Search

Fatigue cracks usually occur at the welded joints of the heavy haul wagon body. To effectively limit the occurrence of such problems at the design stage, this study moves from the structural design strategy to the structural optimization method. First, when analyzing the results of the finite element simulation, the new structural design strategy not only considers the maximum stress, which can reflect the stress concentration position, but also focuses on the stress distributions at the key welded joints. Second, in order to further reduce the stress at the key positions, the study presents an optimization method that integrates an orthogonal experimental design, BP network, and MPSO algorithm. Finally, an 80 t gondola car body is optimized from the whole structure to the local area by the proposed method. Further, it is also verified that the structural optimization method of the heavy haul wagon body is reasonable and feasible.

Background: Correct classifying of analog circuit faults is helpful in the health management of the circuit. It is difficult to be implemented because of the lack of proper feature extraction methods and accurate fault diagnosis models. Objective: T-SNE based core components extraction method and PSO-ELM-based fault diagnosis model are presented to improve the diagnostic accuracy of analog circuit fault diagnosis. Methods: Firstly, circuit output signals are collected, and they are transformed to wavelet coefficients. Then, the high-dimensional wavelet coefficients are processed by t-SNE to generate lowdimensional core components as features. The Extreme Learning Machine (ELM) based diagnosing model is constructed by using the features, and the key parameters of ELM are optimized by using Particle Swarm Optimization (PSO) algorithm. Finally, the constructed PSO-ELM diagnosis model is employed to identify different analog circuit faults. Results: Leapfrog filter circuit and three-phase bridge circuit fault diagnosis experiments are implemented to demonstrate the proposed t-SNE based features extraction method and PSO-ELM based fault diagnosis model. Also, comparisons are performed to verify the high performance of proposed fault diagnosis methods. Conclusion: The proposed t-SNE based core components extraction method and PSO-ELM diagnosis model are effective to improve the fault diagnosis accuracy of the analog circuit.

Vehicle lateral velocity is critical for the high-level automatic driving technology, andyyfusing Global Position System (GPS) in the lateral velocity estimation method can greatly improve the estimation accuracy. However, the method fusing GPS is seldom reported, since the problem of GPS-outage often appears. Accordingly, this article proposes a novel lateral velocity estimation method (VLVEM) based on SBI-LSTM in GPS-outage environment. Additionally, VLVEM integrating Inertial Navigation System-aided GPS (INS-aided GPS) is derived from Federated Kalman Filter (FKF) algorithm. Furthermore, during GPS-outage, induced from the Stack Bidirectional Long Short-Term Memory Recurrent Neural Network (SBI-LSTM RNN), an INS-aided GPS fault reconstructor (IGFR) is designed to reconstruct INS-aided GPS model. Finally, the simulation results show that compared with most of the existing methods which only consider the in-vehicle sensors’ signals, the proposed method fusing GPS has higher lateral velocity estimation accuracy. Besides, when GPS-outage causes INS-aided GPS failure, IGFR can reconstruct INS-aided GPS model, and VLVEM still has high estimation accuracy. Combining the in-vehicle sensors and GPS, VLVEM exhibits great robustness and fault tolerance.

BeiDou navigation satellite system transmits triple-frequency signals. This paper reviews the disadvantages of the traditional least-squares ambiguity decorrelation adjustment algorithm in the ambiguity resolution and exploits the advantages of triple-frequency BDS. A new ambiguity resolution method that is based on adaptive ionosphere estimation constraints for medium-length baselines (from 50 to 100 km) is proposed. We realize the method in two steps. First, double-differencing uncombined observables (B1/B2/B3 observables) are used to obtain float solutions with atmospheric delay estimated as a random walk parameter by using the extended Kalman filter. Second, two ambiguity search equations are constructed between any two uncombined ambiguities. The adaptively estimated ionospheric delay shown in step one is also used to constrain ambiguity search equations. Finally, traversal search of integer ambiguity is performed. The method is decreasingly affected by the cycle slips and reference satellite switching, which can effectively avoid the unstable positioning accuracy of the carrier in the RTK positioning of medium-length baselines. In this study, BDS triple-frequency data collected from two baselines of 53 km and 103 km, respectively, are used to validate the proposed method. Experiments show that the medium-length baseline ambiguity can be rapidly fixed within one single epoch. Although the baseline is over 100 km, the accuracy of coordinate component under dynamic conditions can reach the centimeter level.

Cancer cell genomes originate from single-cell mutation with sequential clonal and subclonal expansion of somatic mutation acquisition during pathogenesis, thus exhibiting a Darwinian evolutionary process (Gerstung et al., 2020; Nik-Zainal et al., 2012). Through next-generation sequencing of tumor tissue, this evolutionary process can be characterized by statistical modelling, which can identify the clonal state, somatic mutation order, and evolutionary process (Gerstung et al., 2020; Mcgranahan & Swanton, 2017). Inference of clonal and subclonal structure from bulk or single-cell tumor genomic sequencing data has a huge impact on studying cancer evolution. Clonal state and mutation order can provide detailed insight into tumor origin and future development. In the past decade, various methods for subclonal reconstruction using bulk tumor sequencing data have been developed. However, these methods had different programming languages and data input formats, which limited their use and comparison. Therefore, we established a web server for Clonal and Subclonal Structure Inference and Evolution (COSINE) of cancer genomic data, which incorporated twelve popular subclonal reconstruction methods. We deconstructed each method to provide a detailed workflow of single processing steps with a user-friendly interface. To the best of our knowledge, this is the first web server providing online subclonal inference based on the integration of most popular subclonal reconstruction methods. COSINE is freely accessible at www.clab-cosine.net.

In order to screen the aroma characteristics of soybean, a new method was established which can quickly quantify the content of 2-acetyl-1pyrroline (2-AP), an important compound related to soybean aroma, using gas chromatography-mass spectrometry (GC-MS). Based on peak profile, total peak area and retention time as test indexes, an accurate identification method for fragrant soybeans was established. The optimum parameters of the protocol consisted of column temperature 70 degrees C, sample injector temperature 180 degrees C, optimum extraction alcohol content 1 mL, NaCl content 0.1 g, ultrasonication time 10 min, and extraction time 1 h, which were established by using the orthogonal test of single factors and three factors with four levels (L-9(3)(4)). 2-AP content of leaves had significant correlations with seeds, which were easier to measure. The protocol was simple and easy to carry out, consumed only small amounts of reagents, and provided accurate and reliable results with good reproducibility. A total of 101 soybean genotypes from different geographical sources were analyzed using this protocol. The results showed that the average content of 2-AP was 0.29 mg L-1, ranging from 0.094 to 1.816 mg L-1, and the genetic diversity index was 0.54. Among all genotypes-tested, they were classified into three grades, including seven elite genotypes identified as "grade one fragrant soybeans", which were Zhonglong 608, Heinong 88, Ha13-2958, Hongmiandou, Heinong 82, Huangmaodou, and Jiyu 21. These results provide both an identification technique and several elite aroma genotypes for gene discovery and good quality breeding in soybean. National Key R&D Program for Crop Breeding [2016YFD0100201]; Scientific Innovation Program of of the Chinese Academy of Agricultural Sciences (CAAS) Published version This work was supported by the National Key R&D Program for Crop Breeding (2016YFD0100201) and Scientific Innovation Program of of the Chinese Academy of Agricultural Sciences (CAAS) . We thank the Core Facility Platform, Institute of Crop Sciences, CAAS for their assistance with GC-MS analyses.

Network virtualization makes it possible to manage multiple virtual networks simultaneously on substrate physical networks. Virtual network embedding (VNE) is the critical step of network virtualization that maps virtual network requests to substrate physical networks. The majority of current virtual network embedding algorithms utilize heuristic algorithm, and manually customize a series of rules and assumptions. Therefore, these experimental results are not particularly convincing. This paper proposes a reinforcement learning based dynamic attribute matrix representation (RDAM) algorithm for virtual network embedding. The RDAM algorithm decomposes the process of node mapping into the following three steps: (1) static representation of substrate physical network. (2) dynamic update of substrate physical network. (3) Reinforcement-Learning-Based algorithm. To our best knowledge, RDAM algorithm is the first algorithm to apply spectral analysis and perturbation theory to virtual network embedding. Meanwhile, the method training virtual network embedding algorithm by reinforcement learning is also non-trivial. Furthermore, we compare RDAM algorithm with three other virtual network embedding algorithms. The results show that RDAM algorithm outperforms the other three algorithms in terms of several evaluation metrics, such as long-term average revenue, long-term revenue consumption ratio, and acceptance ratio.

In this study, we propose a three-dimensional (3D) forward modeling algorithm of surface-to-borehole transient electromagnetic (SBTEM) fields based on an unstructured vector finite-element method to analyze the characteristics of SBTEM responses for complex geoelectrical models. To solve the double-curl diffusion equation for the electric field, we use an unstructured tetrahedral mesh to discretize the model domain and select the unconditionally stable backward Euler scheme to discretize the time derivative. In our numerical experiments, we use a grounded wire as a transmitting source. After validating the algorithm’s effectiveness, we first analyze the diffusion characteristics and detectability of the electromagnetic field. After that, we focus our attention on the distribution and the cause of zero bands for Ex and dBy/dt components with the hope of guiding future field surveys. Finally, by simulating different models, we analyze the capability of the SBTEM method in detecting typical mineral veins so that we can provide a reference for mineral resource exploration in the deep earth.

This study focuses on position tracking control for the networked predictive motion control system with random communication delay. First, the output feedback controller is designed by networked predictive control law to actively compensate the time delay induced by the random channels of the motion control system. A closed-loop model is established for the networked predictive motion control system with random bounded communication delay, modelled by a Markov chain. Then, the sufficient conditions of stability for the networked predictive motion control system are provided, by constructing the Lyapunov–Krasovskii functional, followed by the theoretical proof. Furthermore, the output feedback controller is constructed and the linear matrix inequality method is applied to obtain the designed controller gain. Last, the simulation and experimental results are presented to prove the effectiveness of the proposed method.

Based on the STM32 library function, this design is developed a good docking with the engine, without changing the engine's original ECU. Only on the test bench, this design which connected to the device can enter the injection pulse width, so as to control the engine mixture concentration. Using this experimental device, the fuel adjustment characteristic experiment was completed on the existing engine test bed, which verified that the experimental device developed by this design had good adaptability in use. In this way, universities and research institutes can use this experimental device to easily complete the fuel adjustment characteristic experiment, and can also provide a convenient tool option for the calibration and optimization of the engine pulse spectrum.

Flip-flow screen is a key equipment for dry screening of wet fine coal, and its working load is of great significance to the screening effect and reliability. In order to obtain material mass load ...

In order to achieve the high-precision positioning of underwater target from the close range, in this paper, a special combination of the linear array, which is composed of multiple mutually perpendicular sub-arrays, is proposed to overcome the problem of port and starboard ambiguity of a single linear array. And based on this special combination, an optimized MVDR near-field high resolution positioning method is proposed. Based on the matrix decomposition principle of MUSIC method, the covariance matrix of data received by the array is decomposed to obtain the signal subspace corresponding to the signal component and the noise subspace orthogonal to the signal component. Using the orthogonality of the two subspaces, the covariance matrix of the array output is optimized, and the optimized covariance matrix is used for MVDR focused beamforming to further improve the focusing and positioning performance of MVDR. Theoretical analysis and simulation results show that the combination of the special linear array can effectively overcome the fuzzy problem of target location. The positioning performance of the MUSIC near-field positioning method is higher than that of MVDR focused beamforming positioning method. The further processing of the covariance matrix can improve the performance of MVDR focused positioning method. The sea test further proves the effectiveness of the method.

The detection and recognition of underwater preset structure is an important topic in the detection of potential safety hazards in underwater engineering. At present, the detection of underwater preset structure is mainly based on multi-beam sounding technology. In this paper, the submarine pipeline is taken as the research object, and a new detection method based on 3D real-time imaging sonar technology is proposed, which realizes the underwater environment acoustic high-precision real-time target detection and fine target analysis. First of all, the underwater target detection data is obtained by using 3D imaging sonar, high-precision integrated navigation, tide gauge, CTD and other equipment. Secondly, the attitude correction, sound velocity correction and depth compensation of 3D imaging data is carried out, and the bilateral filtering combined with threshold filtering is used to filter the data. Finally, combined with advanced point cloud data processing algorithms and edge data extraction based on eigenvalues, accurate identification, positioning and status analysis of submarine pipelines are achieved. The results of practical engineering detection show that this method has the advantages of high resolution, real-time imaging and excellent imaging quality. The detection image is stereoscopic and clear, which can provide more detailed description of underwater target contour. The data processing and analysis method adopted in this paper can track and identify the submarine pipeline quickly and accurately, and detect the areas with potential safety hazards such as exposed or suspended pipelines, which provides scientific basis for the timely management of pipelines. This paper provides a new detection method for submarine pipeline detection, which has applicability and popularizing value.

In the power distribution Internet of Things environment, the execution efficiency of user tasks is low when smart node resources are limited. In order to solve this problem, this paper designs an edge computing network architecture under the power distribution Internet of Things environment. The architecture includes three types of devices: smart nodes, network devices, and edge computing nodes. The execution mode of the computing task of the smart node is modeled, the local computing model and the remote computing model are designed, and the task execution time model and energy consumption model are established respectively, as well as the objective function of the edge node resource allocation problem. The four key elements of state space, action set, reward function and search mechanism for solving the optimal resource allocation problem under Q- Learning theory are designed. A Q-Learning-based edge node resource allocation algorithm is proposed, and it is verified that the algorithm in this paper can improve the execution efficiency of user tasks.

The monitoring data of slewing bearing is massive. In order to establish accurate life prediction model from complex vibration signal of slewing bearing, a life prediction method based on manifold learning and fuzzy support vector regression (SVR) is proposed. Firstly, the multiple features are extracted from time domain and time-frequency domain. Then isometric mapping (ISOMAP) is used to reduce high-dimensional features to low-dimensional features that can reflect degeneration of slewing bearing well. Finally, the fuzzy SVR is used to predict the life degradation trend of slewing bearing. The results show that: (1) Multi-feature fusion after ISOMAP can obtain more comprehensive degradation indicator. (2) The complexity of the life prediction model is simplified and the real-time life degradation trend of slewing bearing can be well predicted by fuzzy SVR, so it is very suitable to predict life degradation trend of slewing bearing based on massive data well. The time of prediction on average is reduced by 72.7%. The mean absolute error (MAE) and root mean square error (RMSE) of prediction are reduced by 73% and 59% respectively compared with traditional methods. The accuracy of prediction is greatly improved.

The traditional collaborative filtering (CF) suffers from two key challenges, namely, the normal assumption that it is not robust, and it is difficult to set in advance the penalty terms of the latent features. We therefore propose a hierarchical Bayesian model-based CF and the related inference algorithm. Specifically, we impose a Gaussian-Gamma prior on the ratings, and the latent features. We show the model is more robust, and the penalty terms can be adapted automatically in the inference. We use Gibbs sampler for the inference and provide a statistical explanation. We verify the performance using both synthetic and real datasets

In this paper, under the ellipsoidal channel covariance errors, we investigate robust optimization for artificial noise (AN)-aided wireless information and power transfer with secrecy outage constraint in a two-tier wireless-powered heterogeneous network (HeNet). In this considered network, one femtocell consisting of multiple femtocell users (FUs) is deployed under the coverage of one macrocell consisting of multiple macrocell users (MUs), where FUs can receive information and harvest energy from the femtocell base station (FBS) simultaneously. Meanwhile, multiple eavesdroppers attempt to wiretap the confidential messages intended for FUs. To secure the information and power transfer, AN is injected into the downlink signal of the FBS simultaneously. Moreover, while satisfying the constraints on the signal-to-interference plus-noise ratio requirement at each MU, energy harvesting and secrecy outage probability requirement at each FU the transmit beamforming and AN vectors of the macrocell base station (MBS) and FBS are jointly designed to minimize the total transmit power (MTP). Due to the secrecy outage constraint and channel state information (CSI) errors, this MTP problem is non-convex and contains infinite constraints, which cannot be solved directly. Resorting to the first-order Taylor expansion, Lagrange dual theory and the semi-definite relaxation (SDR) technique, we reformulate the original problem into a solvable form. Then, based on the convex reformulations, we further propose a successive convex approximation (SCA) algorithm for iteratively obtaining the optimal design. Simulation results are finally presented and demonstrate the effectiveness and security of our proposed robust design.

Here, the authors investigate robust secure transmission for wireless information and power transfer in heterogeneous networks (HeNets) under the deterministic and stochastic CSI errors, respectively, where one macrocell base station (MBS) and one femtocell base station (FBS) respectively serve to macrocell users (MUs) and femtocell users (FUs). Meanwhile, FUs acquire energy from the FBS and multiple eavesdroppers attempt to wiretap its downlink information. To secure the wireless information and power transfer, the authors propose an artificial noise (AN)-aided transmission design. Under the deterministic CSI error, the authors jointly design the beamforming and AN vectors of the MBS and FBS to minimise the total transmit power, while satisfying the constraints on the signal-to-interference plus noise (SINR) requirement at each MU, energy harvesting (EH) and secrecy outage probability requirement at each FU. This formulated problem is non-convex. Utilising S-procedure and successive convex approximation technique, the authors successfully transfer it into a series of convex problems. Under the stochastic CSI error, the joint design problem is formulated with a range of outage probabilistic constraints. Further, the authors, respectively, resort to large deviation and Bernstein-type inequalities, S-procedure to convert it into solvable forms. Simulation results demonstrate the effectiveness of the authors' proposed design.

The network security of power monitoring systems is related to national security, and illegal cross-area linkages are one of its key risks. Therefore, how to effectively and accurately locate illegal cross-area linkages in power monitoring systems is studied deeply in this paper. In this paper, through research and extensive data collection, the types of computer services commonly used in power monitoring systems and their characteristics acquisition methods are determined. Then the uniqueness of service characteristics among different devices is given. On this basis, the weighted Hamming distance is used to calculate the similarity between the service feature sets of devices on isolated network areas, where the weight is directly proportional to the uniqueness of the service feature. Through testing in the real environment of thermal power plants, wind power plants, hydropower plants, substations, and dispatching master stations with 40 to 260 devices, under the premise of reasonable setting of service feature weights, the scope of suspicious cross-area connected devices can be reduced to less than 3. The method proposed in this paper is suitable for on-site inspection of illegal cross-area linkages in power monitoring systems, which improves the accuracy and efficiency of risk investigation.

With the development of the city, traffic congestion has become one of the important factors that affect the construction of smart city. The development of BSP (bus signal priority) system can effectively alleviate the traffic congestion problem and accelerate the construction of new smart city. The BSP system based on GNSS (Global Navigation Satellite System) electronic fence can effectively realize the function of bus signal priority application. It sets the virtual trigger area of traffic signal at traffic junction by using GNSS electronic fence technology, innovatively proposes a trigger judgment algorithm improved by the ray method for the bus entering the intersection electronic fence, and aiming at the problem of packet loss and floating point of bus real – time information, the method of judging the correctness of the electronic fence entry strategy is proposed, which ensures the high accuracy rate of the bus priority system. The system has been formally applied in Nanjing, China. Through the statistical analysis of data from the feedback operating data, bus priority effect is significant. Based on the electronic fence technology, the BSP system can be used in the original urban traffic system software upgrades. It is a more economical, convenient and efficient bus priority solution.

Modern science and technology is a weapon against natural disasters. Shenzhen is a city of technology and innovation and we should endeavor to address flood control, drought relief and wind mitigation (named “three-prevention” in this context). In particular, we should vigorously promote the improvement of technology, and make use of modern information technologies such as big data, internet of things, cloud computing and artificial intelligence to create intelligent three-prevention system. This paper described in detail the characteristics and the construction status of the intelligent three-prevention system as well as the future development direction, in order to achieve the construction goals of the overall situation probing of the three-prevention, real-time decision-making assistance, flat command and control and urban resilience development. Since the intelligent three-prevention system’s launch in April 2020, it has started more than 20 times of emergency response against typhoon and flood. The whole process functions of front-end intelligent perception, fine dynamic simulation, real-time forecast and early warning, emergency command and post-disaster assessment have been preliminarily realized.

As a efficient random access scheme, contention resolution diversity slotted ALOHA (CRDSA) improves the system throughput greatly, although deadlock problem limits the maximum throughput of system. To address this issue, in this paper, a non-orthogonal multiple access (NOMA) technique is proposed with random access (RA) scheme for supporting multiple terminals access in satellite IoT-oriented networks, by utilizing a code domain NOMA called Sparse Code Multiple Access (SCMA). In the proposed scheme, the throughput is mainly improved by physical (PHY) decoding and iterative decoding. Through mathematical analysis, the throughput lower bound of our scheme is shown and the simulated result proves the proposed scheme is efficient in enhancing RA throughput.

Nowadays, information resources have become the most important part of resources in the Internet era, and the problems of information security have become the research hotspot. Channel coding information hiding technology is a new type of information hiding technology, which has been widely studied in the field of communication, but there is little research on the application of spacecraft. Based on this, this paper studies a channel coding information hiding technology based on spacecraft application. In the research, this paper mainly uses the good randomness and security performance of chaotic sequences, applies them to the random coding of encrypted information, and realizes the hiding of encrypted information. In the simulation analysis, it can be found that the channel coding information hiding technology proposed in this paper has good encryption performance, and when the channel bit error rate is set to 5 × 10−3, the recovered information error rate is 0, and the encrypted information can be recovered without error.

The power supply subsystem of satellite is the core constant, throughout the whole flight stage. The solar array-battery power system is the main selection for the regulated bus power supply system architecture. The power conditioning unit is responsible for bus voltage regulating and energy management, thus supplying the high-performance steady bus to spacecraft payload. The architecture of the power subsystem must consider system overall efficiency, bus static output characteristics, reliability, modularity, power density, even control methods and other technical specifications. In this paper, the working principle and performance of different power supply system architectures are analyzed, concerning the solar array power regulation, the charging and discharging regulation of battery, the battery charging control strategy, and the bus voltage stabilization. Therefore, the power system designers can make choice of a power system architecture according to the different tracks, load characteristics, solar array configuration, battery configuration, satellite’s weight, and thermal control system.

In order to reduce the pilots overhead and improve the estimation accuracy of doubly selective sky wave channel estimation based on compressed sensing (CS), a new compressed sensing estimation scheme is proposed by exploiting the temporally correlated sparsity of channel coefficient in this paper. Combining with the transmission model of High-frequency (HF) orthogonal frequency division multiplexing (OFDM) system and the sparsity of sky wave channel in transformed domain, the channel estimation problem is formulated under the framework of compressed sensing. On this basis, the temporal correlation of channel coefficient sparsity is theoretically verified by proving that the channel coefficients in several successive OFDM symbols share the common support set. Since the temporal correlation enhances the sparsity of channel coefficients in time dimension, then higher reconstruction accuracy of compressed sensing and the reducuction of pilot overhead can be expected. Simulation results verify the effectiveness of the proposed scheme.

In tag-enhanced video recommendation systems, videos are attached with some tags that highlight the contents of videos from different aspects. Tag ranking in such recommendation systems provides personalized tag lists for videos from their tag candidates. A better tag ranking model could attract users to click more tags, enter their corresponding tag channels, and watch more tag-specific videos, which improves both tag click rate and video watching time. However, most conventional tag ranking models merely concentrate on tag-video relevance or tag-related behaviors, ignoring the rich information in video-related behaviors. We should consider user preferences on both tags and videos. In this paper, we propose a novel Graph neural network based tag ranking (GraphTR) framework on a huge heterogeneous network with video, tag, user and media. We design a novel graph neural network that combines multi-field transformer, GraphSAGE and neural FM layers in node aggregation. We also propose a neighbor-similarity based loss to encode various user preferences into heterogeneous node representations. In experiments, we conduct both offline and online evaluations on a real-world video recommendation system in WeChat Top Stories. The significant improvements in both video and tag related metrics confirm the effectiveness and robustness in real-world tag-enhanced video recommendation. Currently, GraphTR has been deployed on WeChat Top Stories for more than six months. The source codes are in https://github.com/lqfarmer/GraphTR.

With the improvement of science and technology, the application of ARM is more and more extensive, Cortex-M3 is a new product launched by ARM. Cortex-M3 is a 32-bit processor core, using ARMv7-M architecture, which has high advantages in low cost and low power consumption. In this paper, the ARM Cortex-M3 and STM32 microcontrollers are studied. By selecting a semiconductor STM32 F103ZET6 development board, the basic module GPIO of STM32 is studied, and then the timer is used to generate PWM (pulse width modulation) output. PWM is an effective technique that uses the digital output of a microprocessor to get command of analog circuits, and is widely applied in many fields from measurement and communication to rate of work command and conversion. Therefore, the study of PWM technology has very important practical significance [1] [2]. The time of the timer is the period of the PWM. By calculating the square wave frequency and duty cycle, the timer and IO port are configured, and finally the PWM output is completed.

Ultrasound passive cavitation mapping (UPCM) technique can feed back the distribution information and scattering intensity of ultrasound cavitation in space, and has been proposed for the purpose of ultrasound treatment monitoring. The traditional UPCM reconstruction algorithm based on time exposure acoustics (TEA) introduces the artifacts due to large axial extension and high sidelobes when using the array transducers for diagnostic purposes as passive receivers. We recently applied dual apodization with cross-correlation (DAX) technique to enhance UPCM performance. In this study we propose an enhanced version by using multi-apodization with cross-correlation (MAX), where multiple receive apodization functions divided from DAX apodization functions are used. The obtained results on single- and multi-source simulations indicate that MAX-UPCM algorithm increases the image contrast by 20.2 dB and 23.4 dB over TEA-UPCM algorithm and by 10.1 dB and 7.7 dB over DAX-UPCM algorithm. The robust performance of the proposed UPCM algorithm in this study may enhance the detectability of ultrasound-induced cavitation and benefit the treatment monitoring.

Hemispherical-array passive acoustic mapping (PAM) has been proposed to monitor transcranial ultrasound treatments in real time by enabling three-dimensional (3D) mapping of cavitation activity in the brain. The traditional hemispherical-array PAM is implemented based on delay-and-sum processing, resulting in low resolution and high-level interference due to the low-frequency limit and the interaction between sound sources. Inspired by our previous work, in this study we apply dual apodization with cross-correlation (DAX) technique in hemispherical-array PAM to achieve better image quality with reduced interference artifacts. The proposed algorithm first designs two complementary receive apodizations, and then measures the similarity between two beamformed signals with the cross-correlation coefficient, finally weights the combined signal with the correlation coefficient. The algorithm is tested through numerical simulations, where a 30-cm and 256-element hemispherical array is used to receive the signal from each element (Gaussian-windowed sinusoidal signal at 600 kHz) in three different source configurations without and with noise. The image enhancement is assessed with signal-to-interference ratio (SIR). The obtained results on simulated data indicate that in the noise-free (noisy) case, the proposed algorithm improves the SIR by 12.0 dB (19.5 dB), 14.3 dB (20.4 dB), and 11.8 dB (18.4 dB) in single-, multi-, and random-source configurations. The proposed hemispherical-array PAM algorithm presented in this study has the potential to provide high-quality 3D cavitation image during transcranial ultrasound treatment.

Sequential recommendation and group recommendation are two important branches in the field of recommender system. While considerable efforts have been devoted to these two branches in an independent way, we combine them by proposing the novel sequential group recommendation problem which enables modeling group dynamic representations and is crucial for achieving better group recommendation performance. The major challenge of the problem is how to effectively learn dynamic group representations based on the sequential user-item interactions of group members in the past time frames. To address this, we devise a Group-aware Long- and Short-term Graph Representation Learning approach, namely GLS-GRL, for sequential group recommendation. Specifically, for a target group, we construct a group-aware long-term graph to capture user-item interactions and item-item co-occurrence in the whole history, and a group-aware short-term graph to contain the same information regarding only the current time frame. Based on the graphs, GLS-GRL performs graph representation learning to obtain long-term and short-term user representations, and further adaptively fuse them to gain integrated user representations. Finally, group representations are obtained by a constrained user-interacted attention mechanism which encodes the correlations between group members. Comprehensive experiments demonstrate that GLS-GRL achieves better performance than several strong alternatives coming from sequential recommendation and group recommendation methods, validating the effectiveness of the core components in GLS-GRL.

Real-world integrated personalized recommendation systems usually deal with millions of heterogeneous items. It is extremely challenging to conduct full corpus retrieval with complicated models due to the tremendous computation costs. Hence, most large-scale recommendation systems consist of two modules: a multi-channel matching module to efficiently retrieve a small subset of candidates, and a ranking module for precise personalized recommendation. However, multi-channel matching usually suffers from cold-start problems when adding new channels or new data sources. To solve this issue, we propose a novel Internal and contextual attention network (ICAN), which highlights channel-specific contextual information and feature field interactions between multiple channels. In experiments, we conduct both offline and online evaluations with case studies on a real-world integrated recommendation system. The significant improvements confirm the effectiveness and robustness of ICAN, especially for cold-start channels. Currently, ICAN has been deployed on WeChat Top Stories used by millions of users. The source code can be obtained from https://github.com/zhijieqiu/ICAN.

Every day, the power system receives massive I/O logs. The amount of data in these logs is so large that it takes huge computational resources to analyze. Therefore, it is necessary to reduce the size of the massive I/O logs and only analyze the key log data, thereby reducing the workload of invalid analysis. This paper takes the I/O log of the substation as the research object, and studies the dimension reduction method of the massive I/O data flow log, which reduces the computational load brought by the high-dimensional I/O data flow log data and reduces the massive I/O data flow log. This paper proposes a method of secondary dimensionality reduction. Firstly, the high dimensional I/O log data stream is classified, so that the data is transformed from high-dimensional to low-dimensional. Then, the dimension is reduced again in each category, so that the most simplified massive I/O logs are achieved. Through theoretical analysis, we can come to the conclusion that the computational time complexity of the data after dimension reduction is reduced by more than 80%.

Recently, the integration of Edge computing and Wireless power transmission has provided low delay information processing and sustained green energy in IOT system. However, the constraint of system energy consumption remains the key factor which limits performance of IOT system. In this letter, we propose an energy-efficiency based resource allocation and power control algorithm (EE-RAPC) to Wireless Power Transfer and Edge Computing system (WPT-EC-IoT) in time division duplex-orthogonal frequency division multiple access mode (TDD-OFDMA). The algorithm optimizes distribution of subchannels, assigns different power to each subchannel to obtain local optimum of system energy consumption under the condition of specific time allocation coefficient and then adjust time allocation coefficient to get the global optimum. Numerical results demonstrate that the proposed algorithm maintains good performance in energy consumption of IOT system.

Session-based target behavior prediction aims to predict the next item to be interacted with specific behavior types (e.g., clicking). Although existing methods for session-based behavior prediction leverage powerful representation learning approaches to encode items' sequential relevance in a low-dimensional space, they suffer from several limitations. Firstly, they focus on only utilizing the same type of user behavior for prediction, but ignore the potential of taking other behavior data as auxiliary information. This is particularly crucial when the target behavior is sparse but important (e.g., buying or sharing an item). Secondly, item-to-item relations are modeled separately and locally in one behavior sequence, and they lack a principled way to globally encode these relations more effectively. To overcome these limitations, we propose a novel Multi-relational Graph Neural Network model for Session-based target behavior Prediction, namely MGNN-SPred for short. Specifically, we build a Multi-Relational Item Graph (MRIG) based on all behavior sequences from all sessions, involving target and auxiliary behavior types. Based on MRIG, MGNN-SPred learns global item-to-item relations and further obtains user preferences w.r.t. current target and auxiliary behavior sequences, respectively. In the end, MGNN-SPred leverages a gating mechanism to adaptively fuse user representations for predicting next item interacted with target behavior. The extensive experiments on two real-world datasets demonstrate the superiority of MGNN-SPred by comparing with state-of-the-art session-based prediction methods, validating the benefits of leveraging auxiliary behavior and learning item-to-item relations over MRIG., Comment: Short paper with 7 pages, WWW'2020

Energy storage systems (ESSs) have recently been incorporated into low-voltage distribution networks (DNs) for promoting hosting capacity of photovoltaic (PV) systems. On the other hand, applications of information communications technologies enhance the self-regulation ability of LVDNs and are playing an important role to resolve integration problems of PV systems. Although the two aspects influence each other, they are generally configured separately in LVDNs. This study proposes a collaborative configuration scheme based on a bi-level optimisation model for distributed ESSs and cyber systems in LVDNs with high penetration of PV systems. In the upper level, the ESSs and wireless control terminals are deployed to minimise investment and operation cost with consideration of specific control of ESSs. In the lower level, communication topology is further optimised according to the upper-level results to promote control performance and balance communication topology complexity. The economy and effectiveness of the proposed configuration scheme are demonstrated by comparing with separated cyber and physical configuration schemes in case studies.

After the ZS five-terminal flexible DC transmission demonstration project in Zhejiang Province put into operation, the stability characteristics of ZS Power Grid have been changed. Based on the ZS equivalent grid with flexible DC connection for offshore wind power, the maximum output of offshore wind power considering voltage stability is analyzed in this paper. The details are as follows: Firstly, based on the conventional system equivalent method, the ZS Power Grid is equivalent to a small system with the same grid structure, thus simplifying the research object. Secondly, the branch voltage stability indexes are briefly introduced, and their adaptabilities to the equivalent grid are verified. Finally, based on the branch voltage stability index, the maximum output of the wind farm considering voltage stability is given.

In text categorization, a document is usually represented by a vector space model which can accomplish the classification task, but the model cannot deal with Chinese synonyms and polysemy phenomenon. This paper presents a novel approach which takes into account both the semantic and statistical information to improve the accuracy of text classification. The proposed approach computes semantic information based on HowNet and statistical information based on a kernel function with class-based weighting. According to our experimental results, the proposed approach could achieve state-of-the-art or competitive results as compared with traditional approaches such as the k-Nearest Neighbor (KNN), the Naive Bayes and deep learning models like convolutional networks.

Automated platooning of trucks has its beneficial effects on energy saving and traffic flow efficiency. The vehicles in a platoon, however, need to maintain an extremely short headway to achieve these goals, which will result in a heavily blocked front view for the driver in a following truck. Monitoring surrounding traffic environment and foreseeing upcoming hazardous situations becomes a difficult, yet safety-critical task. This exploratory study aims to investigate whether providing platoon drivers with additional visual information of the traffic environment can influence their monitoring pattern and increase awareness of the upcoming situation. 22 professional truck drivers participated in the driving simulator experiment, either following a see-through lead truck (i.e., with projection of forward scene attached to the rear of the lead truck), or a normal lead truck until the automation system failed unexpectedly in a critical situation. Results showed that when provided with front view projection, the participants spent 10% more time monitoring the road, and responded less severely to a critical situation, suggesting a positive effect of the “see-through” technology.

Photovoltaic (PV) power generation is affected by the availability and quality of sunlight. Therefore, PV power generation systems must be frequently started-up when the illumination intensity is not stable. To reduce the number of startup and shut-down events, to prolong the service time of PV inverters, and to reduce the associated impact on the power grid, a novel startup optimization control strategy for distributed PV power generation systems is proposed. The proposed algorithm can implement start–stop inverter control according to different PV power generation conditions without modifying the existing hardware architecture, thus minimizing the startup and shut-down events due to insufficient PV power generation. This protects the key equipment in the PV inverter and enhances the ability of the power grid to adapt to new power generation inputs. The proposed optimization control algorithm is verified experimentally and is found to provide a novel solution for the efficient and reliable control of distributed PV power generation systems. The research presented in this paper is expected to be applicable to numerous engineering projects.

The nature of free communication results in increased challenges with respect to the reliability of user interaction in online social network. It is difficult to ensure that the users in a large-scale network are familiar with one another and that they interact without risk. Users behaviors might therefore be vulnerable to attacks or frauds. Trust relationship detection scheme is a feasible solution. This study proposes an original trust link detection scheme, named Trust Traversal, which includes three aspects, i.e., trust scheme, trust link category, and traversal scheme, which are used to build up strong or weak trust links among users. Firstly, we address overview of Trust Traversal scheme and its related formal definitions. Then, trust scheme model is proposed for managing calculation methods of subjective trust, reputation, and indirect link. Subjective trust is calculated based on past interaction data, while reputation is defined as collective objective trust with three factors: effective accumulation, time attenuation, and intensity factor. Additionally, trust of indirect link in trust scheme is computed with three rules: trust transitivity rule, reputation weight rule, and level attenuation rule. To describe the relationship with both reliability and closeness, we present a trust link category, which comprise three kinds of trust links: strong trust link, weak trust link, and untrusted link, and meanwhile, gives the classifying method based on two factors: mutual trust factor and interaction active factor. Moreover, Trust Traversal scheme for both single and multiple target users is proposed to detect strong or weak trust links, which implies trustable and close relationships among users. In Trust Traversal scheme, each step is determined by trust based probability, and thus the user with the strongest trust link would have the highest probability of being the traversal user in the trust link. And also, terminal mechanism of Trust Traversal scheme is discussed. Finally, we provide an examination to determine further the performance of our proposed scheme.