Search

This correspondence discusses recognition of traffic modes and behavior patterns based on Global Navigation Satellite System (GNSS) data. The traffic modes (e.g., walk, car, train, etc.) are firstly inferred, and then their behavior patterns (e.g., left-turn, right-turn, turn-around, etc.) are further identified. Because both traffic modes and behavior patterns are strongly influenced by geographical circumstances, their recognitions are enhanced by geographical layer information (e.g., building, road, water, etc.). At one specific GNSS point, its surrounding area is uniformly sliced as grids, and the probabilities for grid centers belonging to six different geographical layers are calculated based on whether these centers lie inside or outside of the minimum rectangles containing polygons indicating different geographical objects. Finally, the six-dimensional probability matrix is processed and compressed as a geographical information vector by the convolutional neural network (CNN). The latter is then combined with kinematic metrics such as velocity, acceleration, and moving direction from GNSS data, and serially input into a long short-term memory (LSTM) network to predict traffic modes and behavior patterns. Experimental results validate that the geographical information does enhance the performances of two recognition tasks. The CNN+LSTM framework retains the powers of CNN and LSTM, and outperforms classical machine learning algorithms.

Recent 3-D scanning techniques can produce various kinds of digitized 3-D data. Most of these scanned data are in a format of unstructured point clouds. Such low-level representation of 3-D data usually contains only geometric properties (point positions), while lacking higher level structure cues, for example, feature lines. Feature lines can be defined as a visually prominent characteristic of the shape, including edges, ridges, and valley lines in multiple scales, which can support a lot of downstream applications, such as shape reconstruction and analysis. We present a two-phase algorithm for extracting line-type features on point clouds. To extract both large-scale and shallow feature lines, we first define a statistical metric to detect all potential feature points while immune to the noise to some extent. Then, for correctly reconstructing the feature lines from these identified coarse feature points, we introduce an anisotropic contracting scheme to force feature points lying on the underlying real feature lines. To illustrate the reliability of our method, various experiments have been conducted on both synthetic and raw data. Both visual and quantitative comparisons show that our method is robust to noise and can correctly extract multiscale feature lines. In addition, our method is generally applicable to robotic picking.

Large scale aerial deployment of miniature sensors in tough environmental conditions requires a deployment device that is lightweight, robust, and steerable. We present a novel samara-inspired autorotating craft that is capable of two flight modes (autorotating mode and diving mode) with an average glide angle of 28.9 $^{\circ }$ (1.81 m lateral distance per 1 m loss of altitude) in the former mode. The bidirectional transition between the two modes and directional control is achieved by using only a single actuator. Also, in order to minimize its glide angle, a design optimization methodology is presented for our prototype, diving samara autorotating wing, along with a new cyclic control strategy for directional control of autorotating descent. The dynamic model, simulated in a six degrees-of-freedom environment using the blade element theory, is integrated with genetic algorithm to derive parameters for the wing geometry, flap angle for autorotation, and the proposed cyclic control. The physical prototype autorotates at a descent velocity of 1.43 m/s and rotation speed 4.17 Hz, and is able to transit to diving mode in an average duration of 272 ms to increase its descent velocity by at least 17.6 times. At any point during the dive, it is able to transit back into autorotation in an average duration of 327 ms. Semioutdoor experiments were used to investigate the bidirectional transitions and verify the glide angle (28.9 $^{\circ }$ ), which is much improved from the previous prototype (SAW+, 58.4 $^{\circ }$ ). Lastly, as a demonstration of a real-life deployment scenario and environmental conditions, the prototypes were dropped from a fixed-wing unmanned aerial vehicle at a suburban test site.

Robotic hand exoskeletons can assist people who suffer from hand functional disabilities caused by a stroke. However, currently existing hand exoskeletons remain inadequate in terms of user-friendly design, lightweight structure, and accurate modeling of hand motion. In this study, a large displacement compliant mechanism is developed as a finger exoskeleton. First, a topology configuration is developed which contains information on hand biomechanics and thus yields an efficient force-deflection mechanism for a finger. A pseudo-rigid-body model is then established using a large-deformation flexible hinge to solve the point flexure problem to create a finger exoskeleton that functions as intended. Finally, the mechanism synthesis method is used for parameter optimization to ensure that the output port of the device matches the motion of the finger. Using the proposed method, an integrated compliant finger exoskeleton is developed, and the performance of the exoskeleton is evaluated. The experimental result shows that the finger exoskeleton prototype provides suitable movement guidance for the index finger, which demonstrates that the proposed method may be a good way in the design of hand exoskeletons.

Indoor localization as one of the visions of the Internet of Things (IoTs) receives much attention in Wireless Sensor Networks (WSNs). Many prior indoor localization technologies concentrated on multiple Access Points (APs), so developing a system that uses a single AP with high accuracy is indispensable, especially for households or small businesses. To this end, we present HiLoc, a single AP-based indoor localization system with sub-meter accuracy. The key insight of HiLoc is to estimate the relative Time of Flight (ToF) between the target and the distributed antennas and localize the target using the convex optimization accurately. Specifically, HiLoc utilizes the combination of Orthogonal Matching Pursuit (OMP) and Maximum Likelihood (ML) to estimate relative ToF. Then, by applying the semi-definite relaxation technique, the localization process can be transformed into an SDP problem. Unlike the methods that use the iterative least-square, in which an accurate guess of the target location is required, we can find the target location accurately without any initializations. Furthermore, we design the prototype AP using Software-Defined Radio (SDR) platform with Orthogonal Frequency Division Multiplexing (OFDM) signals. Our implementation in extensive indoor environments shows that HiLoc achieves a median localization error of 0.72m at 100MHz signal bandwidth, comparable to the state-of-the-art prior systems.

In this paper, we develop a pedestrian inertial navigation (PIN) that uses an inertial measurement unit (IMU) mounted on a shoe to locate a pedestrian in an infrastructure-free environment. In the PIN, stand-alone navigation is possible by calculating the navigation solution using the inertial navigation system (INS) mechanism. For accurate navigation, however, it is necessary to compensate for accumulated errors over time. In a foot-mounted PIN, the error can be compensated through zero-velocity (ZV) correction method by detecting the moment the foot touches the ground. One important point here is to detect ZVs accurately. In most of the literature related to this, ZV detection techniques for normal gait have been proposed. In this paper, we first propose a powerful ZV detection algorithm for various gait types. This algorithm consists of a new buffering method of the sensor output power, an exact ZV detection method through glitch removal, and a reinforcement method for ZV detection based on auxiliary step detection. Second, we design a PIN/ZV integration filter, and describe the real-time operation plan of all these functions. The performance of the PIN using the proposed algorithm is verified through experiments.

Wi-Fi Received Signal Strength (RSS, fingerprints) based indoor localization is promising and widely investigated with the pervasive deployment of Wi-Fi Access Points. However, the process to collect RSS, also known as site survey, is labor-intensive. Thus, we propose and demonstrate an efficient fingerprint crowdsourcing method in this paper. Specifically, RSS measurements are obtained and annotated with location tags while a participant is walking along a chosen path with a smartphone at hand. In the localization stage, we adopt the Gaussian Process based solution and propose a novel mean function selection method. Extensive experiments show that the path-based site survey can achieve a comparable localization performance to the point-based site survey, but takes less survey time. We find that fingerprints collected while walking are more suitable for localizing moving pedestrians. In addition, due to the sparsity of fingerprints collected through crowdsourcing, the proposed mean function selection strategy is advantageous and can reduce localization errors significantly compared to a baseline solution.

This paper addresses sampled-data control of 2D Kuramoto-Sivashinsky equation over a rectangular domain. We suggest to divide the 2D rectangular into N sub-domains, where sensors provide spatially averaged or point state measurements to be transmitted through communication network to the controller. Note that differently from 2D heat equation, here we manage with sampled-data control under point measurements. We design a regionally stabilizing controller applied through distributed in space characteristic functions. Sufficient conditions ensuring regional stability of the closed-loop system are established in terms of linear matrix inequalities (LMIs). By solving these LMIs, we find an estimate on the set of initial conditions starting from which the state trajectories of the system are exponentially converging to zero. A numerical example demonstrates the efficiency of the results.

Many CNN-based segmentation methods have been applied in lane marking detection recently and gain excellent success for a strong ability in modeling semantic information. Although the accuracy of lane line prediction is getting better and better, lane markings' localization ability is relatively weak, especially when the lane marking point is remote. Traditional lane detection methods usually utilize highly specialized handcrafted features and carefully designed postprocessing to detect the lanes. However, these methods are based on strong assumptions and, thus, are prone to scalability. In this work, we propose a novel multitask method that: 1) integrates the ability to model semantic information of CNN and the strong localization ability provided by handcrafted features and 2) predicts the position of vanishing line. A novel lane fitting method based on vanishing line prediction is also proposed for sharp curves and nonflat road in this article. By integrating segmentation, specialized handcrafted features, and fitting, the accuracy of location and the convergence speed of networks are improved. Extensive experimental results on four-lane marking detection data sets show that our method achieves state-of-the-art performance.

The area of this study will cover the location-wise seismic response variation of an electrical cabinet in nuclear power point (NPP) based on classical reliability analysis. The location-based seismic ground motion (GM) selection is carried out with the help of probabilistic seismic hazard analysis using PSHRisk-tool, where the variation of reliability analysis can be understood from the relation between the reliability index and intensity measure. Two different approaches such as the first-order second moment method (FOSM) and Monte Carlo Simulation (MCS) are helped to evaluate and compare the reliability assessment of the cabinet. The cabinet is modeled with material uncertainty utilizing Steel01 as the material model and the fiber section modeling approach is considered to characterize the section's nonlinear reaction behavior. To verify the modal frequency, this study compares the FEM result with recorded data using Least-Squares Complex Exponential (LSCE) method from the impact hammer test. In spite of a few investigations, the main novelty of this study is to introduce the reader to check and compare the seismic reliability assessment variation in different seismic locations and for different earthquake levels. Alongside, the betterment can be found by comparing the result between two considered reliability estimation methods.

View-synthesis-based methods have shown very promising results for the task of unsupervised depth estimation in single images. Most existing approaches synthesize a new image and employ it as the supervision signal for depth and pose prediction. There are two problems in these approaches: 1) There are many combinations of pose and depth that can synthesize a certain new image; therefore, reconstructing the depth and pose based on the view-synthesis method from only two images is an inherently ill-posed problem; 2) The model is trained under the photometric consistency assumption that the brightness or gradient is constant when applied to the video sequences. However, this assumption is easily violated in realistic scenes due to light changes, reflective surfaces and occlusions. To overcome the first drawback, we exploit the point cloud consistency constraint to eliminate ambiguity. To overcome the second drawback, we use threshold masks to filter dynamic and occluded points and introduce matching point constraints that implicitly encode the geometry relationship between two matched points to improve the precision of depth prediction. In addition, we employ epipolar constraints to compensate for the instability of the photometric error in textureless regions and varying illumination conditions. The experimental results on the KITTI, Cityscapes and NYUv2 datasets show that the method can improve the accuracy of depth prediction and enhance the robustness of the model in handling textureless regions and illumination changes. The code and data are available at https://github.com/XTUPRLAB/FixUnDepth.

This paper studies a class of distributed attack strategies against state estimation of wireless sensor networks. The attack objective is to corrupt the least numbers of sensors so that the state estimate error approaches to a target bias while avoiding being detected and a high attacking lost. First, with precise knowledge of measurement model and applying a sparsity projection operation with average linear complexity rather than the popular brute force search and approximation techniques, a distributed optimization scheme is proposed which provides to an exact and low-complexity algorithmic solution to finding the optimal attack strategy. Furthermore, by introducing dead-zone type projection and region projection operators, a distributed optimization algorithm for robust sparse undetectable attacks is proposed with imprecise model knowledge. A distinguishing point of the proposed algorithms is that a trade-off between computational complexity and attacks impact can be achieved by properly designing a cost function. Simulation results on an unmanned ground vehicle are presented to substantiate the algorithm.

In this paper, an optimized rectangular corrugated tube is used for the super-critical water reactor (SCWR) cooling, numerically. Al2O3, CuO, Fe3O4 and TiO2 nanoparticles in different shapes (spherical, brick, platelet and cylindrical) are considered as additives to SCW for cooling purposes, and it is tried to introduce the most suitable cases from the maximum heat transfer view point. Piecewise temperature-dependent properties are considered for the SCW and different correlations such as Brinkman and Pak equations were examined for nanofluid properties. The governing equation are solved based on control volume method (CVM) software, where the SCW properties are introduced to it by a user defined function (UDF). As the main outcome, Pak and Cho correlation had the best fitted curves with other presented functions. Also, the corrugated tube filled by spherical alumina-SCW had the best performance for SCWR cooling.

This paper is concerned with the recovery of (approximate) solutions to parabolic problems from incomplete and possibly inconsistent observational data, given on a time-space cylinder that is a strict subset of the computational domain under consideration. Unlike previous approaches to this and related problems our starting point is a regularized least squares formulation in a continuous infinite-dimensional setting that is based on stable variational time-space formulations of the parabolic partial differential equation. This allows us to derive a priori as well as a posteriori error bounds for the recovered states with respect to a certain reference solution. In these bounds the regularization parameter is disentangled from the underlying discretization. An important ingredient for the derivation of a posteriori bounds is the construction of suitable Fortin operators which allow us to control oscillation errors stemming from the discretization of dual norms. Moreover, the variational framework allows us to contrive preconditioners for the discrete problems whose application can be performed in linear time, and for which the condition numbers of the preconditioned systems are uniformly proportional to that of the regularized continuous problem. In particular, we provide suitable stopping criteria for the iterative solvers based on the a posteriori error bounds. The presented numerical experiments quantify the theoretical findings and demonstrate the performance of the numerical scheme in relation with the underlying discretization and regularization.

The efficiency calculated from the conventional slacks-based measure (SBM) model is the minimum in terms of efficiency scores, and the corresponding target point requires more effort for an inefficient decision making unit (DMU) to become efficient. To find the closest target point to the assessed DMU, several models have been proposed via calculating the maximum efficiency. This paper proposes a set of models to measure the maximum SBM efficiency for DMUs with multi-period data. It is found that, when the non-Archimedean number is not in effective in claculating the radial efficiency, the maximum SBM efficiency is greater than or equal to the radial efficiency. The corresponding target point is easier for an inefficient DMU to reach. This efficiency is more reliable than radial efficiencies because the latter may be affected by the non-Archimedean number. Furthermore, the SBM efficiency can be decomposed into the product of the input and output efficiencies. Under the different-multiplier models, the DMU efficiency can also be decomposed into a weighted average of the individual year efficiencies. The factor and year that cause the unsatisfactory performance can thus be detected. A case of Taiwanese commercial banks is used as an illustration.

Traditional fault diagnosis methods are easy to be affected by different working conditions. This paper proposed a motor fault diagnosis method based on visual knowledge, to reduce the impact of changes in working conditions and improve the feature extraction ability. The mapping relationship between actual faults and image intuitive features by symmetrized dot pattern (SDP) and scale invariant feature transform (SIFT) is established in this paper. The fault state is obtained by statistics of the matching point with the dictionary templates generated from signals of normal and unnormal motors. Compared with other machine learning algorithms, this method does not need too much data training and learning. The efficiency of this method is validated by experiments, and the data image processing technology has great industrial application value in the field of motor fault detection or monitoring in the age of intelligence.

In this study, we provide a characterization of the class of proportional bargaining solutions introduced by Kalai. Our result differs from earlier axiomatizations in that we use a property that we label individual disagreement point concavity. This property is a weakening of disagreement point concavity used by Chun and Thomson. An application illustrates the potential usefulness of our new property in a strategic setting.

Deep neural networks have been scaled up to thousands of layers with the intent to improve their accuracy. Unfortunately, after some point, doubling the number of layers leads to only minor improvements, while the training difficulties increase substantially. In this paper, we present an approach for constructing high accuracy deep evolutionary networks and train them by activating and freezing dense networks (AFNet). The activating and freezing strategy enables us to reduce the classification error of test and reduce the training time required for deeper dense networks. We activate the layers which are being trained and construct a Freezing Box to freeze the idle and pre-trained network layers in order to minimize the memory consumption. The training speed in the early stage is not fast enough because many layers are activated for training. As the epochs gradually increase, the training speed becomes faster and faster since fewer and fewer layers are activated. Our method improves the convergence to the optimal performance within a limited number of epochs. Comprehensive experiments on a variety of datasets show that the proposed model achieves better performance when compared to the other state-of-the-art network models.

To date, numerous cooperative merging techniques have been developed, with multiple studies demonstrating their potential to improve traffic safety, efficiency, and environmental impact at merging points. However, many of these techniques make assumptions that will almost certainly be questionable in the real world, namely: 1) relative positioning between vehicles will be sufficiently accurate at all times; 2) the merging application-equipped vehicles will not change lanes before reaching the merge point, nor will other vehicles change lanes into their lane; and 3) vehicles with merging-assist technology will not mix with vehicles without the technology. This paper proposes a cooperative merging application which is capable of functioning even when the vehicles positioning accuracy is severely degraded and all vehicles are free to change lanes into and out of the merging traffic streams. Furthermore, our strategy demonstrates safety, efficiency, and environmental benefits for the overall traffic stream even with technology penetration rates as low as 20%.

The design and analysis of megawatt power electronics systems calls for an accurate and efficient simulation approach that can consider the switching transients. However, the state-of-the-art simulation methods and tools exhibit low efficiency in megawatt power electronics systems, even with ideal switch model. It is hardly possible to perform physical-switch-model-based simulation due to the time cost and the convergence issue. This paper proposes a switching transient simulation approach based on the discrete state event-driven approach. The proposed method is composed of a switch-pair-based modeling method and a time-scale-decoupling simulation method. In the example of a 2 MW power electronics transformer composed of 576 semiconductor switches, the proposed method enables the switching transient simulation of a 0.2-second dynamic within 10 minutes and is free of convergence problems. The speed is 20-fold faster than commercial software adopting ideal switch model. The accuracy is attested to by comparisons with both experimental and commercial software results. Meanwhile, this paper demonstrates the application in system efficiency evaluation. Due to the speed gain and the ability of capturing switching transients, the proposed method enables simulation of the efficiency curve, and hence the locating of the maximum efficiency point and optimized design of the conversion system.

For the landing legs with single air chamber in the buffer structure of the reusable landing vehicle, the geometric topological models and the dynamic model associated with the hard points of the landing legs are established. The geometric constraint relationship in the design of the landing legs is also obtained. The whole vehicle dropping test is conducted, and the test results agree well with that of the simulation model, which validates the dynamic model. Based on the verified model, the effect of hard point positions on the performance of the landing system is analyzed. The multidisciplinary collaborative optimization algorithm and archive-based micro genetic algorithm (CO-AMGA) are used to optimize the design parameters that contain the hard points and the damper. Compared with artificial iteration, the maximum landing impact acceleration response of the vehicle and the buffer struct maximum force are reduced by 30.70% and 14.51% respectively, and the maximum length of retractable pillar decreases by 8.54% while the design margin increases by 69.11%. The proposed optimization method is efficient and can greatly facilitate the design of landing legs.

WiFi fingerprinting-based indoor positioning system is vulnerable to the dynamic environment, which makes the positioning accuracy decrease and the fingerprint map invalid. To address these issues, An Accurate Indoor Positioning System (AIPS) with fingerprint map adaptation is proposed. For online positioning, it treats the received signal strength (RSS) from each access point (AP) individually and can be divided into two steps, coarse and fine positioning. In coarse positioning, a novel clustering algorithm based on RSS attenuation is proposed. In fine positioning, signal noise is considered to construct AP ring, and the reference point (RP) contained by the largest number of rings is selected as nearest RP, then the RPs with larger number are searched out by region growing algorithm to estimate the location of test point (TP). For the fingerprint map adaptation, K-means is adopted to divide APs into two types, based on the number of rings, and to find which APs’ information has been changed by dynamic environment, and then update them through Gaussian process regression. Experimental results show that the positioning algorithm in AIPS can improve the positioning accuracy compared with other algorithms, and the fingerprint map adaptation scheme in AIPS can reduce the online running time while keep accuracy.

In view of the present situation that most aggregation methods of fuzzy preference information are extended or mixed by classical aggregation operators, which leads to the aggregation accuracy is not high. The purpose of this paper is to develop a novel method for spatial aggregation of fuzzy preference information. Thus we map the fuzzy preference information to a set of three-dimensional coordinate and construct the spatial aggregation model based on Steiner-Weber point. Then, the plant growth simulation algorithm (PGSA) algorithm is used to find the spatial aggregation point. According to the comparison and analysis of the numerical example, the aggregation matrix established by our method is closer to the group preference matrices. Therefore, the optimal aggregation point obtained by using the optimal aggregation method based on spatial Steiner-Weber point can best represent the comprehensive opinion of the decision makers.

We consider a resource allocation and offloading decision-making problem in a mobile edge computing (MEC) network. Since the locations of user equipments (UEs) vary over time in practice, we consider a dynamic network, where the UEs could leave or join the network coverage at any location. Since the joint offloading decision that minimizes the network cost also varies with the topology, the expected best offloading decision for the previous topology would not match the new topology. Consequently, the system suffers from recurring cost peaks due to the topology change. Thus, we propose a robust distributed hierarchical online learning approach to enhance the algorithm’s robustness and reduce the cost peaks. Specifically, the UEs learn the utility of each offloading decision via deep Q networks (DQNs) from their interaction with the MEC network. Meanwhile, the computational access points (CAPs) train their deep neural networks (DNNs) online with the real-time data collected from the UEs to predict their corresponding Q-value vectors. Therefore, the UEs and CAPs form a hierarchical collaborative-learning structure. When the topology changes, each UE downloads its Q-value vector as the Q-bias vector and learns its difference from the actual Q-value vector via its DQN. With different agents learning distributedly, both the peak and sum costs are reduced as the joint offloading decision could start from a near-local-optimal point. In simulations, our robust approach successfully reduces the peak cost and sum cost by up to 50% and 30%, respectively. This demonstrates the need for a robust learning algorithm design in a practical dynamic MEC network.

An analytical modeling approach is developed to compute the unsteady force on a foil, with any specified chordwise thickness profile, due to a passing point vortex. The model uses a generalized Jou...

In this paper we provide a general Prekopa-Leindler type inequality for Sugeno integral and any measurable functions. In addition, we apply novel inequalities to study some properties of monotone measures. As a by-product, we point out that the result from [22] , stating that the classical Prekopa-Leindler inequality is not valid for the Sugeno integral with respect to the Lebesgue measure, is not true and we provide its correction.

This article presents a timing error detection and correction (EDaC) technique optimized for near-/sub-threshold operation to recover energy lost in the conventional signoff margins. The presented EDaC requires no modifications to the processor pipeline and avoids imposing additional hold constraints on monitored paths by instantaneously checking for late activity. Furthermore, two correction methods are discussed: a simple clock gating method and a low cycle overhead clock stretching method. Both provide robust last-minute error prevention. The EDaC is applied in a near-/sub-threshold implementation of the CoolFlux DSP processor and infers only a 2.8% and 2.1% area overhead for the detection and correction, respectively. Silicon measurements validate the EDaC system from 0.25 to 0.7 V (1-200 MHz) and show that it recovers all voltage margins in the near-/sub-threshold region. The design achieves a minimum energy point (MEP) of 8.1 pJ/cycle at 0.34 V and 10 MHz. Here, the EDaC technique reduces the energy consumption by 48%-17.6% with respect to the signoff margins depending on their conservatism, and it enables the processor to operate with only 12% energy overhead compared to its ideal non-margined critical operation point.

We present well known properties related to the topology of Steiner minimal trees and to the geometric position of Steiner points, and investigate their application in the exact algorithms that have been proposed for the Euclidean Steiner problem. We discuss the difficulty in the application of properties that were very successfully applied to solve the problem in the plane, when the dimension of the space increases, and point out that the application of some geometric conditions for Steiner points is hindered when the well known implicit enumeration scheme proposed by Smith in 1992 is considered. Finally, we mention mathematical-optimization methods as a direction to explore in the search for good formulations of inequalities that would allow the application of these restrictive geometric conditions.

Multi-sided faces arise in polyhedral modeling through introduction of features not aligned with the regular grid structure, e.g. when trimming off corners or merging primary shapes. A standard first step is to split the n -gon into n quadrilaterals that join at the n -gon’s centroid. A canonical example is the ‘face point rule’ of Catmull–Clark subdivision. We show that rules negatively impact shape — already in the single, first refinement step. A new alternative face (and edge) rule improves the shape of surfaces controlled by the refined mesh noticeably. The improved face rule has a moderately increased footprint compared to the Catmull–Clark rule but remains simple in that it is characterized by just one scalar weight for each n .

In this article, a comprehensive method for parametric representation of wire trajectories, allowing accurate geometric description of complex and arbitrarily oriented wire bundles, is introduced. This is the starting point to develop a computationally efficient numerical transmission-line (TL) model for predicting the radiated susceptibility of arbitrarily oriented bundles of wires, illuminated by (possibly) nonuniform electromagnetic fields. The proposed method foresees solution of the field-to-wire coupling problem through suitable discretization and sampling of the bundle geometry and the incident electromagnetic field. Differently from previous models, where bundles parallel to ground were assumed, the proposed model allows for arbitrary bundle orientation by exploiting, first, exact projection of the external field onto the bundle direction, and second, evaluation of the actual wire length (instead of the longitudinal one) of each TL section. Accuracy and computational efficiency of the proposed method are assessed versus full-wave simulation for two application examples, involving parabola-shaped and trefoil knot-shaped wiring structures above ground. Although the strong nonuniformity affecting these structures forces TL theory to work very close to its limits, the achieved agreement is satisfactory and the significant reduction of computational times makes the proposed method suitable for approximate yet efficient prediction of radiated susceptibility characteristics of complex wire bundles.

In this analysis, the cubic B-spline method is employed for constructing the approximate solutions of a class of fractional two-point boundary value problems. These fractional problems are expressed in terms of the conformable fractional derivative approach. More deeply, a class of conformable Lane-Emden model is considering and modifying with singularities. Many numerical applications are presented and discussed to exhibit the feasibility and efficiency of the procedure involved in both linear and non-linear cases. The numerical findings are quite similar to those of the exact solutions as well as require relatively less computational work. Latterly, remarks and future research work are provided.

We prove new explicit asymptotic formulae between (geometric) eigenvalues in max-algebra and classical distinguished eigenvalues of nonnegative matrices, which are useful tools for transferring results between both settings. We establish new inequalities for both types of eigenvalues of Hadamard products and Hadamard weighted geometric means of nonnegative matrices. Moreover, a version of the spectral mapping theorem for the distinguished spectrum is pointed out.

In this note, we point out that Theorem 3 (a version of Egoroff's theorem for monotone set-valued measures) shown in the paper “Lusin's theorem for monotone set-valued measures on topological spaces” (Fuzzy Sets and Systems 364 (2019) 111-123) is not valid, and we present two revised versions. We also point out that the proofs of Lemma 2 and Theorem 1 are defective, thus the truth of their conclusions cannot be affirmed. Thereby, an open problem for the characteristics of monotone measures is raised.

The ability to improve both the size and efficiency of multilevel single-phase converters is a key to uplift them as an attractive solution for industries, while the high number of switches and complex modulation techniques understandably make them unattractive. Five-level active neutral point clamped (ANPC) converter, due to its inherent advantages such as employing different switching frequencies and using different switch technologies, presents an ideal candidate for study. This article performs a comprehensive analysis of the converter to highlight the advantages of it. This analysis results in a modified hybrid modulation that effectively regulates the neutral point (NP) of the dc link. Consequently, the combination of the topology and the modified modulation make the converter ideal to utilize two different switch technologies - in this article, silicon carbide (SiC) and Si MOSFET. To evaluate the analysis and the effectiveness of modulation, a 2-kW hybrid five-level ANPC is built. The converter current behavior is analyzed to determine the power loss in the filter and switches. The efficiency measurement is performed and compared with the calculated efficiency. There is a close coherency between the measurement and the calculated results and a peak efficiency of 98.4% is achieved.

Line-of-sight (LOS) is a critical requirement for mmWave wireless communications. In this work, we explore the use of access point (AP) infrastructure mobility to optimize indoor mmWave WiFi network performance based on the discovery of LOS connectivity to stations (STAs). We consider a ceiling-mounted mobile (CMM) AP as the infrastructure mobility framework. Within this framework, we propose two heuristic algorithms (basic and weighted) derived from Hamming distance computation and a machine learning (ML) solution fully exploiting available network state information to address the LOS discovery problem. Based on the ML solution, we then propose a systematic solution WiMove, which can decide if and where the AP should move to for optimizing network performance. Using both ns-3 based simulation and experimental prototype implementation, we show that the throughput and fairness performance of WiMove is up to 119% and 15% better compared with single static AP and brute force search.

The search for more realism in renderings led to increased interest in tabular bidirectional reflectance distribution function (BRDF) samples captured from real-world materials. This representation provides a high degree of realism at the cost of a long acquisition time and high storage space. Because of those drawbacks, the literature presents few BRDF databases captured from real materials, and the available ones present a limited number of samples. Aiming to create new and realistic-looking materials, this work proposes an approach to generate new BRDFs from a tabular BRDF database. Our goal is to provide a pipeline where the user expresses appearance necessities by selecting materials from a BRDF database. We generate an appearance-driven space shaped by the user’s interests, in which it is feasible to navigate and retrieve new tabular BRDFs. This new space should enable smooth navigation and, to provide visual richness and realism, not rely only on the materials selected by the user. To this end, we first preprocess a tabular BRDF database by keeping BRDF samples with relevant reflectance features. A low-dimensional representation of this preprocessed database is obtained by applying a dimensionality reduction method, and a clustering algorithm is applied to obtain clusters of materials with similar features. Based on the materials selected by the user, we define one or more clusters of materials to build the appearance-driven space. Any point in this lower-dimensional space can be mapped to the original BRDFs space. Therefore, new BRDFs that present similar properties to the ones which the user selected are created. We compare the performance of linear (Multidimensional Scaling - MDS) and nonlinear (Isometric Feature Mapping - ISOMAP) dimensionality reduction methods in this task and analyze the obtained materials.

This letter proposes a fast Unmanned Aerial Vehicle (UAV) trajectory planning algorithm based on the Fermat-point theory in maritime Internet of Things systems. Specifically, we first construct Delaunay triangles based on the deployment of unmanned surface vehicles. Then, the Fermat point of each Delaunay triangle is calculated and as a hovering point of UAVs to improve the channel condition. Finally, the trajectory planning problem can be transformed into a vehicle routing problem with pickup, which is solved by the proposed algorithm based on C-W saving method. Simulation results show the proposed algorithm can efficiently increase collected data volume of UAVs.

In this paper, we employ multiple wireless-powered relays to assist information transmission from a multi-antenna access point to a single-antenna receiver. The wireless relays can operate in either the passive mode via backscatter communications or the active mode via RF communications, depending on their channel conditions and energy states. We aim to maximize the overall throughput by jointly optimizing the access point's beamforming and the relays' radio modes and operating parameters. Due to the non-convex and combinatorial structure, we develop a novel optimization-driven hierarchical deep deterministic policy gradient (H-DDPG) approach to adapt the beamforming and relay strategies dynamically. The optimization-driven H-DDPG algorithm firstly decomposes the binary relay mode selection into the outer-loop deep Q-network (DQN) algorithm and then optimizes the continuous beamforming and relaying parameters by using the inner-loop DDPG algorithm. Secondly, to improve the learning efficiency, we integrate the model-based optimization into the DDPG framework by providing a better-informed target estimation for DNN training. Simulation results reveal that these two special designs ensure a more stable learning and achieve a higher reward performance, up to nearly 20%, compared to the conventional DDPG approach., Comment: 29 pages, 7 figures, 2 tables, submitted to IEEE Transactions on Wireless Communications

Camera and LIDAR are both important sensor modalities for real-world applications, especially autonomous driving. The sensors provide complementary information and make sensor fusion possible. However, the progress of early-fusion for both modalities is very slow due to the limitations of viewpoint misalignment, feature misalignment and data volume alignment, so that its performance is also very low. In this work, we propose a novel fusion pipeline: an early-fusion method of range image and RGB image to enhance 3D object detection. It takes full advantage of LIDAR’s range view, point view, bird’s eye view, and RGB view of the camera: First, it uses a large-scale 2D detection data set to pre-train a 2D detection network for extracting high-level semantic features from the RGB image; Second, it uses a specially designed 2D convolution network to extract high-level geometric features from the range image; Third, it fuses the semantic and geometric features through the point view of point clouds; Finally, the point view is transferred to bird’s eye view for 3D object detection. Both the range image and the RGB image are front views, so the multi-modal features are more matched. Experiments show that the proposed method has large improvements on the method of using only the point cloud, and outperforms the state-of-the-art methods on the self-built actual data set. In ablation, we study the dependence of the fusion method on 2D features and the effect of different fusion positions on the performance.

We propose a new method for simulating pedestrian crowd movement in a virtual environment. A crowd consists of groups of different number of people with different attributes such as gender, age, position, velocity, and energy. Each group has its own intention used to generate a trajectory for each pedestrian navigating in the virtual environment. Additionally, an agent - based model is introduced to simulate pedestrian behaviours in the groups, where various steering behaviours are introduced and combined into a single steering force to allow pedestrians in each group to walk toward their destination point. Based on the proposed method, every single pedestrian in each group can continuously adjust their attributes. Moreover, pedestrians optimize their path independently toward the desired goals, while avoiding obstacles and other pedestrians in the scene. This method takes into account the safety-space around each pedestrian to avoid collisions among pedestrians. The proposed method was implemented for several simulation scenarios under various conditions for a wide range of different parameters. Statistical analysis is carried out to evaluate the performance of the proposed method for simulating the crowd movement in the virtual environment. Results indicate that our method can generate each pedestrian’s trajectories in each group independently to reach several goal points within a reasonable computational time. Moreover, the obtained results reveal that the mean value of the computational time is not increased significantly with the increasing of the number of pedestrians in the crowd.

We show that on the generic degree-d hypersurface in P n + 1 , when d ≥ n + 2 , every closed point of degree d is separable, and has monodromy group S d . We study related Franchetta-type questions on generic objects, and give affirmative answers in specific cases.

Automatic registration of point clouds captured by terrestrial laser scanning (TLS) plays an important role in many fields including remote sensing (e.g., transportation management, 3-D reconstruction in large-scale urban areas and environment monitoring), computer vision, and virtual reality and robotics. However, noise, outliers, nonuniform point density, and small overlaps are inevitable when collecting multiple views of data, which poses great challenges to 3-D registration of point clouds. Since conventional registration methods aim to find point correspondences and estimate transformation parameters directly in the original point space, the traditional way to address these difficulties is to introduce many restrictions during the scanning process (e.g., more scanning and careful selection of scanning positions), thus making the data acquisition more difficult. In this article, we present a novel 3-D registration framework that performs in a ``middle-level structural space'' and is capable of robustly and efficiently reconstructing urban, semiurban, and indoor scenes, despite disturbances introduced in the scanning process. The new structural space is constructed by extracting multiple types of middle-level geometric primitives (planes, spheres, cylinders, and cones) from the 3-D point cloud. We design a robust method to find effective primitive combinations corresponding to the 6-D poses of the raw point clouds and then construct hybrid-structure-based descriptors. By matching descriptors and computing rotation and translation parameters, successful registration is achieved. Note that the whole process of our method is performed in the structural space, which has the advantages of capturing geometric structures (the relationship between primitives) and semantic features (primitive types and parameters) in larger fields. Experiments show that our method achieves state-of-the-art performance in several point cloud registration benchmark datasets at different scales and even obtains good registration results for data without overlapping areas.

Unsupervised domain adaptation (UDA) for the semantic segmentation of remote sensing images is challenging since the same class of objects may have different spectra while the different class of objects may have the same spectrum. To address this issue, we propose a class-aware generative adversarial network (CaGAN) for UDA semantic segmentation of multisource remote sensing images, which explicitly models the discrepancies of intraclass and the interclass between the source domain images with labels and the target domain images without labels. Specifically, first, to enhance the global domain alignment (GDA), we propose a transferable attention alignment (TAA) procedure to add more fine-grained features into the adversarial learning framework. Then, we propose a novel class-aware domain alignment (CDA) approach in semantic segmentation. CDA mainly includes two parts: the first one is adaptive category selection, which is to alleviate the class imbalance and select the reliable per-category centers in the source and target domains; the second one is adaptive category alignment, which is to model the intraclass compactness and interclass separability from source-only, target-only, and joint source and target images. Finally, the CDA plays as a penalty of GDA to train GaGAN in an alternating and iterative manner. Experiments on domain adaptation of space to space, spectrum to spectrum, both space-to-space and spectrum-to-spectrum data sets demonstrate that CaGAN outperforms the current state-of-the-art methods, which may serve as a starting point and baseline for the comprehensive applications of semantic segmentation in cross-space and cross-spectrum remote sensing images.

A novel compact fan-shaped folded waveguide resonator (FSFWGR) is proposed in this article. The resonator can reduce the size by 87.5% compared with the waveguide cavity while retaining a high unloaded quality ( $Q_{{u}})$ value (>1500). Due to these advantages, a high-quality (-Q) fourth-order bandpass filter centered at 6.0 GHz with a 6.67% bandwidth is designed first. Novel slots cross-coupling is presented to produce a pair of transmission zeros (TZs) while avoiding support structures that traditional metal probes often used. Then compact deformed folded antenna is designed to integrate with the filter. Compared with published works where the last resonator of the reference filter often needs to be detuned, all parameters here do not need to change by introducing a metal probe between the feeding point of the filter and antenna. Therefore, the integrating process is more straightforward. The integrated filter/antenna achieves the same filtering function as the reference filter and a high gain of 6.7 dBi. For experimental demonstration, metal 3-D printing technology was used to fabricate models with only two pieces that significantly reduced the assembly loss. The final measured results agreed well with simulated ones.

In this paper, we investigate a nonlinear generalized fractional differential equation with two-point and integral boundary conditions in the frame of $ \kappa $-Hilfer fractional derivative. The existence and uniqueness results are obtained using Krasnoselskii and Banach's fixed point theorems. We analyze different types of stability results of the proposed problem by using some mathematical methodologies. At the end of the paper, we present a numerical example to demonstrate and validate our findings.

Phase unwrapping is one of the key problems in interferometric fiber sensors, which usually acts as the system performance bottleneck. Compared with the two-dimensional phase unwrapping, the one-dimensional phase unwrapping suffers more seriously from noise. Because modern phase unwrapping algorithms need to make the best use of all the adjacent phase points when evaluating the true phase at a given point. But the available adjacent points in one-dimensional phase unwrapping are very limited. A two-step one-dimensional phase unwrapping algorithm is proposed in this work to combat the above problems. In the first step, the one-dimensional phase is encoded into two-dimensional array based on the quasi-Gramian matrix, and in the second step the deep convolutional neural network (DCNN) is adopted for phase unwrapping. Both simulation and actual experiment results show that the unwrapped phase quality by using our method obviously outperforms the traditional methods with the signal-to-noise ratio (SNR) of lower than 4 dB, and it can still work stably even for negative SNR.

This letter deals with the joint design of transmit waveform and receive filter bank for airborne multiple-input multiple-output (MIMO) radar under the target uncertainties. Assuming that the spatial angle and the Doppler frequency of the target are unknown, we formulate the maximin joint design problem by maximizing the worst-case signal-to-interference-plus-noise ratio (SINR) under the energy constraint, flexible modulus constraint, and similarity constraint on the \textcolor{red} {transmit} waveform. To tackle this problem, we develop a computationally efficient algorithm based on iterative feasible point pursuit successive convex approximation (FPP-SCA). Numerical results are provided to demonstrate the effectiveness and robustness of the proposed algorithm.

Deep learning in 3-D classification tasks focuses on the designs of comprehensive local aggregation operators. Airborne laser scanning (ALS) point clouds have their own characteristics: 1) the object overlaps in the vertical direction; 2) the spatial density is uneven; and 3) the objects present large-scale variations between different categories. However, the dynamic graph convolutional neural network (DGCNN) ignores the inherent properties of ALS point clouds. This study modifies the DGCNN network and proposes a new neural network module called Webconv, which densely connects point pairs in multiscale local regions to learn contextual information. One modified cross entropy loss function with variable weight is proposed to solve the problem of uneven category distributions in ALS points. Because preprocessing such as block partition ignores the context information of the whole region, the conditional random field is used to refine the point cloud. Our approach achieves state-of-the-art performance on the dataset of the 2019 IEEE GRSS Data Fusion Contest 3-D Point Cloud Classification Challenge and can be widely used for ALS point classification.