Search

Aiming at the received signal strength index (RSSI) in wireless positioning system, an improved weighted KNN fingerprint positioning algorithm is proposed in this paper. The algorithm pre-processes fingerprint data in offline stage that including eliminating outliers and Kalman filtering first, in order to improve the accuracy of data acquisition. Secondly, the fingerprint data is partitioned by using RSSI to attenuate obstacles such as walls. Then, points with significant RSSI differences in each region are selected as regional feature points, and the distance between RSSI of test points and feature points in each region is calculated respectively to determine the region in which the test points are located. Geometric method is used to analyse and define the correlation degree, and KNN is re-weighted to achieve accurate positioning in the region. Finally, experiments were carried out in the indoor environment to complete the establishment of the fingerprint database. Compared with the existing NN, KNN and WKNN, the experimental analysis results show that the accumulated error and average error are better than the traditional algorithm with the increase of measurement points, which has reference value for the complex environment positioning technology. [ABSTRACT FROM AUTHOR]

Due to the presence of numerous branches in the pressurized water pipeline network (PWPN), leak-caused negative pressure wave (NPW) has non-unique propagation path, so leak detection and positioning for PWPN is highly challenging. This paper proposes a leak monitoring and positioning method for PWPN based on the time delay matching of NPW propagation. Via introducing the NPW velocity equation, the PWPN is divided into short segments with equal NPW propagation duration, and then the shortest NPW propagation path from each discrete point, namely hypothetical leak point, to each deployed pressure transmitter, is efficiently numerically searched and then used to calculate the NPW arrival time and establish a contrastive time delay library (CTDL) in advance. A set of signal processing algorithm are developed to capture the NPW arrival time and compare these measured values with the pre-established CTDL to accurately position leak point in the PWPN. The effectiveness of this method is verified via leak localization experiments on a field PWPN with an area of 12 × 12 km and 1478 branches. The proposed method can achieve leak positioning with errors less than 100 m throughout the entire pipeline network when only six pressure transmitters are effective. [ABSTRACT FROM AUTHOR]

Tensegrity structures, which consist of tension cables and compression rods, are widely used in various fields. It is particularly important to find the tensegrity with definite geometric configuration for the following research and application. This paper presents a traversal topology-finding method for tensegrity structures based on dynamic programming algorithm and ground structure method. After the designer has given the cable topology, the dynamic programming is adopted to realize the traversal selection of the rod. In order to reduce the traversal space and realize the constraint of rod length type, the rod is classified according to length. Compared with existing topology-finding methods, the algorithm can not only output all feasible topologies but also incorporate structural constraints, such as rod length type and prestressed stability, into the calculation process. Four tensegrity numerical examples illustrate the feasibility and effectiveness of the topology-finding method. [ABSTRACT FROM AUTHOR]

With the progress of remote sensing sensors, the quality of optical remote sensing image is significantly improved, and target detection on it can extract rich feature information. However, due to the characteristics of remote sensing image with various target sizes and a large proportion of the number of small targets, increasing the difficulty in target detection for it. In response to this challenge, this paper proposes an improved YOLOv8 algorithm for multi-scale target detection of optical remote sensing images. First, we propose a PSPPF module, which improves the model's ability to adapt to different data distributions; Second, DSConv is introduced into the Backbone structure of YOLOv8 to reduce the complexity of the network while maintaining the performance of model detection; Finally, we replace the original loss function CIoU with MPDIoU to improve the localization accuracy of the prediction box. Applying the improved algorithm to the public dataset NWPU VHR-10, the mAP value of the our algorithm is 95.1%, which is 3.0% higher than that of the original YOLOv8, indicating that the proposed algorithm is able to effectively detect multi-scale targets in optical remote sensing images. [ABSTRACT FROM AUTHOR]

Aiming at the characteristics of remote sensing images such as a complex background, a large number of small targets, and various target scales, this paper presents a remote sensing image target detection algorithm based on improved YOLOv8. First, in order to extract more information about small targets in images, we add an extra detection layer for small targets in the backbone network; second, we propose a C2f-E structure based on the Efficient Multi-Scale Attention Module (EMA) to enhance the network's ability to detect targets of different sizes; and lastly, Wise-IoU is used to replace the CIoU loss function in the original algorithm to improve the robustness of the model. Using our improved algorithm for the detection of multiple target categories in the DOTAv1.0 dataset, the mAP@0.5 value is 82.7%, which is 1.3% higher than that of the original YOLOv8 algorithm. It is proven that the algorithm proposed in this paper can effectively improve target detection accuracy in remote sensing images. [ABSTRACT FROM AUTHOR]

The CDT subproblem, which minimizes a quadratic function over the intersection of two ellipsoids, is a classical quadratic programming problem. In this paper, we study a method of solving CDT with a positive Lagrangian duality gap. An efficient splitting algorithm is proposed for finding the global optimal solutions. A cutting plane is firstly added to divide the feasible set of CDT into two subsets, and then two new quadratic programming problems with ellipsoidal and linear constraints are generated accordingly. Using the newly developed technique — second-order cone constraints to enhance the efficiencies of the SDP relaxation-based algorithms on the two subproblems, an optimal solution of CDT can be acquired by comparing the objective values of the two subproblems. Numerical experiments show that the new algorithm outperforms the two recent SDP relaxation-based algorithms, the two-parameter eigenvalue-based algorithm and the solver Gurobi 9.5 for certain types of CDT. [ABSTRACT FROM AUTHOR]

Summary: Graph clustering is a primitive operation for graph data mining. It plays an important role to reveal community clusters, hubs, and outliers in complex networks. There are several graph clustering algorithms have been proposed based on the well‐studied SCAN algorithm in recent years. However, SCAN and its improved sequential variants are prohibitively slow due to their iterative computations. The parallel variants are focusing on improving the efficiency of graph clustering by utilizing multi‐cores computer architectures on single computing node with complex optimization techniques. Therefore, SCAN and its variants are not suitable for processing very large graphs due to the limitations of memory size and storage volume on a single node. In this article, we proposed a distributed parallel structural clustering algorithm using MapReduce. In order to improve the efficiency further, we proposed optimization techniques including partition based clustering and simplified combination with labels to accelerate the operations. We conducted extensive experiments on real world datasets. The experimental results showed our algorithm is high efficiency and scales well under different settings. [ABSTRACT FROM AUTHOR]

When dealing with the mutual storage relationship of behavioral features in long time sequence video, the convolutional neural network is easy to miss important feature information. To solve the above problems, this paper proposes a super automatic algorithm combining nonlocal convolution and three-dimensional convolution neural network. The algorithm uses sparse sampling to segment the long time sequence video to reduce the amount of redundant information, and integrates non-local convolution into the residual neural network, thus forming a super automatic full variational - L1 algorithm. Experimental results show that the proposed method can significantly improve the efficiency of behavior recognition. [ABSTRACT FROM AUTHOR]

Mobile edge computing (MEC) driven by 5G cellular systems has recently emerged as a promising paradigm, enabling mobile devices (MDs) with limited computing resources to offload various computation-intensive tasks (such as autopilot, online game) to edge servers to enhance the data processing capabilities of MDs. However, the uncertainty of wireless channel state and data volume of offloading tasks, as well as the data security privacy of offloading tasks, bring serious challenges to computation offloading in MEC. In this article, we consider a time-varying MEC scenario and formalize the delay and energy consumption during the computation offloading process as a joint optimization problem. Then the optimization problem is decomposed into two sub-problems: intelligent task prediction and resource allocation. Different from traditional methods, we improve the federated learning (FL) algorithm and propose a thoughtful cloud-edge-client FL task prediction mechanism based on Bidirectional Long Short-Term Memory. Each participating MD trains the model locally without uploading data to the server, and periodically aggregates the model in the edge and in the cloud. The algorithm both eliminates the need to solve complex optimization problems and ensures user privacy security. Finally, experimental results show that our proposed algorithm significantly outperforms other benchmark algorithms in energy efficiency. [ABSTRACT FROM AUTHOR]

Industry and research communities have widely studied Blockchain technology, and the consortium blockchain is currently the most used category with a wide range of applications. However, issues, such as the performance of consensus mechanisms, have become essential constraints on promoting and applying the consortium blockchain. To improve the performance of the consortium blockchain consensus, we use the practical Byzantine fault tolerance (PBFT) consensus widely used in consortium blockchains to reduce the number of consensus nodes to optimize performance. Using the PBFT consensus, we screen high-performance nodes and obtain a reliable and limited number of consensus nodes. We propose a genetic algorithm-based blockchain consensus algorithm improvement scheme, design the fitness function of blockchain nodes and the genetic algorithm to iterate out consensus node groups with excellent indicators continuously, and finally iterate the nodes participating in the consensus. This algorithm can increase the speed and efficiency of the consensus, block generation, and computation. The algorithm in this article is tested on the FISCO BCOS (i.e., a consortium blockchain platform built by the FISCO open-source working group), and controlled experiments and the experimental results illustrate the safety and practicability of the method. [ABSTRACT FROM AUTHOR]

The range‐angle coupling characteristics of traditional frequency diverse array (FDA) will deteriorate target localization performance. To address this problem, several decoupling methods for FDA have been developed in the past decade, mainly including deterministic methods and global optimization methods. However, most of these methods are based on frequency offset design and do not fully utilize other design degrees of freedom, such as excitation and element spacing, which limits obtained pattern performance. In this letter, a low‐sidelobe FDA pattern synthesis method is proposed by jointly optimizing frequency offset, array spacing and excitation. The entire optimization process is treated as two sub‐optimization problems, that is, a non‐convex sub‐problem involves frequency offset and array spacing, and a convex sub‐problem involves excitation. By integrating the mayfly algorithm and convex programming techniques, this synthesis problem can be effectively solved. Compared with the current state‐of‐the‐art counterparts, the proposed method takes full advantage of the FDA's design freedom and generates a doted beampattern with lower sidelobes. [ABSTRACT FROM AUTHOR]

Aiming at the inconvenience of inspection and monitoring of coal mine pump room equipment in a narrow and complex environment, this paper proposes and designs a two-wheel self-balancing inspection robot based on laser SLAM. Using SolidWorks, the three-dimensional mechanical structure of the robot is designed, and the overall structure of the robot is analyzed by finite element statics. The kinematics model of the two-wheel self-balancing robot was established, and the multi-closed-loop PID controller was used to design the two-wheel self-balancing control algorithm of the robot. The 2D LiDAR-based Gmapping algorithm was used to locate the robot and construct the map. Through the self-balancing test and anti-jamming test, it is verified that the self-balancing algorithm designed in this paper has a certain anti-jamming ability and good robustness. By using Gazebo to build a simulation comparison experiment, it is verified that the selection of the particle number is of great significance for improving the map accuracy. The actual test results show that the constructed map has high accuracy. [ABSTRACT FROM AUTHOR]

In order to reduce the impact of disguised missing value on data analysis, this paper proposes a new algorithm - The Detection Algorithm for Disguised Missing Value Based on Filter-Kmeans. The algorithm identifies mainly disguised missing value by the clustering effect, and is applied to the data set with certain probability of data points. Firstly, the suitable data object points are selected using the silhouette coefficient method and Bernoulli's law of large numbers. And then the weighted average distance is used to control the cluster traversal. Finally, the filtering operation is performed during the process of cluster traversal. According to the experimental results, the algorithm achieves better improvement in the precision ratio, recall ratio and F1-measure on the open dataset. [ABSTRACT FROM AUTHOR]

As China's distributed energy is still in the development stage, energy transmission loss will inevitably occur in the transmission process from the source end to the load end. To reduce transmission energy loss, we should also beware of electricity theft. The principle of common electricity theft methods is analyzed to improve the accuracy of established electricity theft characteristics and electricity theft detection. The ReliefF multivariate characteristics selection algorithm optimizes the electricity theft characteristics. The back propagation (BP) neural network-based electricity theft detection model is built, and the optimized characteristics are selected as the model's input. The experiment results show that the detection model has better electricity theft identification accuracy using the optimized characteristics for electricity theft detection. [ABSTRACT FROM AUTHOR]

By controlling the null distribution in the range‐angle dimension, the frequency diverse array (FDA) can handle range‐dependent interference, which is usually difficult to solve for the traditional phased array. However, the current investigations focus more on the design of weights for null control, ignoring the unique design dimension of FDA, that is, the frequency offsets. Therefore, the array potential cannot be fully exploited to maximise the range‐dependent interference suppression capability. In this paper, a novel FDA synthesis method with large null depth is proposed. By incorporating the mayfly algorithm and convex programing, frequency offsets and transmit weights are jointly optimised to generate maximum null depths in a given region while maintaining the specified array response of the desired target. Other null control methods for FDA are listed as comparison, including weights‐only optimisation method, frequency offsets‐only optimisation and frequency offsets and weights sequentially optimisation method. Numerical examples demonstrate the superiority of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Localizing previously visited places during long‐term localization and mapping, that is, loop closure detection (LCD), is a crucial technique to correct accumulated inconsistencies. In common bag‐of‐words (BoW) model, a visual vocabulary is built to associate features for detecting loops. Currently, methods that build vocabularies off‐line determine scales of the vocabulary by trial‐and‐error, which results in unreasonable feature association. Moreover, the detection precision of the algorithm declines due to perceptual aliasing given that the BoW‐based method ignores the positions of visual features. To build the optimal vocabulary automatically and eliminate human heuristics, we propose a natural convergence criterion based on the comparison between the radii of nodes and the drifts of feature descriptors in vocabulary construction. Furthermore, a novel topological graph verification method is proposed for validating loop candidates, which can effectively distinguish visual ambiguities by involving geometrical position of features and thus improve the precision of LCD. Experiments on various public datasets verify the effectiveness of our proposed approach. [ABSTRACT FROM AUTHOR]

In this work, we investigate how the choice of the key-expansion algorithm and its interaction with the round function affect the resistance of Simon-like ciphers against rotational-XOR cryptanalysis. We observe that, among the key-expansion algorithms we consider, Simon is most resistant, while Simeck is much less so. Implications on lightweight ciphers design are discussed and open questions are proposed. [ABSTRACT FROM AUTHOR]

A high-definition map of the autonomous driving system was built with the target points of interest, which were extracted from a large amount of unordered raw point cloud data obtained by Lidar. In order to better obtain the target points of interest, this paper proposes an improved RandLa-Net algorithm incorporated with NDT registration, which can be used to automatically classify and extract large-scale raw point clouds. First, based on the NDT registration algorithm, the frame-by-frame raw point cloud data were converted into a point cloud global map; then, the RandLa-Net network combined random sampling with a local feature sampler is used to classify discrete points in the point cloud map point by point. Finally, the corresponding point cloud data were extracted for the labels of interest through numpy indexing. Experiments on public datasets senmatic3D and senmatickitti show that the method has excellent accuracy and processing speed for the classification and extraction of large-scale point cloud data acquired by Lidar. [ABSTRACT FROM AUTHOR]

The front-row shading reduction coefficient is a key parameter used to calculate the system efficiency of a photovoltaic (PV) power station. Based on the Hay anisotropic sky scattering model, the variation rule of solar radiation intensity on the surface of the PV array during the shaded period is simulated, combined with the voltage-current characteristics of the PV modules, and the shadow occlusion operating mode of the PV array is modeled. A method for calculating the loss coefficient of front shadow occlusion based on the division of the PV cell string unit and Hay anisotropic sky scattering model is proposed. This algorithm can accurately evaluate the degree of influence of the PV array layout, wiring mode, array spacing, PV module specifications, and solar radiation on PV power station system efficiency. It provides a basis for optimizing the PV array layout, reducing system loss, and improving PV system efficiency. [ABSTRACT FROM AUTHOR]

In view of the problem that both the antenna array and system complexity of large distributed synthetic microwave radiometer system increase with the improvement of imaging resolution, an imaging algorithm based on data fusion for synthetic aperture microwave radiometer with high-resolution is proposed. Firstly, the data fusion model is established. And then the data fusion algorithm is discussed in detail. Finally, a series of ground validation imaging tests are carried out and some imaging results are presented. Both the theoretical and experimental results demonstrate that the proposed algorithm is feasible and can effectively improve the imaging resolution. Moreover, it provides a new solution for enhancing the spatial resolution of microwave radiometric remote sensing from the perspective of data dimension. [ABSTRACT FROM AUTHOR]

Aiming at the problems of low mapping accuracy, slow path planning efficiency, and high radar frequency requirements in the process of mobile robot mapping and navigation in an indoor environment, this paper proposes a four-wheel drive adaptive robot positioning and navigation system based on ROS. By comparing and analyzing the mapping effects of various 2D-SLAM algorithms (Gmapping, Karto SLAM, and Hector SLAM), the Karto SLAM algorithm is used for map building. By comparing the Dijkstra algorithm with the A* algorithm, the A* algorithm is used for heuristic searches, which improves the efficiency of path planning. The DWA algorithm is used for local path planning, and real-time path planning is carried out by combining sensor data, which have a good obstacle avoidance performance. The mathematical model of four-wheel adaptive robot sliding steering was established, and the URDF model of the mobile robot was established under a ROS system. The map environment was built in Gazebo, and the simulation experiment was carried out by integrating lidar and odometer data, so as to realize the functions of mobile robot scanning mapping and autonomous obstacle avoidance navigation. The communication between the ROS system and STM32 is realized, the packaging of the ROS chassis node is completed, and the ROS chassis node has the function of receiving speed commands and feeding back odometer data and TF transformation, and the slip rate of the four-wheel robot in situ steering is successfully measured, making the chassis pose more accurate. Simulation tests and experimental verification show that the system has a high precision in environment map building and can achieve accurate navigation tasks. [ABSTRACT FROM AUTHOR]

This study introduces active control into a Levy cable dome structure, designing and fabricating a model with active struts and length-adjustable cable. The research investigates the structure's capacity for internal force optimization and control under various load conditions through experiments and simulations. Initially, internal forces are optimized using the intelligent algorithms, from which the optimal actuation plans are determined through comparison. Experiments verify the feasibility and the correctness of these plans. Results show that, after assembly and tensioning, the prestress distribution of the refined test model aligns closely with theoretical values, with a maximum error of only 5.06%. Under symmetric and asymmetric overload conditions, average internal force reductions from experiments and simulations are about 40.00% and 50.00%, respectively. However, a notable increase in total actuation lengths is observed in asymmetric overload conditions compared to symmetrical one, highlighting the necessity for higher actuation quantity under asymmetric loads to achieve equivalent internal force optimization and control. • The structure's internal force optimization and control capabilities were studied through experiments and simulations. • Effective optimization and control of internal forces in the cable dome structure were achieved under various overload conditions. • Larger actuation quantities are needed under asymmetric overload conditions to achieve the same level of internal force optimization and control. [ABSTRACT FROM AUTHOR]

Aiming at the problem of vehicle model tracking error and overdependence in traditional path planning of intelligent driving vehicles, a path planning method of intelligent driving vehicles based on deep reinforcement learning is proposed. Firstly, the abstract model of real environment is extracted. The model uses deep reinforcement learning end-to-end strategy (DRL-ETE) and vehicle dynamics model to train the reinforcement learning model which approaches the optimal intelligent driving. Secondly, the real scene problem is transferred to the virtual abstract model through the model transfer strategy, and the control and trajectory sequences are calculated according to the trained deep reinforcement learning model in the environment. Finally, the optimal trajectory sequence is selected according to the evaluation function in real environment. Because the storage mode of experience playback mechanism of Deep Q-Network algorithm is FIFO, and the sampling mode of later playback training is average sampling, the efficiency of experience playback is low. These two problems lead to the slow process of intelligent driving vehicle to target and route finding. And because of greedy strategy, the information of exploration map is incomplete, and IDQNPER algorithm model is proposed. When storing samples, the samples are given weight, and sent to the network in priority order for sample training. Meanwhile, the importance data sequence is retained in the experience playback cache area, and the sequence with high similarity is removed. The total reward value is about 10% higher than the reward value of original Deep Q-network, which proves that the accuracy of intelligent driving vehicles tends to target points is higher. In order to further realize the autonomous decision-making of intelligent driving vehicles and solve the problem of relying too much on map information in the traditional human planning framework, an end-to-end path planning method is proposed based on the depth reinforcement learning theory, which maps the action instructions directly from the sensor information and then issues them to the intelligent driving vehicles. Firstly, CNN and LSTM are used to process radar and camera information. By comparing the advantages of DQ, Double DQN, Dueling DQN and PER algorithm, IDQNPER algorithm is used to train the automatic path planning of intelligent driving vehicles. Finally, the simulation and verification experiments are carried out in the static obstacle environment. The test results show that IDQNPER algorithm is adaptable to intelligent vehicles in different environments. The method can deal with the continuous input state and generate the continuous control sequence of the corner control, which can reduce the lateral tracking error. At the same time, the generalization performance of the model can be improved and the overdependence problem can be reduced by experience playback. [ABSTRACT FROM AUTHOR]

SAR slow-moving target detection and velocity estimation in ultra-high speed platform and strong clutter are a serious problem. The existing slow-moving target velocity estimation methods including fractional Fourier transform (FrFT), bidirectional imaging mode (DiBi) and the dual-beam interferometry mode (DBI) cannot suppress the clutter, which reduces the accuracy of velocity estimation. To solve the problem, we propose an accurate azimuth velocity estimating method refocusing iterative optimization (RIO) based on the quad-beam mode. Firstly, we use the back projection (BP) algorithm for the quad-beam imaging, and perform accurate registration and clutter suppression for the two forward (fore-) and two afterward (aft-) beams respectively. After that, the azimuth offset in fore- and aft-beam images is utilized to achieve the coarse azimuth velocity estimation. Then we put forward the RIO algorithm to compensate the error in velocity estimation. Through iterative search and compensation the azimuth velocity, the azimuth offset is gradually reduced to disappear, thus a more accurate estimation of azimuth velocity is realized. The experiments validate the accuracy of azimuth velocity estimation of the proposed method and the error is less than 0.01 m s − 1 , which is much better than the other three typical methods (FrFT, DiBi, DBI). [ABSTRACT FROM AUTHOR]

In this paper, a novel distributed model predictive control (DMPC) strategy based on voltage observer for multiple energy storage systems (ESs) is firstly proposed to achieve a tradeoff between voltage regulation and power sharing. In order to reduce the impact of communication delay on voltage observer, an improved DMPC consensus algorithm is proposed. This scheme improves the robustness of the system to the delay. Besides, a small-signal dynamic model of a DC microgrid with the improved DMPC controllers is established to analyze its dynamic performance. The time-domain simulation results are presented to verify that, under the constraints of the DC bus voltage, a flexible trade-off between voltage regulation and power distribution can be achieved. Besides, the communication delay problems can be reduced. [ABSTRACT FROM AUTHOR]

To evaluate the separate impacts on human health and establish effective control strategies, it is crucial to estimate the contribution of outdoor infiltration and indoor emission to indoor PM2.5 in buildings. This study used an algorithm to automatically estimate the long‐term time‐resolved indoor PM2.5 of outdoor and indoor origin in real apartments with natural ventilation. The inputs for the algorithm were only the time‐resolved indoor/outdoor PM2.5 concentrations and occupants' window actions, which were easily obtained from the low‐cost sensors. This study first applied the algorithm in an apartment in Tianjin, China. The indoor/outdoor contribution to the gross indoor exposure and time‐resolved infiltration factor were automatically estimated using the algorithm. The influence of outdoor PM2.5 data source and algorithm parameters on the estimated results was analyzed. The algorithm was then applied in four other apartments located in Chongqing, Shenyang, Xi'an, and Urumqi to further demonstrate its feasibility. The results provided indirect evidence, such as the plausible explanations for seasonal and spatial variation, to partially support the success of the algorithm used in real apartments. Through the analysis, this study also identified several further development directions to facilitate the practical applications of the algorithm, such as robust long‐term outdoor PM2.5 monitoring using low‐cost light‐scattering sensors. [ABSTRACT FROM AUTHOR]

Feature selection (marker gene selection) is widely believed to improve clustering accuracy, and is thus a key component of single cell clustering pipelines. Existing feature selection methods perform inconsistently across datasets, occasionally even resulting in poorer clustering accuracy than without feature selection. Moreover, existing methods ignore information contained in gene-gene correlations. Here, we introduce DUBStepR (Determining the Underlying Basis using Stepwise Regression), a feature selection algorithm that leverages gene-gene correlations with a novel measure of inhomogeneity in feature space, termed the Density Index (DI). Despite selecting a relatively small number of genes, DUBStepR substantially outperformed existing single-cell feature selection methods across diverse clustering benchmarks. Additionally, DUBStepR was the only method to robustly deconvolve T and NK heterogeneity by identifying disease-associated common and rare cell types and subtypes in PBMCs from rheumatoid arthritis patients. DUBStepR is scalable to over a million cells, and can be straightforwardly applied to other data types such as single-cell ATAC-seq. We propose DUBStepR as a general-purpose feature selection solution for accurately clustering single-cell data. Cell-type-specific genes are often strongly correlated in expression - an informative yet underexplored property of single-cell data. Here, the authors leverage gene expression correlations to develop DUBStepR, a feature selection method for accurately clustering single-cell data. [ABSTRACT FROM AUTHOR]

To protect the converters and minimize the power transmission interruption during dc line faults, it is necessary to detect the faults at ultra high speed for the modular multilevel converter (MMC)-based HVdc grids. This article reviews the state-of-the-art of dc fault protection methods of MMC-HVdc grids, and summarizes the underlying principles of each method. On this basis, the dc fault characteristics analysis in terms of modal-domain, time-domain, and frequency-domain are analyzed, which direct the protection design. Typical boundary protection and nonboundary protection schemes are reviewed. The advantages and disadvantages of existing fault protection methods are compared. A two-terminal MMC-HVdc prototype is developed to test the effectiveness of three fault protection methods. Comprehensive quantitative assessments of the protection methods discussed above are carried out in a four-terminal MMC-HVdc grid. Finally, the future trends of the protection schemes are discussed and the findings are concluded. [ABSTRACT FROM AUTHOR]

The advancement of new energy vehicles has led to more demanding standards for detecting defects in cylindrical coated lithium batteries. The current research lacks robustness and has low performance. This paper seeks to provide real-time defect identification in cylindrical coated lithium batteries and improve the object detection method of the YOLOv5s model. This paper presents an MGSEC3 module with multi-scale feature extraction and integration of the SENet network in the YOLOv5 Backbone network. This module aims to reduce computational burden and enhance feature extraction efficiency as much as possible. At the same time, the CARAFE operator is utilized to enhance the up-sampling operator in order to reduce the loss of feature information. In addition, enhancements to the loss function enhance both the detection performance and convergence speed. The enhanced YOLOv5s model achieved an average detection rate of 82.4% on the custom cylindrical coated lithium battery dataset, 2.3% higher than the original YOLOv5s. This paper significantly enhances the precision and efficiency of flaw identification in cylindrical coated lithium batteries. [ABSTRACT FROM AUTHOR]

We simulate the self-assembly of active colloidal molecules from binary mixtures of spherical particles using a Brownian dynamics algorithm. These particles interact via phoretic interactions, which are determined by two independently tunable parameters, surface activity and surface mobility. In systems composed of equal-size particles, we observe the formation of colloidal molecules with well-defined coordination numbers and spatial arrangement, which also display distinct dynamic functions, such as resting, translating, and rotating. By changing the size ratio to 2 : 1 between the two species, we further observe the formation of colloidal molecules with new structures arising from breaking the size symmetry. By tuning the mutual interactions between the smaller species via their surface mobility, we are able to control their spacing as well as the coordination number of the colloidal molecules. This study highlights the importance of tuning surface parameters and size asymmetry in controlling the structure and the active dynamics of colloidal molecules. [ABSTRACT FROM AUTHOR]