Search

The virtual synchronous generator (VSG) control is regarded as an effective solution for operating converters in weak grid conditions due to its excellent grid support functions. However, successful implementation of the VSG control of an AC/DC converter relies on the existence of a steady DC voltage at the DC side, while this is not easy to achieve for the cascaded converter system of this work, i.e., the Modular Multilevel Converter-based High-Voltage Direct-Current (MMC-HVDC) transmission systems with offshore wind farm integration. To address this issue, a novel grid-forming control strategy with real-time inertia support and direct DC-link voltage regulation is proposed for the Receiving End Converter (REC) of the MMC-HVDC. The intrinsic power balancing regime of the internal energy stored in MMC's submodule (SM) capacitors is utilized for grid synchronization rather than emulating the swing equation as requested by the VSG control. For being more robust to sudden power variations, proper insertion strategies of the REC are developed to decouple the DC-link voltage of MMC-HVDC from the voltage of SM capacitors. Furthermore, in terms of the grid fault-ride-through issue of the REC, an associated FRT control strategy is proposed aiming for power angle stability and the stability of SM capacitor voltage. Finally, simulation results in PSCAD/EMTDC show that the proposed control can provide fast inertia support with satisfactory control of DC-link voltage and FRT capability, etc.

A crucial problem of neural networks is to select an architecture that strikes appropriate tradeoffs between underfitting and overfitting. This work shows that L1 regularizations for two-layer neural networks can control the generalization error and sparsify the input dimension. In particular, with an appropriate L1 regularization on the output layer, the network can produce a tight statistical risk. Moreover, an appropriate L1 regularization on the input layer leads to a risk bound that does not involve the input data dimension. The results also indicate that training a wide neural network with a suitable regularization provides an alternative bias-variance tradeoff to selecting from a candidate set of neural networks. Our analysis is based on a new integration of dimension-based and norm-based complexity analysis to bound the generalization error.

DC technology has gained considerable interest in the medium voltage applications due to the benefits over the AC counterpart. However, to utilize the full capacity of this development, selection of a suitable power electronic converter topology is a key aspect. From the pool of voltage source converters (VSCs), it is unclear which topology is suitable for multi-megawatt applications at medium voltage dc (MVdc) levels. To address this, the paper proposes a selection guideline based on reliability and optimum redundancy levels of VSCs for MVdc applications. This will be combined with other functional factors such as operational efficiency and return-on-investment. Three candidate multi-level topologies namely three-level neutral point clamped converter (3L-NPC), modular multi-level converter (MMC) and cascaded 3L-NPC (which is being used for the first MVdc link in the UK) have been evaluated over two-level-VSC from 10 kV to 50 kV. Results show that with the increase of MVdc voltage level MMC shows better performance whereas at low MVdc levels 3L-NPC is the prominent topology.

Mobile manipulators are increasingly applied to improve the efficiency in industrial manufacturing. As a typical system using multi-sensor fusion technology, accurate extrinsic calibration of manipulator’s exteroceptive sensors like camera and 2D-LiDAR is essential for mobile manipulators to perform complicated task such as mobile assembly. However, most existing camera-LiDAR calibration methods require sophisticated artificial calibration targets, leading to implementation restrictions. In this paper, by reconstructing an arbitrary trihedron that existed in the general human-made environment, a novel method is presented to estimate the transformation between the manipulator camera and the 2D-LiDAR coordinate system. The proposed method is based on the use of point, line, and plane geometry constraints between the segmented 2D-LiDAR scan and the reconstructed trihedron features. Considering the metric of reconstruction, the extended hand-eye calibration framework is implemented to recover the scale factor and hand-eye parameters. Then, a new optimization model is presented to reconstruct the key features of arbitrary trihedron in a preset global coordinate system. Finally, with only one 2D-LiDAR measurement, camera-LiDAR transformation can be calculated by constructing point-to-line and line-to-plane geometric constraints. Further, the transformation between the manipulator kinematic base and 2D-LiDAR can also be calibrated. Both simulation and real-world experiments show that the proposed method can provide robust and accurate results.

A comprehensive analysis of the motivations, gap acceptance, duration, and speed adjustment of heavy vehicle lane changes (LC) is conducted in this paper. An rich data set containing 433 discretionary LC trajectories of heavy vehicles is used in this study and the data set is divided into two data sets based on the LC direction (LC to the left lane [LCLL] and LC to the right lane [LCRL]) for comparison. It is seen that LCLL and LCRL have significantly different motivations, which also results in different gap acceptance behavior. However, the LC direction does not significantly influence the LC duration. The navigation speed significantly influences the LC duration of heavy vehicles and the LC duration will decrease with the increase of speed, indicating the substantial impact of traffic conditions on LC duration. An obvious speed synchronization phenomenon is found in the process of LCLL, which is not the case in LCRL. The results of this study highlight the distinct characteristics of the LC of heavy vehicles and produce a better understanding of the lane-changing behaviors of heavy vehicles. The fitted distributions of LC duration and further investigation into gap acceptance behaviors may be used for microscopic traffic simulation and auto driving.

Abnormal climate event is that some meteorological conditions are extreme in a certain time interval. The existing methods for detecting abnormal climate events utilize supervised learning models to learn the abnormal patterns, but they cannot detect the untrained patterns. To overcome this problem, we construct a dynamic graph by discovering the correlation among the climate time series and propose a novel dynamic graph embedding model based on graph entropy called EDynGE to discriminate anomalies. The graph entropy measurement quantifies the information of the graphs and constructs the embedding space. We conducted experiments on synthetic datasets and real-world meteorological datasets. The results showed that EdynGE model achieved a better F1-score than the baselines by 43.2%, and the number of days of abnormal climate events has increased by 304.5 days in the past 30 years.

Face super-resolution (SR) has become an indispensable function in security solutions such as video surveillance and identification system, but the distortion in facial components is a great challenge in it. Most state-of-the-art methods have utilized facial priors with deep neural networks. These methods require extra labels, longer training time, and larger computation memory. In this paper, we propose a novel Edge and Identity Preserving Network for Face SR Network, named as EIPNet, to minimize the distortion by utilizing a lightweight edge block and identity information. We present an edge block to extract perceptual edge information, and concatenate it to the original feature maps in multiple scales. This structure progressively provides edge information in reconstruction to aggregate local and global structural information. Moreover, we define an identity loss function to preserve identification of SR images. The identity loss function compares feature distributions between SR images and their ground truth to recover identities in SR images. In addition, we provide a luminance-chrominance error (LCE) to separately infer brightness and color information in SR images. The LCE method not only reduces the dependency of color information by dividing brightness and color components but also enables our network to reflect differences between SR images and their ground truth in two color spaces of RGB and YUV. The proposed method facilitates the proposed SR network to elaborately restore facial components and generate high quality 8x scaled SR images with a lightweight network structure. Furthermore, our network is able to reconstruct an 128x128 SR image with 215 fps on a GTX 1080Ti GPU. Extensive experiments demonstrate that our network qualitatively and quantitatively outperforms state-of-the-art methods on two challenging datasets: CelebA and VGGFace2., Neurocomputing'2021

The single-phase-to-ground (SPG) fault at the valve side of the converter transformer is one of the most critical faults threatening the operation of bipolar half-bridge modular-multilevel-converter-based multi-terminal direct-current (HB-MMC MTDC) systems. Following a valve-side SPG fault, the upper-arm submodule (SM) capacitors in the faulty converter may suffer a serious overvoltage, and the grid-side fault currents may experience no zero-crossings. The existing protection methods generally rely on tripping the grid-side alternate-current circuit-breaker (ACCB), which cannot fully isolate the fault. The healthy poles and stations may encounter severe disturbance and be all blocked, which will lead to a power outage of the entire MTDC system. This paper proposes an ACCB-independent scheme based on thyristor-pairs and damping SMs to address these problems. Antiparallel thyristor-pairs are connected in series within the upper and lower arms in each phase. Serial damping SMs are connected within the lower arms to help the lower-arm thyristor pairs cut off the fault currents. Although the proposed additional devices inevitably cause conduction losses, they can effectively protect the faulty converter and healthy circuit in the MTDC system. Simulations in PSCAD/EMTDC have been conducted to verify the effectiveness of the proposed strategy.

Recently the hybrid multi-port DC circuit breaker (MP-DCCB) is becoming popular in protecting HVDC grids, thanks to their re-duction of power electronics devices. In this paper, an enhanced multi-port current-limiting DCCB (MP-CLCB) for multiple line protection is proposed. The integrated fault current limiter (FCL) inside the MP-CLCB can clear the fault faster with slightly in-creased costs. To reduce the energy dissipation requirement for the surge arresters caused by the newly added current-limiting path, an energy transfer path which provides a loop with the in-ductors during the current decay stage is designed. The theoreti-cal analysis of the pre-charging, current-limiting, fault interrup-tion and energy dissipation of the MP-CLCB is carried out. Moreover, the design principles of the energy dissipation and the key parameters of the MP-CLCB are provided. The proposed approaches are verified through simulations in PSCAD/EMTDC. The results show that the MP-CLCB can replace multiple DCCBs, accelerate the fault current interruption and reduce the energy dissipation requirement for the surge arresters.

In this paper, we propose a novel strategy that identifies the dynamic relationship pattern for abnormality detection on financial time series. In particular, we select the basis indices that affect financial time series to discover the spurious relationships and construct a dynamic relationship matrix to model these. Then, we propose a graph embedding model by modifying the structural deep network embedding model to map these relationships into an embedding space. The abnormality is detected by using the outlier detection methods. To evaluate the proposed model, we have conducted the experiments by using the real financial time series (e.g., Apple, Amazon, Coke, Starbucks, and McDonald’s). The results showed that the proposed model achieved higher accuracy than the baselines by 4%.

Safe and effective autonomous navigation in dynamic environments is challenging for four-wheel independently driven steered mobile robots (FWIDSMRs) due to the flexible allocation of multiple maneuver modes. To address this problem, this study proposes a novel multiple mode-based navigation system, which can achieve efficient motion planning and accurate tracking control. To reduce the calculation burden and obtain a comprehensive optimized global path, a kinodynamic interior-exterior cell exploration planning method, which leverages the hybrid space of available modes with an incorporated exploration guiding algorithm, is designed. By utilizing the sampled subgoals and the constructed global path, local planning is then performed to avoid unexpected obstacles and potential collisions. With the desired profile curvature and preselected mode, a fuzzy adaptive receding horizon control is proposed such that the online updating of the predictive horizon is realized to enhance the trajectory-following precision. The tracking controller design is achieved using the quadratic programming (QP) technique, and the primal-dual neural network optimization technique is used to solve the QP problem. Experimental results on a real-time FWIDSMR validate that the proposed method shows superior features over some existing methods in terms of efficiency and accuracy.

In large-scale structured scenes, efficient and reliable global localization is hard to achieve for mobile robots due to the substantial increase in the search space. This article addresses this issue by proposing an optimized branch and bound (BnB) based global localization using multiscale/resolution maps, which can efficiently find the global optimal initial pose with a high success rate. Multiscale feature maps (MSFMs) are first constructed to reflect the explicit point/line information of the structured scene. Then, to narrow the position space, robust features are extracted from LiDAR data at different scales and matched with MSFMs instantaneously using the hash function. Utilizing the attitude and heading reference system, a yaw angle estimation considering environmental interference is achieved to reduce the orientation space and improve the localization reliability in ambiguous environments. Besides, with the optimal resolution level and sparse point cloud sampled from interval filtering, an optimized BnB-based search is presented by using multiresolution maps. Finally, the effectiveness of the proposed method is verified by mobile robot real-world and simulated experiments in large-scale structured environments.

Semantic information from images can be used to improve the performance of deep learning methods in recognizing human emotions. In this paper, we propose a novel framework based on the graph convolutional network for emotion recognition by utilizing the semantic relationships of different regions. First, we extract the salient image regions within video frame clips by using the bottom-up attention module to construct the node features of a graph. Then, we build the graphs containing the node features and the semantic correlations of nodes by using the graph convolutional network. For refinement, each node feature of graph vectors is enhanced via a gated recurrent unit consisting of gate and memory units to remove redundant feature information. Experimental results show that our proposed method achieves superior performance over state-of-the-art approaches for the emotion recognition on the CEAR and AFEW datasets.

Recently, reconciling social networks, identifying the accounts belonging to the same individual across social networks, receives significant attention from both academic and industry. Most of the existing studies have limitations in the following three aspects: multiplicity, comprehensiveness and robustness. To address these limitations, we rethink this problem and, for the first time, robustly and comprehensively reconcile multiple social networks. In this paper, we propose two frameworks, MASTER and MASTER+, i.e., across Multiple social networks, integrate Attribute and STructure Embedding for Reconciliation. In MASTER, we first design a novel Constrained Dual Embedding model, simultaneously embedding and reconciling multiple social networks, to formulate this problem into a unified optimization. To address this optimization, we then design an effective NS-Alternating algorithm and prove it converges to KKT points. To further speed up MASTER, we propose a scalable framework, namely MASTER+. The core idea is to group accounts into clusters and then perform MASTER in each cluster in parallel. Specifically, we design an efficient Augmented Pre-Embedding model and Balance-aware Fuzzy Clustering algorithm for the high efficiency and the high accuracy. Extensive experiments demonstrate that both MASTER and MASTER+ outperform the state-of-the-art approaches. Moreover, MASTER+ inherits the effectiveness of MASTER and enjoys higher efficiency.

With the rapid development of voltage source converter-based high-voltage direct current (HVdc) transmission, it is an irresistible trend that HVdc grid will come into being. High-voltage and high-power dc/dc converters will serve as dc transformers in HVdc grid to interconnect dc lines with different voltage ratings. This article proposes a monopolar symmetrical dc/dc converter which is composed of cascaded half-bridge submodules and series-connected insulated-gate bipolar transistors. This hybrid topology features low capital costs, high efficiency, small footprint, and bidirectional power transfer capability. Operation principle, parameter design, and the control strategies of this topology are introduced. A 480-MW, ±500 kV/±160 kV monopolar symmetrical dc/dc converter is simulated to verify its performance and evaluate the efficiency. In addition, a downscaled prototype rated at 2.4 kW, ±300 V/±100 V has been built and tested. Experimental results further validate the effectiveness of the proposed dc/dc converter.

Electrical load forecasting is a vital process for balancing the electricity supply and demand sections. In order to make a precise prediction, many intelligent methods aiming at forecasting future load of different horizons have been proposed in recent years. Regardless of the great progress in this field, yet two major problems widely exist in the field of midterm load forecasting, which are pseudo midterm forecast and divergence of the forecasting error. This article proposes a multistep forecasting approach, mainly based on phase space reconstruction and support vector machine (SVM) methods, to solve these two problems. The relations between the forecasting step in the future and the historical data points that need to be fed into the SVM model are derived, which endue the model with the real ability to forecast adjustable steps of future load without the divergence of error and, therefore, have a notable engineering application significance. The proposed methodology is implemented on the European Network on Intelligent Technologies dataset. Compared with previous methods, the results show that the multistep implemented model is of more accurate prediction and stronger robustness. Moreover, the method is easy to implement and able to be combined with other intelligent methods to get better performance.

In this paper, a new generalized step-up multilevel DC-AC converter is proposed, which is suitable for applications with low-voltage input sources such as photovoltaic power generation and electric vehicles. This inverter can achieve a high voltage gain by controlling the series-parallel conversion of DC power supply and capacitors. Only one DC voltage source and a few power devices are employed. The maximum output voltage and the number of output levels can be further increased through the switched-capacitor unit's extension and the submodule cascaded extension. Moreover, the capacitor voltages are self-balanced without complicated voltage control circuits. The complementary operating mechanism between each pair of switches simplifies the modulation algorithm. The inductive-load ability is fully taken into account in the proposed inverter. Additionally, a remarkable characteristic of the inverter is that the charging and discharging states among different capacitors are synchronous, which reduces the voltage ripple of the frontend capacitors. The circuit structure, the working principle, the modulation strategy, the capacitors and losses analysis are presented in detail. Afterwards, the advantages of the proposed inverter are analyzed by comparing with other recently proposed inverters. Finally, the steady-state and dynamic performance of the proposed inverter is verified and validated by simulation and experiment. 1

Multiple-input multiple-output (MIMO) is a favorable technique that can improve system capacity and performance through spatial multiplexing. However, the performance gets degraded over the correlated wireless channels. In this letter, we consider a point-to-point MIMO system and jointly optimize both the precoding matrix and the difference between two transmit vectors to resist transmit correlation. The simulation results demonstrate that the proposed method can get average 75% gain comparing with the existing method in terms of the minimum pairwise error probability.

Multiblock data, where multiple groups of variables from different sources are observed for a common set of subjects, are routinely collected in many areas of science. Methods for joint factorization of such multiblock data are being developed to explore the potentially joint variation structure of the data. While most of the existing work focuses on delineating joint components, shared across all data blocks, from individual components, which is only relevant to a single data block, we propose to model and estimate partially joint components across some, but not all, data blocks. If covariates, with potential multiblock structures, are available, then the components are further modeled to be driven by the covariate information. To estimate such a covariate-driven, block-structured factor model, we propose an iterative algorithm based on thresholding, by transforming the problem of signal segmentation into a grouped variable selection problem. The proposed factorization provides accurate estimation of individual and (partially) joint structures in multiblock data, as confirmed by simulation studies. In the analysis of a real multiblock genomic dataset from the Cancer Genome Atlas project, we demonstrate that the estimated block structures provide straightforward interpretation and facilitate subsequent analyses.

We propose Unicoder-VL, a universal encoder that aims to learn joint representations of vision and language in a pre-training manner. Borrow ideas from cross-lingual pre-trained models, such as XLM (Lample and Conneau 2019) and Unicoder (Huang et al. 2019), both visual and linguistic contents are fed into a multi-layer Transformer (Vaswani et al. 2017) for the cross-modal pre-training, where three pre-trained tasks are employed, including Masked Language Modeling(MLM), Masked Object Classification(MOC) and Visual-linguistic Matching(VLM). The first two tasks learn context-aware representations for input tokens based on linguistic and visual contents jointly. The last task tries to predict whether an image and a text describe each other. After pretraining on large-scale image-caption pairs, we transfer Unicoder-VL to caption-based image-text retrieval and visual commonsense reasoning, with just one additional output layer. We achieve state-of-the-art or comparable results on both two tasks and show the powerful ability of the cross-modal pre-training.

This paper concentrates on the disturbances estimation and attenuation problem of a general class of the fractional-order systems. For this purpose, this paper proposes a fractional sliding-mode disturbance observer (FSMDO)-based control scheme, which is capable of mitigating the unknown disturbances asymptotically. Meanwhile, by incorporating a novel fractional sliding-mode controller into the proposed control scheme, the asymptotical convergence of trajectory tracking errors is guaranteed. The developed control scheme owns three attractive highlights: (1) it is suitable for the general fractional-order systems, including nonlinear systems and incommensurate systems; (2) the proposed FSMDO can estimate the unknown exogenous disturbances and system uncertainties timely and precisely; (3) nominal performance can be retained by compensating the observed disturbances in a feedforward manner, and thus, the robustness of the system can be strengthened. Illustrative examples are provided to show the availability and superiority of the presented FSMDO control method in terms of robust control. As compared with the conventional methods, the dynamic control performance of the closed-loop system can be improved.

Power reversal control strategies for different types of hybrid line-commutated-converter (LCC)/modular multi-level converter (MMC) based high-voltage direct-current (HVDC) systems have been proposed with the consideration of system configurations and MMC's topologies. The studies show that the full-bridge (FB) MMC gives better performance than halfbridge (HB) MMCs in terms of power reversal in hybrid LCC/MMC systems. The modulation method employed in this paper can achieve a smooth online polarity reversal for hybrid LCC/FB-MMC HVDC systems. Additional DC switches and/or discharging resistors may be needed to reverse the DC polarity of LCC/HB-MMC HVDC systems. Based on the proposed strategies, the power reversal processes of the studied systems can be accomplished within several seconds. The speed can be changed according to system operation requirements. The effectiveness of the proposed control strategies has been verified through simulations conducted in PSCAD/EMTDC.

Owing to the existence of high-order motion parameters and vertical velocity, severe range-azimuth coupling and two-dimensional spatial variability of the imaging parameters degrade the image quality of high-squint synthetic aperture radar (SAR) mounted on manoeuvring platforms. To accommodate these issues, a novel sub-aperture SAR imaging algorithm based on keystone transform (KT) and azimuth sub-region deramp is proposed. In the range dimension processing, a correction function is constructed based on the scene centre to realise non-spatial-variant range cell migration (RCM) correction and range-azimuth decoupling, and then the residual spatial-variant RCM is removed by the KT. In the azimuth dimension processing, considering the impact of the range distortion on the azimuth modulated phase, accurate analytical formula of the signal in the azimuth time domain is deduced. On this basis, a fast azimuth compression method based on the azimuth sub-region deramp is proposed, and the theoretical effective imaging area is analysed. The theoretical analysis and the simulations show that the proposed imaging algorithm has high computational efficiency and can effectively expand the imaging area of high-squint SAR mounted on manoeuvring platforms.

Transmission network reinforcements are being undertaken to meet renewable energy targets towards a low carbon\ud transition. High-voltage direct-current (HVDC) links and seriescompensated ac lines are frontrunners in these developments. Although series capacitor installations can lead to subsynchronous\ud resonance (SSR), HVDC links based on voltage source converters\ud (VSCs) can be used to effectively damp SSR upon occurrence. An\ud active damping technique to mitigate torsional interactions (TIs),\ud a form of SSR, is presented. The damping scheme considers an\ud active wideband filter to ensure positive damping in a wide range\ud of subsynchronous frequencies. A state-space representation of\ud the system has been formulated and eigenanalyses have been\ud performed to assess the impact of the HVDC link on the TIs. A\ud damping torque study for SSR screening is carried out, with\ud results complemented with time-domain simulations to assess\ud the accuracy of the small-signal models. The test system is\ud implemented in a real-time digital simulator and connected\ud to a VSC-HVDC scaled-down test-rig to validate the damping\ud scheme through hardware-in-the-loop experiments. The presented\ud damping method exhibits a satisfactory performance, with timedomain simulations and laboratory experiments showing a good\ud correlation.

Dimensionality reduction is a popular approach to tackle high-dimensional data with low-dimensional nature. Subspace Restricted Isometry Property, a newly-proposed concept, has proved to be a useful tool in analyzing the effect of dimensionality reduction algorithms on subspaces. In this paper, we provide a characterization of subspace Restricted Isometry Property, asserting that matrices which act as a near-isometry on low-dimensional subspaces possess subspace Restricted Isometry Property. This points out a unified approach to discuss subspace Restricted Isometry Property. Its power is further demonstrated by the possibility to prove with this result the subspace RIP for a large variety of random matrices encountered in theory and practice, including subgaussian matrices, partial Fourier matrices, partial Hadamard matrices, partial circulant/Toeplitz matrices, matrices with independent strongly regular rows (for instance, matrices with independent entries having uniformly bounded $4+\epsilon$ moments), and log-concave ensembles. Thus our result could extend the applicability of random projections in subspace-based machine learning algorithms including subspace clustering and allow for the application of some useful random matrices which are easier to implement on hardware or are more efficient to compute., Comment: 40 pages, 2 figures

Semantic segmentation is a process of linking each pixel in an image to a class label, and is widely used in the field of autonomous vehicles and robotics. Although deep learning methods have already made great progress for semantic segmentation, they either achieve great results with numerous parameters or design lightweight models but heavily sacrifice the segmentation accuracy. Because of the strict requirements of real-world applications, it is critical to design an effective real-time model with both competitive segmentation accuracy and small model capacity. In this paper, we propose a lightweight network named DABNet, which employs Depth-wise Asymmetric Bottleneck (DAB) and Point-wise Aggregation Decoder (PAD) module to tackle the challenging real-time semantic segmentation in urban scenes. Specifically, the DAB module creates a sufficient receptive field and densely utilizes the contextual information, and the PAD module aggregates the feature maps of different scales to optimize performance through the attention mechanism. Compared with existing methods, our network substantially reduces the number of parameters but still achieves high accuracy with real-time inference ability. Extensive ablation experiments on two challenging urban scene datasets (Cityscapes and CamVid) have proved the effectiveness of the proposed approach in real-time semantic segmentation.

With excellent mobility and flexibility, mobile manipulators have great potential for loading and unloading tasks of numerical control machine tools (CNC) in manufacturing workshops. However, because of the rough and oily ground, dynamic obstacles and the convex plate of a CNC, harsh manufacturing workshop poses a huge challenge to the localization system of an autonomous mobile manipulator. To address the above problem, this paper presents a hybrid maps enhanced localization system which mainly consists of a global localization method and a pose tracking method. Hybrid maps including hybrid grid map, multi-resolution likelihood fields (MLFs) and hybrid point map are constructed to efficaciously model the harsh environment and to improve localization performance. Our global localization method employs the convex hull sampling to spares dense Lidar data and the MLFs based branch and bound (BnB) search to speed up global search. To achieve real-time localization reliably and accurately, our pose tracking method seamlessly combines the BnB search and the adaptive Monte Carlo localization, and the Iterative Closest Point (ICP) based scan matching using the hybrid point map is adopted for higher accuracy. In addition, a distance filter improved by unscented transform is integrated into the pose tracking process to mitigate the influence of dynamic obstacles. The developed localization system is evaluated through different experiments including two weeks of loading and unloading tasks in a real manufacturing scenario, resulting in superior localization performance.

Tunnel section is the throat of transportation and attracts lots of attentions. This paper proposed a method to evaluate the driver's visual search stability based on the Markov Chain properties of eye movements. Firstly, visual and physiological data about 16 participants driving through 13 urban tunnels were collected. Then, the view area was divided into six AOIs (Area of Interest) by fast clustering of the drivers' fixation points. The one-step fixation transition probability and the stable distribution of different lane changing behavior were obtained based on the division of the view area. The probability of transition from the forward windscreen to the left rearview mirror and other 6 visual parameters were selected as indexes by correlation tests. And the first four principal components which covered 96.1% of all information were extracted. Then an evaluation method for visual search stability was implemented by principal component analysis. In order to validate the method, average lane change times, average speed and SDNN (Standard Deviation of NN Intervals) of the drivers' heart rate were clustered into two categories. According to the consistency between the evaluation results and the clustering results, the evaluation method proposed in this paper has been proven to be reliable. Finally, the score threshold for judging the driver's stability was obtained as $E=0.313$ . The method could be applied to adjustment of tunnel facilities, assistance in driving training and development of auto driving system by assessing whether a driver can take over the control of the vehicle or not.

The rapid response characteristics and high-speed growth of electric vehicles (EVs) demonstrate its potential to provide auxiliary frequency regulation services for independent system operators through vehicle-to-grid (V2G). However, due to the spatiotemporal random dynamics of travel behavior, it is challenging to evaluate the ability of EV cluster to provide ancillary services under the premise of reaching the expected state of charge (SOC) level. To address this issue, a novel calculation model of charge and discharge capacity of EV cluster based on trip chain with excellent parallel computing performance is presented in this work. Following the introduction of the characteristic variables of the proposed trip chain model, the user’s continuous travel behavior in a time scale of several weeks is simulated. In particular, a bidirectional V2G scheduling strategy based on the five-zone map is designed to guide the charging and discharging behavior of EVs, where the expected SOC levels are guaranteed. The results of a 3-week travel simulation verify the effectiveness of the presented model in coordinating the V2G scheme and calculating the charge and discharge capacity of the EV cluster.

In the recent period, significant progress has been achieved towards reconstructing the 3D face model from face image. With the support of the render engines and sufficient data, the reconstruction results are fine in detail. Nevertheless, the research on the 3D face reconstruction with texture from a single unrestricted face image is imperfect. The rebuild process lacks essential structure and texture information in the profile and the craniofacial region. To address this problem, we present a method of creating a 3D full face-and-head model with photorealistic texture from a single “in-the-wild” face image in this paper. To this end, we introduce a pipeline to integrate the highly-detailed face model into the basic model. Specifically, the basic model was built by multilinear optimization, and the highly-detailed face model which represents the facial features generated by constrained illumination distribution. Additionally, to infer the invisible region texture information corresponding to the input face image, we design an effective architecture with the generative adversarial network (GAN) for panoramic UV texture generation. The final results after UV texture mapping were visualized in the experiment, which demonstrates that the model faithfully recovers the photorealistic details in arbitrary perspective. Furthermore, compared to the state-of-the-art facial modeling techniques and existing commercial solutions, our method takes less input and performs better in surface detail.

Conventionally, the current flow control and DC fault protection issues of HVDC grids are supposed to be solved by the DC current flow controller (CFC) and DC circuit breaker (DCCB) separately, which may result in a high capital cost. This paper proposes a CFC topology with DC fault current limiting and interrupting capabilities. The topology and operating principle of the CFC are presented with theoretical analysis. The control strategies under normal and fault conditions are described. In order to reduce the use of IGBTs, an H-bridge inter-line CFC with fault current limiting capability is further proposed based on the first proposed CFC. The proposed CFCs are tested in PSCAD/EMTDC. Simulation results show that the proposed two CFCs can effectively control the current flow of two lines during normal operation and limit and interrupt DC fault currents.

In recent years, biologically active natural products have gradually become important agents in the field of drug research and development because of their wide availability and variety. However, the target sites of many natural products are yet to be identified, which is a setback in the pharmaceutical industry and has seriously hindered the translation of research findings of these natural products as viable candidates for new drug exploitation. This review systematically describes the commonly used strategies for target identification via the application of probe and non-probe approaches. The merits and demerits of each method were summarized using recent examples, with the goal of comparing currently available methods and selecting the optimum techniques for identifying the targets of bioactive natural products.

Deployment of dc circuit breakers (DCCBs) will help to isolate dc faults in dc systems. Conversely, current flow controllers (CFCs) will be employed in dc grids to balance currents among transmission lines. However, the inclusion of these devices may incur significant capital investment. A way to reduce costs is by integrating current control capabilities into DCCBs. This article presents a new device, the CB/CFC, which combines a multiline DCCB with a half-bridge based CFC. The operating principles of the device are analyzed and its operating modes are classified. A level-shift modulation method ensuring that a single bridge of the CB/CFC is modulated for each operating mode is considered. This simplifies the control scheme for CFC operation. For completeness, the CB/CFC is compared with other alternatives available in the literature. It is shown that the presented device reduces the number of semiconductor components compared to other solutions. DC fault isolation and current flow control are verified through simulations conducted in Power Systems Computer Aided Design (PSCAD).

2D-Lidars have been widely used in mobile robotics and autonomous vehicles for their precise measurement of distance. To effectively utilize the measurement of the sensor and perform data fusion, the extrinsic calibration between the equipped sensors is imperatively required. This paper presents a new method to obtain the relative pose of multiple 2D-Lidars automatically. Compared with existed researches which need human intervention by using artificial calibration target or synchronized sensors, our method is designed for the fully automatic calibration. Moreover, the method can also be further promoted to apply to other sensor systems. When the mobile robot is placed in the environment and automatically traverses the environment with arbitrary trajectory, the measurement data collected by 2D-Lidars at different moments and the motion estimation information are incorporated to construct the factor graph. Then a robust optimization is achieved to acquire the optimum solution. The performance of our method is validated by simulation and real-world experiment.

Electronic supply chain vulnerability has been threatened by counterfeits for decades and as a result, the authentication and quality of Electronic Components (ECs) are highly valued by stakeholders for safety-critical industrial systems. Thermography Nondestructive Testing (NDT), offering a rapid inspection while covering a large area within a short time frame, is a promising technique to inspect the integrity of ECs. However, recognising the internal structural layout directly through a packaged EC from thermography images is still a challenge due to its sealed and cramped structure. In this research, we propose a Pulsed Thermography (PT) based data analysis method for enhancement and identification of the layout of die and lead frame in ECs. The proposed solution combines the Sliding Window (SW) concept, Thermographic Signal Reconstruction (TSR) and Principal Component Thermographic (PCT) technique, named SWPCT, to enhance the contrast of die, die substrate and lead frame. Compared to the conventional full-scale PT image analysis, the proposed method presents its superior sensitivity in small thermal features by reducing the spatial thermal information redundancy benefiting from the sliding window PCT analysis. It overcomes the structural thermograms obstruction from the thick packaging, reveals the hidden die and lead frame layout and suppresses the noise simultaneously. By selecting a proper window size, this paper reveals that SWPCT can reconstruct the position and dimension of die and die substrate in ECs, and spatially distinguish the lead frame layout. Two groups of typical OpAmps chip samples with internal structure diversity (X-ray observed) were tested using PT to demonstrate the effectiveness of the proposed method. Simulation data were also used to verify this approach. This research could unlock the challenge of thermography-based inspection for key internal structure in the encapsulated integrated components.

Accurate and efficient electromagnetic transient (EMT) simulation of various types of solid-state transformers (SST) is extremely time-consuming due to the complex module structure, flexible topology connections, large number of electrical nodes and simulation time-steps limited in the range of micro-seconds. Therefore, it is urgent to develop the EMT equivalent modelling and fast simulation of SSTs for system level studies. Taking the modular multilevel converter (MMC) based SST as an example, this paper proposes an accelerated EMT model which focuses on the equivalence of the dual active bridge (DAB) based high-frequency link (HFL) in the SST. Compared with the existing algorithms, two critical factors of the proposed method that contribute the most to the efficiency improvement are the preprocessing of the nodal admittance equation and the conversion of the short-circuit admittance parameters. The proposed model is verified in PSCAD/EMTDC by comparing it with the detailed EMT model. The results show that the accelerated model is one to two orders of magnitude faster than the detailed model without sacrificing the accuracy. The experiment validation also confirms the validity of the proposed model.

With a large of time series dataset from the Internet of Things in Ambient Intelligence-enabled smart environments, many supervised learning-based anomaly detection methods have been investigated but ignored the correlation among the time series. To address this issue, we present a new idea for anomaly detection based on dynamic graph embedding, in which the dynamic graph comprises the multiple time series and their correlation in each time interval. We propose an entropy for measuring a graph’s information injunction with a correlation matrix to define similarity between graphs. A dynamic graph embedding model based on the graph similarity is proposed to cluster the graphs for anomaly detection. We implement the proposed model in vehicular edge computing for traffic incident detection. The experiments are carried out using traffic data produced by the Simulation of Urban Mobility framework. The experimental findings reveal that the proposed method achieves better results than the baselines by 14.5% and 18.1% on average with respect to F1-score and accuracy, respectively.

This study aims to build an accurate merging position model by incorporating heterogeneity into the accelerated hazard model based on random parameter and latent class models. Three kinds o...

Intelligent reflective surface (IRS) is envisioned as a promising transmission technique for next-generation communication systems owing to its ability to reconfigure the wireless propagation environments. In this paper, we aim to maximize the energy-efficiency (EE) for an IRS-assisted downlink multiple-input single-output system. To solve the multi-variable non-convex optimization problem, a joint optimization of the transmit beamforming vector, power, and phase shift matrix at the IRS is performed with the help of the block coordinate descent (BCD) method. Presented simulation results demonstrate that the convergence of the BCD method is remarkably fast and the single iteration based solution has almost the same EE performance (less than 2.5%) as the iterative one. Meanwhile, the proposed scheme can achieve a 70.7% gain in EE when compared with the IRS-free counterpart.

Cloud servers usually face low utilization. One method to mitigate such problem is to collocate multiple applications onto a single server to improve the resources exploitation. However, due to the workload variations and resources constrains, collocating applications interfere with each other, and this incurs performance degradation. The lack of effective resources managing mechanism further drops the utilization as the number of latency-sensitive user-facing applications, such as search engines and online shopping, increases. In this paper, we assume the strategy is to collocate one latency-critical task with one or more batch tasks. We design a reinforcement learning based controller that can simultaneously meet the performance target of the latency-critical task and improve the throughput of batch tasks. We combine the concept of workload characterization to make our design non-intrusive and widely applicable to any applications. The online reinforcement learning helps with capturing workload variations. We implement a prototype of our design and evaluate on real-world applications. The result indicates its effectiveness compared with uncontrolled collocation and the static feedback-based controller.

During emergency situations in buildings (e.g., fire at Grenfell Tower), rapid and safe evacuation is crucial for saving lives of people. Computer models allow determining evacuation times for various emergency conditions. These models offer insights into potential difficulties and offer possible solutions to evacuation challenges (e.g., low visibility, high anxiety, low environmental familiarity) in a short time and at low cost. This project focuses on defining the optimal location of guidance signs on evacuation route networks in building using optimisation and simulation with the aim of avoiding congestion. A bio-inspired optimization is introduced into the Social Force Model in order to avoid high density areas while following the fastest route.

An epileptic seizure is a chronic disease with sudden abnormal discharge of brain neurons, which leads to transient brain dysfunction. To detect epileptic seizures, we propose a novel idea based on a dynamic graph embedding model. The dynamic graph is built by identifying the correlation among the multi-channel EEG signals. Graph entropy measurement is exploited to calculate the similarity among the graph at each time interval and construct the graph embedding space. Since the abnormal electrical brain activity causes the epileptic seizure, the graph entropy during the seizure time interval is different from other time intervals. Therefore, we propose an entropy-based dynamic graph embedding model to cluster the graphs, and the graphs with epileptic seizures are discriminated. We applied the proposed approach to the Children Hospital Boston-Massachusetts Institute of Technology Scalp EEG database. The results have shown that the proposed approach outperformed the baselines by 1.4% with respect to accuracy.

Existing ship inverse synthetic aperture radar (ISAR) imaging algorithms are not applicable, when the signal-to-noise ratio (SNR) is low, for the translational motion that cannot be well compensated by existing algorithms. To achieve ship ISAR imaging with low SNR, a ship ISAR imaging algorithm based on the generalized radon-Fourier transform (GRFT) is proposed in this paper. Considering not only the rotational motion but also the translational motion between the radar and the ship, the proposed algorithm uses the GRFT to simultaneously compensate the time-variant range envelopes and the Doppler phase. Thus, the signal coherence is fully utilized, and the coherent integration of the ship’s multicomponent echo signal is realized. Subsequently, to overcome the problem of the heavy computational load and improve the efficiency of the proposed algorithm, the scheme of cascaded GRFTs that consists of the coarse GRFT and the subsequent fine GRFT is adopted. The coarse GRFT with large search ranges and intervals is aimed at obtaining the real ranges of ship scatter points’ motion parameters. Based on the coarse GRFT result, the fine GRFT with small search ranges and intervals is performed to efficiently obtain the coherent integration result. Then, based on the coherent integration result, the constant false alarm rate (CFAR) detection is performed to obtain the desired scatter points and their amplitudes and motion parameters, and the multicomponent signal is reconstructed. Finally, based on the reconstructed multicomponent signal, the high-quality instantaneous ship ISAR image can be obtained. Computer simulations and experiment results validate the effectiveness of the proposed algorithm.

Deceptive jamming is an effective jamming method for synthetic aperture radar (SAR). Time-delay repeater jamming, which is an important method of deceptive jamming because of its simplicity and effectiveness, can generate false targets in the SAR image and further affect the image interpretation. To detect the false target, this study first establishes the multi-angle SAR jamming signal model. Subsequently, the representations of the false targets and real targets in SAR image are analysed. Based on the aforementioned analysis, a SAR deceptive jamming target detection method based on multi-angle SAR images is proposed. Through the registration of multi-angle SAR images and the change detection, the false targets generated by the time-delay repeater jamming can be detected, and the SAR anti-jamming can be realised correspondingly. Computer simulation verifies the effectiveness of the proposed method.