Search

In order to satisfy higher requirements for data transmission rate and system capacity, millimeter wave (mmWave) with abundant spectrum resources has be exploited and become a research hotspot of the fifth generation (5G) mobile communication systems. However, introducing mmWave technologies into cellular networks also faces many severe challenges, especially the initial user equipment (UE) discovery. In the UE discovery, the narrow-beam directional scanning is often used to search the UE, but it increases the control overhead of the system and brings about a non-negligible search delay. In this paper, we will analyze the average search delay of mmWave systems under two different UE access methods, i.e., the random scanning access method in independent networks and the search method based on UE location in heterogeneous networks. Taking the average search times discovering UEs as a measure and the number of search sectors in beam-space as an optimization parameter, the optimal number of sectors with the maximum capacity is obtained so that the search delay and the capacity can be balanced. The simulation results verify the correctness of the analysis and show that the performance of the user discovery method in heterogeneous networks is better than that of the random scan access method in mmWave independent networks.

The online monitoring of pointing error (PE) caused by the main reflector (MR) deformation under wind disturbance has become a knotty problem for large reflector antennas. This paper presents a real-time estimation algorithm for monitoring wind-induced PE by optimized acceleration measurement points. First, based on an unbiased minimum-variance input-state estimation theory via accelerations measurement, we set up the PE observation model which illustrates the relation between the structural vibration state and PE. Moreover, compared with the frequency domain integration approach, the non-integral observation model without introducing a low-frequency-cut filter can sense pseudo-static and low-frequency PE. Second, a sensor placement method considering sensor constraint spacing (SPMCSCS) is proposed to achieve a high-quality monitoring effect with low information redundancy. In the method, the constraint spacing between sensors is determined by a novel scan recognition algorithm in which regional division, and regional correlation analysis of the modal shape are carried out successively. Finally, the simulation implementation of a 7.3-m antenna under a wind disturbance is used to illustrate the algorithm.

Real-time data collection and analysis in large experimental facilities present a great challenge across multiple domains, including high energy physics, nuclear physics, and cosmology. To address this, machine learning (ML)-based methods for real-time data compression have drawn significant attention. However, unlike natural image data, such as CIFAR and ImageNet that are relatively small-sized and continuous, scientific data often come in as three-dimensional data volumes at high rates with high sparsity (many zeros) and non-Gaussian value distribution. This makes direct application of popular ML compression methods, as well as conventional data compression methods, suboptimal. To address these obstacles, this work introduces a dual-head autoencoder to resolve sparsity and regression simultaneously, called \textit{Bicephalous Convolutional AutoEncoder} (BCAE). This method shows advantages both in compression fidelity and ratio compared to traditional data compression methods, such as MGARD, SZ, and ZFP. To achieve similar fidelity, the best performer among the traditional methods can reach only half the compression ratio of BCAE. Moreover, a thorough ablation study of the BCAE method shows that a dedicated segmentation decoder improves the reconstruction., Comment: 6 pages, 6 figures

In the last decade, 3D printing technology has demonstrated a number of merits, including time- and labor-saving, capability to fabricate complicated shapes, and diversity of materials choices. These advantages offer potential for rapid prototyping of advanced antenna designs, particularly for designs in the microwave region due to the compatibility between antenna dimensions and the printing resolution of 3D printers. Antennas with circular polarization (CP) have been employed as important components in wireless systems, due to their superiorities of mitigating multi-path fading, avoiding polarization mismatch, and providing flexibility in the orientation of mobile antennas. Meanwhile, wideband end-fire antennas with CP are in high demand in various applications, such as satellite and vehicle communications.

To build an actual cloud energy storage system by blockchain for the ancillary service, this paper presents a prospective engineering planning method and design process to build a platform with five functions of cloud energy storage system. To demonstrate the feasibility, the engineering planning method includes the following steps. First, the detail design drawings of planning and feasibility analysis are introduced. The feasibility analysis including the project’s necessity, project alternatives, project impact, design drawings, construction scale are needed. Then, the device hardware design of blockchain platform with five functions is presented. Thirdly, the main material and equipment list, and the total investment are given. Finally, the post project evaluation is introduced. The functions include the peak load regulation, unit primary frequency regulation and automatic generation control, black start, reactive power and voltage regulation, reserve ancillary service for the integrated energy system. The paper can be referenced for planner.

Smoke recognition has been actively studied in the computer vision domain. Due to large variance of smoke color, texture and shapes, smoke recognition is a challenging task. Traditional smoke recognition methods are based on handcrafted features. In the past few years, some methods which are based on convolutional neural networks have been proposed that achieved great improvement of smoke recognition. However, previous methods cannot capture the internal structure of smoke well. Manifolds can represent the internal characteristics of smoke data in lower dimensions, reduce the redundancy of data representation, and obtain more discriminative capabilities. In this paper, an end-to-end deep convolutional neural network which based on manifold structure was proposed to capture more discriminative features for smoke recognition. Experimental results show that the proposed method achieves promising results on the public smoke recognition dataset. The proposed method can obtain high detection rate, high accuracy rate and low false alarm rate in the same time.

In marine mapping, the sound speed profile (SSP) is an essential parameter. The SSP reflects the characteristics of the ocean sound field’s vertical structure. Furthermore, analyzing the SSP has an important significance on the propagation of underwater sound signals. The empirical orthogonal function (EOF) can extract the SSP data’s main features. It is widely used in SSP fitting and other applications. In this paper, the studying area is 116°~117°E,18°~19°N in the South China Sea. Using the Argo data from 2010 to 2020, this paper first visually analyzes SSP, carries out the EOF’s analysis, and then elaborates the basic EOF principle. It confirms that the EOF has a good application value for the South China Sea, and the first sixth order function can accurately describe the sound speed profile. Then this paper analyzed the EOF’s temporal function. The study found that the third mode of temporal function strongly connects with the ocean’s climate condition. And the two severe fluctuations in 2010 and 2014 have a strong relationship with the typhoon passage. In other words, the degree of extreme fluctuations is positively correlated with typhoon intensity. In brief, understanding the SSP’s temporal pattern is beneficial for applying various engineering scenarios. It can improve the accuracy of sound speed correction of marine acoustic instruments, while the study of marine climate environment has an important role.

This paper presents a flexible broadband power divider based on defect compensation that works in the 3-12GHz frequency band. In the design, the cascaded multi-section impedance converter is employed to expand the working bandwidth, at the same time, additional capacitance and inductance are introduced through the defective ground structure (DGS) to compensate for the performance of the splitter caused by bending. This paper analyzes the transmission characteristics of the flexible power divider in different directions and at different angles through numerical calculation and modeling simulation. The results show that the proposed compensation design scheme achieves a maximum insertion loss of -3.1dB within 3-12GHz, the flatness of the most in-band is less than 0.7dB, the isolation is

Nowadays, plug-in electric vehicles (PEVs) not only have on-board level-1 or level-2 chargers for home or office charging, but also off-board level-3 chargers for high-power fast charging gradually. This paper presents an integrated multifunctional battery charger, which merges on-board charger (OBC), auxiliary-battery dc-to-dc converter (ADC) and motor driving inverter (MDI) into a single system. The proposed topology can achieve four operating functions by sharing switching devices, magnetic component and capacitors: charging the traction battery by linking to the three-phase grid [grid to vehicle (G2V)], the traction battery supplying surplus energy to the three-phase grid [vehicle to grid (V2G)], charging the auxiliary battery by the traction battery [traction battery to auxiliary battery (T2A)] and the traction battery discharging to drive the motor [traction battery to motor (T2M)].

The structure and dynamic process characteristics of furnace temperature control system of billet heating furnace are analyzed, and the idea of adaptive control without model is put forward, and the corresponding temperature controller is constructed. In the process of furnace temperature control, the nonlinear, hysteresis and time-varying characteristics of the system are analyzed. In the process of heating and cooling control, the idea of fuzzy control is used to effectively overcome the difficulty that the system parameters are not suitable to cause the temperature fluctuation in the process of temperature rise and fall. The simulation results show that the controller can control the furnace temperature well.

In the field of ear shape clustering, 3D ear modeling and ear related personal product implementation, it is of great significance to retrieve the accurate position of some key physiological curves and key points. Traditional edge extraction method is very sensitive to lighting and pose variations. A method for extracting the key physiological curves of the human ear based on an improved instance segmentation network YOLACT is proposed in this paper. By labeling the human ear dataset, the YOLACT model is trained. By improving the network prediction process, it can accurately segment different regions of the human ear and extract key physiological curves. Compared with the original YOLACT and other segmentation methods, it shows higher segmentation accuracy on the test dataset and can extract more accurate human ear physiological curves. The proposed method also shows more robustness under lighting variation, pose variation and slight occlusion.

The Snake active contour model converts the image segmentation problem into solving the energy functional minimum problem, and uses curve evolution to locate the edge of the target. But the disadvantage is that it depends on the position of the initial contour. In order to improve the defect that the initial contour of the traditional Snake active model is difficult to locate when it is applied and the inaccurate extraction results when the initialization curve is far away from the target edge, follow the steps below:The initial curve is determined by the inscribed circle generated by the Hough transform algorithm in the area where the target is located; According to the different area where the point on the curve located and the point’s neighborhood characteristics, the direction of the point movement on the curve and the automatically changed Step size. The above steps can improve the defect of inaccurate extraction results when the initialization curve is far away from the target edge. The results obtained by the above method are compared with the traditional Snake model segmentation results to verify the feasibility and accuracy of the improved algorithm.

It is difficult to detect grasping position of irregular objects automatically which always leads to the failure of grasping task in the current arm robot application scenario. This paper proposes an adaptive grasping position detection method that uses the modified YOLO to recognize and detect contour of an image of the object which sampled by a camera. The detection contour results are then fed into a special regression network trained by prepared datasets to detect the grasping position of the object. The experimental results have shown that the grasping position detection method has good robustness and environmental adaptability and can meet the most needs of general applications in practical working condition.

Traditional methods of inverse kinematics of robots always adopt analytical approach and numerical approach to solve the continuous state and action problems with experience and experiment mostly, which require much time and work in reality work scene, especially for robots with complex structure. This paper proposes a method based on reinforcement learning TD3 network, which is constructed by PyTorch to find the inverse solution from another point of view. A set of improved distributed multiple rewards which choose the position difference between adjacent joints as the reward standard are designed to optimize the solution, avoid solving unreachable points and prevent the mechanical structure from being damaged also in the environment of five-degree-of-freedom arm robot. The validity of above method is verified by simulation experiment results.

People used to take a series of nearly identical photos about the same subject, but it is usually a tedious chore to select the reversed ones from them. Despite the remarkable progress, most existing studies on image aesthetics assessment fail to fulfill the task of series photo selection. In this paper, we develop a novel deep CNN architecture that aggregates multi-scale features from different network layers, in order to capture the subtle differences between series photos. To reduce the risk of redundant or even interfering features, we introduce the spatial-channel self-attention mechanism to adaptively recalibrate the features at each layer, so that informative features can be selectively emphasized and less useful ones suppressed. Extensive experiments on a benchmark dataset well demonstrate the potential of our approach for series photo selection.

The proper segmentation of breast lesion based on mammography is crucial not only for traditional computer aided diagnosis but also for supervised learning by deep learning, which is significant for further quantitative analysis. To achieve a comprehensive diagnosis with versatile deep learning algorithms, a novel Docker based framework is proposed and designed by incorporating popular CNN based models CDeep3M. The validation is preceded under a unique amount of B-type ultrasound images. The results indicate that CDeep3M have a good capability for breast tumor prediction. Besides, the analysis shows that Docker framework is flexible and extensive for environmental construction for breast cancer real-time prediction, as well as other related applications.

With the fast-growing of fleet size and the railway mileage, the energy consumption of trains is becoming a serious concern globally. The nature of railway offers a unique opportunity to optimize the energy efficiency of engines by the aid of various landforms along railway lines. However, because of the feature of non-linearity, high-dimension, complicated-constraint and time-variant, it is difficult to get the optimal train-driving scheme that saves the most energy. This paper proposes a dexterously augmented CNN model called sequential CNN(S-CNN) to generate automatic train-driving schemes. Forward and backward features are defined to depict the context of the running train and they are processed separately in the input layer. Operations of experienced train drivers are used for the CNN model training. The S-CNN model takes both the time-series and space-series into consideration and pays attention to their ordinal attributes. The proposed techniques are validated in a Hardware-in-Loop platform using realistic data. Experimental results proved that the driving operations obtained by S-CNN model are very similar to experienced human drivers and it saves more than 10% energy when compared with the average performance of human drivers.

In contemporary society, railway networks are usually considered among the most critical undertakers of transportation tasks. Thanks to the born-with preferential conditions of the railway systems, in particular the settings and circumstances in which they run, railroad trains are best positioned to become automatic, vis–vis other transportation systems, such as road traffic. Having said that, one must admit that the transformation towards intelligent man-less train driving system will not be easy, given the requirements in high dimensionality, non-linearity, multi-origin restraints, and other factors that may change over time. Challenges, including energy efficiency, on-time performance, and safety, need to be addressed. The authors of the present paper used a deep reinforcement learning (DRL) method to generate energy-optimized train driving solutions. The neural network representation is based on a Bidirectional Long-Short-Term Memory (BLSTM) and an automatic encoder, which integrate sequential properties of the time-series and inherent non-sequential attributes. It is believed that the learned model has value in assisting decision makers in some representative cases of industrial control, as experiments show that this is perfectly adaptable to meet the basic safety, energy saving, and punctuality requirements in train driving.

Quadrotors have developed rapidly due to low-cost, precise hovering and the ability of vertical take-off and landing (VTOL). As the crucial part of quadrotors, the control system has been researched extensively. In this paper, we present a PID controller for path tracking based on a double-loop controller where the inner loop is for attitude control and the outer loop is for position control, and then combine fuzzy logic and the PID controller to improve the controlling effect. However, the fuzzy rules based on experience cannot guarantee excellent tracking effect and it is difficult to adjust the fuzzy rules manually. So we employ particle swarm optimization (PSO) to optimize fuzzy rules. Finally, we testify the effectiveness of the PID controller by simulation, and show that the fuzzy PID controller with PSO optimization achieves better effect compared with basic PID controller and the fuzzy PID controller without the optimization of fuzzy rules.

This paper presents an adaptive robust control approach for heterogeneous vehicular platoons subject to nonlinear and (possibly fast) time-varying uncertainty. The uncertainty is unknown and bounded. The vehicles in the platoon are modeled as nonlinear uncertain dynamic systems. To keep compact formation and guarantee collision avoidance, two-side inequality constraints are introduced for the spacing error between adjacent vehicles. A mathematical transformation scheme is proposed to convert the bounded state into an unbounded one. Based on backstepping method and Lyapunov stability theory, an adaptive robust controller is designed. A leakage type adaptive law is utilized to identify the bound of uncertainty. Furthermore, the controller renders the uniform boundedness and uniform ultimate boundedness performance for the unbounded state, which in turn guarantees the two-side restrictions for the spacing error. Finally, simulations are performed to illustrate the efficiency of the proposed algorithm.