Finally, numerical simulation experiments verify the effectiveness of our proposed algorithm. [ABSTRACT FROM AUTHOR]

The silicon photonic integrated network based on the thermo-optical or electro-optical effect generally has the disadvantages of volatile, large footprint, and significant energy loss. Here, we propose and demonstrate an all-optical controllable switch device with non-volatile, broadband adjustable, and bistable fast switching performance by integrating the phase-change material (Ge2Sb2Te5) on the optical fiber. By adjusting the energy of the laser pulse train exerting on the optical fiber, the phase state of the phase-change material can be switched, thereby realizing the “ON” and “OFF” function of the optical switch. As a result, a non-volatile optical switch with a wide bandwidth working range (up to 60 nm) and a significant transmission contrast (2 dB) is demonstrated. This device is expected to help realize fast and broadband optical routing in fiber optical communication networks as the optical switch. Furthermore, its logic operation function can impart optical fiber with new storage and computing functionalities except for transmission and sensing. It can be combined with an optical field-programmable gate array (OFPGA) and play a part in optical fiber photonic circuitry in the future. [ABSTRACT FROM AUTHOR]

Since LLC converter is capable of naturally creating the soft-switching condition to improve the overall efficiency, it is widely used in dc–dc applications. Using LLC resonant tank to realize soft switching for three-phase ac–dc converters would be a very promising solution to high efficiency and low cost ac–dc conversion, an LLC-based three-phase single-stage ac–dc converter was configured by taking advantage of three-phase voltage feature $ u_{\mathrm a}$ + $ u_{\mathrm b}$ + $ u_{\mathrm c}$ = 0 and incorporating three bidirectional switches. Because of time-varying feature of three-phase voltages, the input voltage of the converter to the LLC resonant tank is no longer a symmetrical square wave, but a time-varying and asymmetrical three-level waveform, which makes the analysis of the resonant circuit behavior much complicated. Some traditional analysis methods, such as fundamental harmonic approximation and its variants, are impossible to deal with. To overcome this analysis obstacle, this article proposes a numerical analysis method to analyze the time-varying three-level LLC resonant tank for this converter. According to the monotonous change of LLC input voltage in one fundamental cycle, it calculates and determines the time durations of three phases and the corresponding operation modes in each switching cycle. And this numerical analysis method is used to calculate the numerical relationship between the two control variables (switching frequency $ f_{\mathrm s}$ and duty cycle $ D_{\mathrm M}$) of this converter and the voltage gain and power level, which is a guideline for the design of the resonant tank parameters and control strategy of the LLC-based three-phase single-stage ac–dc converter. Finally, the time-varying waveforms of $ f_{\mathrm s}$ and $ D_{\mathrm M}$ obtained by numerical calculation well match the PSIM simulation results, and a 3.3-kW three-phase 220Vac input and 48Vdc output prototype was built to verify the effectiveness of the proposed numerical analysis method and the feasibility of the control strategy. [ABSTRACT FROM AUTHOR]

In an era of ubiquitous large-scale evolving data streams, data stream clustering (DSC) has received lots of attention because the scale of the data streams far exceeds the ability of expert human analysts. It has been observed that high-dimensional data are usually distributed in a union of low-dimensional subspaces. In this article, we propose a novel sparse representation-based DSC algorithm, called evolutionary dynamic sparse subspace clustering (EDSSC). It can cope with the time-varying nature of subspaces underlying the evolving data streams, such as subspace emergence, disappearance, and recurrence. The proposed EDSSC consists of two phases: 1) static learning and 2) online clustering. During the first phase, a data structure for storing the statistic summary of data streams, called EDSSC summary, is proposed which can better address the dilemma between the two conflicting goals: 1) saving more points for accuracy of subspace clustering (SC) and 2) discarding more points for the efficiency of DSC. By further proposing an algorithm to estimate the subspace number, the proposed EDSSC does not need to know the number of subspaces. In the second phase, a more suitable index, called the average sparsity concentration index (ASCI), is proposed, which dramatically promotes the clustering accuracy compared to the conventionally utilized SCI index. In addition, the subspace evolution detection model based on the Page-Hinkley test is proposed where the appearing, disappearing, and recurring subspaces can be detected and adapted. Extinct experiments on real-world data streams show that the EDSSC outperforms the state-of-the-art online SC approaches. [ABSTRACT FROM AUTHOR]

The cybersecurity breaches expose surveillance video streams to forgery attacks, under which authentic streams are falsified to hide unauthorized activities. Traditional video forensics approaches can localize forgery traces using spatial-temporal analysis on relatively long video clips, while falling short in real-time forgery detection. The recent work correlates time-series camera and wireless signals to detect looped videos but cannot realize fine-grained forgery localization. To overcome these limitations, we propose Secure-Pose, which exploits the pervasive coexistence of surveillance and Wi-Fi infrastructures to defend against video forgery attacks in a real-time and fine-grained manner. We observe that coexisting camera and Wi-Fi signals convey common human semantic information and forgery attacks on video streams will decouple such information correspondence. Particularly, retrievable human pose features are first extracted from concurrent video and Wi-Fi channel state information (CSI) streams. Then, a lightweight detection network is developed to accurately discover forgery attacks and an efficient localization algorithm is devised to seamlessly track forgery traces in video streams. We implement Secure-Pose using one Logitech camera and two Intel 5300 NICs and evaluate it in different environments. Secure-Pose achieves a high detection accuracy of 98.7% and localizes abnormal objects under playback and tampering attacks. [ABSTRACT FROM AUTHOR]

This paper focuses on the spatial-spectral terahertz (THz) networks, where transmitters equipped with leaky-wave antennas send information to their receivers at the THz frequency bands. As a directional and nearly planar antenna, the leaky-wave antenna allows for information transmissions with narrow beams and high antenna gains. The conventional large antenna arrays are confronted with challenging issues such as scaling limits and path discovery in the THz frequencies. Therefore, this work exploits the potential of leaky-wave antennas in the dense THz networks, to establish low-complexity THz links. By addressing the propagation angle-frequency coupling effects, the transmission rate is analyzed. The results show that the leaky-wave antenna is efficient for achieving the high-speed transmission rate. The co-channel interference management is unnecessary when the THz transmitters with large subchannel bandwidths are not extremely dense. A simple subchannel allocation solution is proposed, which enhances the transmission rate compared with the same number of subchannels with the equal allocation of the frequency band. After subchannel allocation, a low-complexity power allocation method is proposed to improve the energy efficiency. [ABSTRACT FROM AUTHOR]

The key advantage of silicon photonics comes from its potential for large scale integration, in a low-cost and scalable fashion. This has sustained the growth in the area despite disadvantages such as the lack of a monolithic light source, or the absence of a second order non-linear response (χ(2)). Thus far, the work in the field has focused on reporting individual devices from a single die, with excellent performances. Wafer-level results, an area which has not been addressed sufficiently, is a critical aspect of silicon photonics and will provide the community with information regarding scalability and variation, which will be the key differentiating advantage of silicon photonics over other photonic platforms. In this work, we report the development of a low-loss, high-bandwidth C-band silicon photonic platform on a 200 mm CMOS-compatible process line, demonstrating wafer-level performance in the process. Ultra-low waveguide propagation loss with median values as low as 0.43 dB/cm has been achieved. Silicon Mach-Zehnder and microring modulators with median bandwidth of 38.5 and 43 GHz respectively are presented. Finally, germanium waveguide-integrated photodetectors with median bandwidth of 43 GHz are reported. The results reported in this work are comparable to prior demonstrations concerning individual devices. The baseline designs on this platform presented in this work can be accessed commercially from CompoundTek. [ABSTRACT FROM AUTHOR]

The generation of motor adaptation in response to mechanical perturbation during human walking is seldom considered in an exoskeleton system. Reshaping step trajectory over consecutive gait cycles for a walking exoskeleton is investigated in this article. Step adjustment of a walking exoskeleton can adapt to human walking intention by shaping step trajectory. This work develops an admittance adaptive fuzzy control strategy for a walking exoskeleton robot to provide assistance for human lower limb movement. Considering human walking intention and utilizing an admittance model, it shapes a reference trajectory to ensure that the walking exoskeleton follows it according to the human–robot force produced by its wearer. Considering a nonlinear and dynamic model with uncertainties, this work designs an integral-type Lyapunov function controller to track a reference trajectory. A disturbance observer is integrated into the controller design to compensate for uncertain disturbance in order to achieve an effective tracking performance. Finally, this work conducts experiments on two healthy subjects with the proposed method on a walking exoskeleton to validate its effectiveness. The results show that it can be applied to walking exoskeletons to enhance human mobility. [ABSTRACT FROM AUTHOR]

Soft startup is essential to the operation of the star-connected cascaded H-bridge (CHB) STATCOM. The conventional startup stage involves only uncontrolled rectifier process in which the dc voltages are charged by inserting startup resistors to limit the inrush current. Nevertheless, another inrush current is induced at the beginning of normal operation due to the low charged dc voltages during the startup stage. In this article, to eliminate this inrush current, a novel startup control is proposed for further charging the dc voltages after the uncontrolled rectifier process. During this controlled charging process, the startup resistors are not bypassed so as to smoothen the dc voltages and limit the inrush current. In this scenario, the overall dc voltage control with constant charging current is proposed by establishing a relationship between the overall dc voltage and the positive sequence modulation index. Besides, the cluster dc voltage balancing control based on the negative sequence voltage injection is proposed by redistributing the active power of the STATCOM. Moreover, the charging current is designed to reduce the active power of the startup resistors and increase the charging speed. Due to the flexibility of the startup control, it can be applied to any voltage-source-based ac/dc grid-connected converter. Finally, the performance of the proposed control is verified by the simulation and experimental results. [ABSTRACT FROM AUTHOR]

The quasi-optical mode converter (QOMC) for high-power fusion gyrotron is a typical beam transformation and shaping device in near-field region. In this letter, the authors report their progress in designing a compact QOMC with only two reflecting mirrors. For achieving the high-efficiency conversion comparable to conventional multimirror converters, the authors exercise modification and optimization on the first mirror based on the classic quasi-elliptical design. It is shown that the optimized first mirror can fulfill multifunctions that have to be achieved by a chain of mirrors in conventional designs. Numerical results show the effectiveness of the optimization, which can be revealed by the achieved 95.97% Gaussian content in the output beam with power transmission efficiency of 99.0%. [ABSTRACT FROM AUTHOR]

The method proposed in this letter for synthetic aperture radar (SAR) image classification has two main stages. In the first stage, a convolutional neural network (CNN) is trained for normal SAR image classification task. After training, the sample features can be obtained by extracting the output of middle layer in the forward propagation process of CNN. In the second stage, an end-to-end metric network is trained to measure the relations between sample features. The method proposed in this letter is tested with some of the larger targets in OpenSARShip data set which is collected from Sentinel-1 satellite, and it is also tested with the MSTAR data set which is created by the U.S. Air Force Laboratory. The experimental results show that our method can get a higher recognition accuracy than normal CNN structure. [ABSTRACT FROM AUTHOR]

A state-of-the-art 3.3-kV/450-A hybrid power module for the next generation traction inverter of rolling stock is reported in this paper, combining the silicon (Si) insulated-gate bipolar transistor (IGBT) and silicon carbide Schottky barrier diodes (SBDs) chips. Compared with the existing hybrid technology at the same voltage level, this module is characterized by a half-bridge topology, in which 6 IGBT and 12 SBD chips are integrated in each switch. The outnumbering of the diodes represents a promising mitigation to the low availability of SBDs at this voltage level. Both static and dynamic test of this module and an equivalent Si-based module are carried out comparatively. Apart from describing the features of compactness, low-inductance, and good current distribution among chips, this module is characterized by low turn-on current overshooting and turn-on loss of IGBTs, negligible diode reverse recovery time and loss, as well as flexible allowance of IGBT turn-on current rising rate $\boldsymbol{dI}/\boldsymbol{dt}$. A parameterized study is carried out to benchmark the advantage of this new topology. Based on the experimental results, the performance of the hybrid module in a three-phase traction inverter circuit is also evaluated by means of electro-thermal simulation. The hybrid module distinguishes itself by describing much lower power loss and junction temperature than its Si-based counterpart. [ABSTRACT FROM AUTHOR]

Effective and reliable machinery health assessment and prognostic methods have been highly demanded in modern industries. In the past years, promising prognostic results have been achieved by the intelligent data-driven approaches. However, the existing methods generally rely on the availability of the complete system information. In the real industries, due to practical restrictions of mechanical structure, sensor installation etc., the collected data may only cover partial system health information, which is denoted as partial observation problem. The existing data-driven methods are basically less effective in such scenarios with data incompleteness and disturbances. In order to address this issue, a deep learning-based remaining useful life prediction method is proposed in this article, where supervised attention mechanism is introduced. The informative data with more significant degradation features can be focused on for prognostics, and the data with nondiscriminative features can be ignored. A cutting wheel degradation image dataset is contributed to validate the proposed method, which is prepared from the real industrial manufacturing process. The experimental results suggest the proposed method offers an effective and promising approach on the prognostic problems with partial observations. [ABSTRACT FROM AUTHOR]