Search

In this paper, a kind of fuel cell voltage monitoring (CVM) system with negative voltage direct detection function is proposed for the reverse polarity phenomenon of fuel cell stack. This system uses optocoupler relays array to realize sampling channel switching and negative voltage conversion. The CVM prototype shows that the voltage measurement range of the proposed system is larger than that of the superimposed bias voltage method, and the sampling accuracy of cell voltage is 3mV.

The fault protection of DC microgrids restricts the popularization and development of DC grids. DC circuit breaker is an effective solution to this problem. DC circuit breaker is usually connected in series with the converter to realize the function of DC protection. But the volume and weight of DC circuit breaker is also an important factor restricting the development of circuit breaker. Therefore, how to simplify the structure of circuit breaker and realize the integration of circuit breaker and converter is one of the focuses of research. In this paper, a topology integrating boost converter and circuit breaker is proposed, to realize power conversion and fault current isolation. This topology reduces the size and weight of the device through the reuse of passive components. Firstly, the working principle of the proposed topology is analyzed, including the steady working state and the fault state. Then, the process of parameter selection is derived. Finally, the functions of power conversion and fault isolation are verified by MATLAB simulation.

At present, Solid State Power Controller (SSPC) integrated with short-circuit protection function is gradually replacing the traditional mechanical protection device and playing an important role in airborne fuel cell system (FCS). SSPC can quickly isolate short-circuit faults and avoid damage to FCS. However, the response of traditional SSPC is still slow for short-circuit protection, leading to a large fault current. In this paper, a bidirectional DC SSPC for airborne proton exchange membrane fuel cell systems (PEMFC) is proposed. This SSPC adopts a protection scheme based on the current rising rate, which is implemented with high-speed analog devices. Its simulation verification in the power supply system for a fuel-cell-powered unmanned aerial vehicle (UAV) is provided with Simulink. The results show that the protection speed of the scheme is considerably faster than the traditional protection scheme based on the current amplitude. Also, a circuit-level simulation is performed in Saber to show other advantages of the scheme, such as low energy consumption, allowing capacitive load, immunity to noise interference, and the effectiveness of FCS protection.

Joint 2D object detection and 3D reconstruction is an essential computer vision task to get more accurate detection and representation model of the target object. We proposed a novel joint 2D object detection and 3D reconstruction model that enhances the ability of the 2D object detection and the 3D reconstruction, called Adversarial Fusion Mesh Region Convolutional Neural Networks (AFM R-CNN). Our proposed model introduces the Deep Convolutional Generative Adversarial Network (DCGAN) to generate adversarial images and input the real and adversarial images into the object detection module GA-RPN to determine the position and anchor box of the target object. Next, to make better use of the two-dimensional information of the image, the voxel conversion and Fusion model Pix2Vox is introduced to fuse the two types of image features and generate coarse voxels. Afterwards, to differentiate the voxel information more efficiently, we use the Principal Neighborhood Aggregation network (PNA) model in 3D model refinement. The contrast experimental results on the open domain dataset (Pix3D) with baseline models demonstrate the effectiveness of AFM R- CNN in joint 2D object detection and 3D reconstruction task.

The resolution limitation is a main problem in the time-reversed field shaping system. In this paper, a subwavelength field shaping approach is proposed by combining time reversal (TR) technique and defective resonant metasurfaces. The subwavelength 2D and 3D field patterns can be effectively realized in both single-target and multi-target field-shaping systems. The resolution in both single-target and multi-target field-shaping systems aided with defective metasurfaces is approximately $0.1~\lambda $ , which is corresponding to the dimension of the applied metasurface cell. The proposed technique can be used for wireless power transfer, industrial heating, and localized wireless communications.

A new dual-mode bandpass filter (BPF) which is mainly formed by a ring resonator with an open stub perturbation is introduced in this article. The ring resonator inward extend in order to meet the need of the compact size. Meanwhile the orthogonal input and output feed-lines were introduced to generate coupled orthogonal modes. Two transmission zeros are generated which further improve the frequency selectivity and the isolation of the BPF. The simulated and measured results show that the proposed filter has good frequency response characteristics.

This paper designs a battery management system for the isolated DC/DC balance circuit, studies the active balance architecture based on the multi-winding flyback circuit, builds a MATLAB simulation model for balance simulation, and uses LTC6804 and LTC3300 to design prototype of a battery management system. Tests show that the sampling accuracy of battery cell voltage can reach 2mV, the balance current can reach 2A, and the management system can effectively perform balance control.

This article introduced an active frequency selective surface (AFSS) in which the state of the unit cell can be controlled independently. The structure of the cell consists of a periodic metal pattern printed on the dielectric substrate and a bias circuit. The PIN diode is applied in the metal pattern. In addition, This paper also analyzes the influence of metal pillars produced by bias circuit on resonance characteristics. It works as a absorber in the X band when the diode is fully on, and works in the S-band when the diode is completely off.

The recently emerging beyond fifth-generation (B5G) and sixth-generation (6G) mobile communications demand flexible phase shift controlling ability to enable its multi-channel multi-functional transmit/receive systems. In this work, slow-wave half-mode substrate integrated waveguide (SW-HMSIW) patterned with non-uniform polylines of various lengths for flexible phase shifting application is presented. By periodically patterning polylines into the conventional HMSIW, product of the equivalent permittivity and permeability of this composite medium can be enhanced effectively, thus to influence the phase constant of the transmissible waves in it, and the slow-wave effect and associated size reduction will be achieved. Hence, by varying lengths and widths of segments of the polylines differently, the equivalent permittivity and permeability of the SW-HMSIW can be changed, and consequently its phase constant will be differently as well, namely phase shift function is realized. To demonstrate this idea, a size-miniaturized broadband SW-HMSIW phase shifter is implemented. Measured results agree with the simulation well, which indicates its suitability for phase controlling applications in future communication systems.

In this paper, a neural network-based adaptive impedance control (NNAIC) scheme is proposed for controlling electrically-driven robotic systems. The neural network (NN) is used to compensate the robotic model uncertainties and the motor dynamic uncertainties, and a robust compensator term is derived to compensate the disturbances and approximation errors of the NN. In this way, the performances of the joint positions and force tracking can be then improved. Based on the Lyapunov stability theorem, it is proved that the control system is stable and all the signals in closed-loop system are bounded. Simulation tests on a 2-link electrically-driven robotic manipulator are presented to show the effectiveness of the proposed intelligent impedance control method.

The application of the knowledge graph in computer vision is a new trend in deep learning. Facial video-based emotion analysis and recognition are critical topics of research in the mental healthcare field. In this paper, we proposed a novel noncontact intelligent framework to represent the knowledge of facial features and heart rate (HR) features for predicting the emotional states of objects. The framework is divided into two parts: knowledge modeling and knowledge reasoning. In the first step of knowledge modeling, 3D-CNN is utilized to model the spatiotemporal information from the facial and forehead regions based on the remote photoplethysmography technique, separating the blood volume pulse (BVP) signal and extracting the HR from the forehead image sequence. Finally, the multichannel features are integrated and transformed into structured data and put into the knowledge graph as much as possible. Knowledge reasoning is an inferential process that associates the deep learning model with structured knowledge to predict continuous values of the emotional dimensions (pleasure, arousal, and dominance) from facial videos of subjects. Experiments conducted on the DEAP database demonstrate that this approach leads to improved emotion recognition performance and significantly outperforms recent state-of-the-art proposals. The result proved that prior knowledge from the knowledge graph ground truth on deep learning is an efficient means of emotional recognition in vision modality. Our artificial intelligence models can be popularized and applied in daily healthcare.

Chinese liquor culture has a long history. The traditional liquor-making technology belongs to the labor-intensive industry with the characteristics of high labor intensity and low production efficiency. Especially in the process of feeding for liquor distillation, factitious material filling method is laborious and inefficient. Therefore, an intelligent control system is helpful for the process of material filling. The accuracy of steam detection result is the key factor determining the precision of adding material to the caldron. Corresponding to the characteristics of the moving robot arm and the light color and easy diffusion of the detected target steam in the process of adding material to the caldron, a steam detection method for adding material process is proposed in this paper. In the steam detection method, Firstly, suspected steam areas are detected by extracting motion characteristics, then, a feature fusion based BP neural network model is proposed to detect the steam in the suspected area. The proposed method is applied to the material filling process and the effectiveness is demonstrated.

In this paper, a high-frequency frequency scanning antenna based on metasurface is designed. The antenna proposed here consists of three parts: an 1 × 16 unequal divider, 16 BJ900 waveguides and 160 metasurface. The proposed antenna is fed by a standard waveguide BJ900 with periodic metasurface covered on it. The working band is 76–81 GHz. To realize two-dimension beam scanning, an unequal divider is designed. After the final optimization, we can get from the simulation results that the working angle is 30° in both two dimension and the radiation efficiency can reach -0.1 dB. The total realized gain of the antenna can reach 20 dB.

Energy consumption braking is widely used in the field of air winch. When the BLDCM (Brushless Direct Current Motor) is in the reverse braking state, the energy-consuming braking circuit reduces the bleeder power of the bleeder resistor and the volume of the energy-saving brake bleeder resistor by adjusting the duty cycle of the switching tube. In this paper, the Matlab simulation model is used to verify that reducing the duty cycle can reduce the heating power of the bleeder resistor to reduce the volume of the bleeder resistor, thereby increasing the controller power density.

Myoelectric control based on pattern recognition technology is largely influenced by intrinsic and extrinsic factors over time in practical application, which often result in performance degradation. The aim of this study is to develop an adaptive classification scheme based on linear discriminant analysis (LDA) for myoelectric control system. The proposed adaptive LDA classifier based on training dataset selection and probability weighting is applied to improve the performance of myoelectric control system in changing environment. Preliminary study on the long-term use in real clinical implementation is conducted to verify the performance of proposed adaptive LDA classifier. The correct recognition rates of motion recognition of the proposed LDA classifier and conventional LDA classifier are compared. The experimental results show that compared with the conventional LDA, the proposed LDA significantly improves the classifier performance. Experiment results show the proposed scheme is a practical approach for developing myoelectric control systems with strong robustness.

In radar signal processing, the detection and parameter estimation of high-speed maneuvering targets, which often utilize a coherent pulse train signal with linear frequency modulation, have been receiving increasing attention. Fundamentally and significantly, the Cramer–Rao lower bound (CRLB), as a cornerstone for evaluating the estimation performance of high-order kinematic parameters, has been derived and investigated here. In this paper, a 2-D echo signal model expressed in the fast-frequency and slow-time domains is adopted. The scaled orthogonal Legendre polynomials are deliberately introduced to solve the inverse problem of the Fisher information matrix, and then, a linear mapping relationship between different polynomial parameters can be used to obtain the analytical CRLB expressions. The main contributions included are: 1) the CRLBs, which are exact and closed form, have been extended to arbitrary motion model orders and reference time instants; 2) the influences of the motion model order, the reference time instant, as well as the radar parameters on the CRLBs are exploited comprehensively; and 3) some specific cases, including four low-order motion models and two preferred reference times, are also presented to better demonstrate the CRLB performance relationships. It highlights the fact that the reference time instant corresponding to the middle of the pulse train is a reasonable and compromised choice for parameter estimation, although it is not necessarily optimal for the kinematic parameters of all models and orders. The above research results are illustrated with numerical simulations and further verified using the maximum likelihood estimation method combined with Monte Carlo experiments.

As a novel financing method, peer-to-peer (P2P) lending has drawn extensive attention as it provides those financers who cannot participate in the traditional financial market with funds. In P2P lending marketplaces, one of the crucial challenges that P2P online lending platforms are facing is to accurately predict the default risk of each loan by tapping into default prediction models, thus effectively helping P2P lending companies avoid credit risks. That traditional credit risk prediction models fail to meet the demand of P2P lending companies for default risk prediction, which is because of the uneven distribution of credit data samples in the P2P lending marketplaces (i.e., the default sampled data are scarce). In this paper, we designed a multi-round ensemble learning model based on heterogeneous ensemble frameworks to predict default risk. In this model, an extreme gradient boosting (XGBoost) is initially used for ensemble learning, and the XGBoost, deep neural network, and logistic regression are then regarded as heterogeneous individual learners to undergo a linear weighted fusion. To verify the designed default risk prediction model, real credit data from a famous P2P online lending marketplace in China were used in a test. The results of the experiment indicate that this model can effectively increase the predictive accuracy compared with traditional machine learning models and ensemble learning models.

A time-reversal imaging system for far-field super-resolution imaging of scatterers is proposed. First, a grating plate, similar to the optical far-field super lens, is designed to convert near-field evanescent waves to propagative waves with wavevector modulation. Then, the scattered signals are processed by a time-reversal (TR) technique and retransmitted back. Finally, the images of the targets are reconstructed by doing demodulation operations in the k-domain to the TR field at the refocusing time. Imaging results show that two scatterers separated by 0.1 λ can be distinguished clearly.

Minimally invasive surgery (MIS) is a surgical treatment with small scar and quick recovery. In MIS, an endoscope is usually carried by surgeon or assistant to observe surgical instruments and focal points. This paper proposes a new camera tracking system to help surgeons observe the operating environment in MIS. Held by the mirror-holding robot, the camera follows the specified tool and focal points intelligently and autonomously. Based on the target in the picture taken by the camera, the system adjusts the camera position, which is dynamic operation field tracking. It helps surgeons observe the operation in real time. The tracking algorithm can ensure the accuracy, safety and real-time tracking of the motion of the camera. The simulation results verified the effectiveness of the algorithm.