Search

With the continued proliferation of mobile Internet and geo-locating technologies, carpooling as a green transport mode is widely accepted and becoming tremendously popular worldwide. In this paper, we focus on a popular carpooling service called ride hitching, which is typically implemented using a publish/subscribe approach. In a ride hitching service, drivers subscribe ride orders published by riders and continuously receive matching ride orders until one is picked. The current systems (e.g., Didi Hitch) adopt a threshold-based approach to filter out ride orders. That is, a new ride order will be sent to all subscribing drivers whose planned trips can match the ride order within a pre-defined detour threshold. A limitation of this approach is that it is difficult for drivers to specify a reasonable detour threshold in practice. In addressing this problem, we propose a novel top-k subscription query called Top-k Ride Subscription (TkRS) query, which continuously returns the best k ride orders that match drivers' trip plans to them. We propose two efficient algorithms to enable the top-k results maintenance. We also design a novel hybrid index and a two-level buffer to efficiently track the top- $k$ results. Finally, extensive experiments on real-life datasets suggest that our algorithms can achieve desirable performance in practical settings.

This paper concerns the adequacy of utilising a chimney drain, horizontal filter and cutoff wall in reducing seepage through an earthen dam. To this end, anti-seepage systems comprising a chimney drain, a horizontal filter and a cutoff wall were investigated. Variations of lengths, angles and locations of these features were studied. The amount of seepage was determined using SEEP/W software. The results indicated that by expanding the vertical chimney drain and the horizontal filter from the embankment toe to half or three-quarters of the downstream slope, the seepage flow from the dam body and its foundation will decrease. On the contrary, when the horizontal filter's length is extended to more than three-quarters of the dam's downstream slope, the amount of seepage flow that may pass through the dam's cross-section rapidly increases. Additionally, it was observed that seepage from the earthen dam and its foundation reduced by increasing the cutoff depth. Finally, with respect to the minimum factor of safety for the downstream slope, obtained using SLOPE/W software, a suitable model for seepage control within the body and foundation of the dam was found to be the hybrid model introduced at the end of this study.

A 50-year-old man with atrial fibrillation without anticoagulation presented to the hospital after tonic–clonic seizures, followed by pulseless cardiopulmonary arrest with return of spontaneous circulation after 2 min. Imaging demonstrated a left superior cerebellar thromboembolic ischaemic

Dynamic spectrum sharing allows the 4G-LTE and 5G-NR to use the same frequency spectrum band for communication and the spectrum allocator allocates the spectrum to the users based on their demands. In this paper we have analyzed existing OFDM (orthogonal Frequency division multiplexing) Technique in the physical layer for the spectrum sharing and proposed a FBMC (Filter Bank Multi Carrier) technique with Long duration Pulse Shaping filter and designed a Sequence Based Spectrum Allocator for designing a high spectral efficient network. Results were validated based on the parameters like Secondary User Score and Primary User Satisfaction Ratio and Secondary user throughput for overlay spectrum sharing mode in cognitive radio.

Sequential Bayesian filters in non-linear dynamic systems require the recursive estimation of the predictive and posterior distributions. This paper introduces a Bayesian filter called the adaptive kernel Kalman filter (AKKF). With this filter, the arbitrary predictive and posterior distributions of hidden states are approximated using the empirical kernel mean embeddings (KMEs) in reproducing kernel Hilbert spaces (RKHSs). In parallel with the KMEs, some particles, in the data space, are used to capture the properties of the dynamical system model. Specifically, particles are generated and updated in the data space, while the corresponding kernel weight mean vector and covariance matrix associated with the feature mappings of the particles are predicted and updated in the RKHSs based on the kernel Kalman rule (KKR). Simulation results are presented to confirm the improved performance of our approach with significantly reduced particle numbers, by comparing with the unscented Kalman filter (UKF), particle filter (PF) and Gaussian particle filter (GPF). For example, compared with the GPF, the proposed approach provides around 5% logarithmic mean square error (LMSE) tracking performance improvement in the bearing-only tracking (BOT) system when using 50 particles., Comment: The manuscript has been accepted for publication as a regular paper in the IEEE Transactions on Signal Processing

Mission profile emulation for submodules (SMs) in the modular multilevel converter (MMC) is a crucial step before a full-scale system is put into field operation. Existing methods suffer from limitations of the testing accuracy due to the staircase-like arm voltages which result in distorted loading current and require large filter inductance. This is especially the case when limited number of SMs are under test. In this paper, an efficient MPE method is proposed to emulate the loading behaviors of multiple SMs in MMC. The testing circuit allows two arms of MMC with flexible SM number, operating in inverting and rectifying mode under nearest level modulation (NLM), to be tested simultaneously with reduced DC supply voltage. A common H-bridge circuit is introduced in the testing circuit, with each arm of the H-bridge circuit functioning as the voltage compensator by using feed-forward control. In this way, the current distortion can be significantly suppressed and testing accuracy can be guaranteed in a cost-efficient manner even though no large inductor is included in the testing circuit. The proposed method allows SMs under test to be loaded as in the actual back-to-back MMC system. Simulations and experimental measurements are given to validate the proposed method.

The multipoint dynamic aggregation (MPDA) problem of the multirobot system is of great significance for its real-world applications such as bush fire elimination. The problem is to design the optimal plan for a set of heterogeneous robots to complete some geographically distributed tasks collaboratively. In this article, we consider the dynamic version of the problem, where new tasks keep appearing after the robots are dispatched from the depot. The dynamic MPDA problem is a complicated optimization problem due to several characteristics, such as the collaboration of robots, the accumulative task demand, the relationships among robots and tasks, and the unpredictable task arrivals. In this article, a new model of the problem considering these characteristics is proposed. To solve the problem, we develop a new genetic programming hyperheuristic (GPHH) method to evolve reactive coordination strategies (RCSs), which can guide the robots to make decisions in real time. The proposed GPHH method contains a newly designed effective RCS heuristic template to generate the execution plan for the robots according to a GP tree. A new terminal set of features related to both robots and tasks and a cluster filter that assigns the robots to urgent tasks are designed. The experimental results show that the proposed GPHH significantly outperformed the state-of-the-art methods. Through further analysis, useful insights such as how to distribute and coordinate robots to execute different types of tasks are discovered.

This article aims at proposing an adaptive neural control strategy for a class of nonlinear time-delay systems with input delays and unknown control directions. Different from previous researches that investigated delays and constraints separately, the novelty of this article lies in that it simultaneously considers delays (state and input delays) and input constraints (magnitude and rate constraints) for a class of uncertain nonlinear systems. In this article, the uncertain states and input delays are handled by integrating a constructed auxiliary system that functions as an observer with neural networks (NNs), with which the adverse effects caused by the uncertain states and input delays can be approximated and compensated. By involving smooth hyperbolic tangent functions in the designed auxiliary system, the problem of magnitude and rate constraints of the control input is fully addressed. Then, the backstepping technique runs through the entire control designing process, which allows the designed adaptive neural control strategy to handle the input constraints and delays at the same time. Furthermore, Nussbaum functions are employed to resolve the problem of unknown control directions. Due to the introduction of an input-driven filter, only the output of the system is required to be measured as the control feedback, which promotes the applicability of the designed controller. Under the proposed control scheme, semiglobal, uniform, and ultimate boundedness of all signals of the closed-loop system is realized with uncertain control directions, input and state delays, and guaranteed magnitude and rate constraints of control inputs. Finally, simulation results are illustrated to verify the effectiveness of the presented control method.

High frequency power electronic devices in grid-connected converters excited vast switching harmonics and electromagnetic interference (EMI) noises. Conventional LCL-type harmonic filters can well attenuate switching harmonics, but have a poor EMI suppression capability, especially for common-mode (CM) noise. Symmetrical filter structure and an additional EMI filter are generally needed to effectively suppress EMI in power converters, but this will inevitably increase the size and cost of the output filter. In order to tackle this issue without sacrificing filter performance, this paper proposes a decoupled magnetic integration of symmetrical LCL filter with a CM inductor for single-phase grid-connected converters. The harmonic inductors of symmetrical LCL filter and CM inductor are integrated on the side-limbs and middle-limb of an EE-type magnetic core, respectively. Due to the decoupled magnetic structure and the reasonably assignment of the windings, the magnetic coupling between converter-side inductors and grid-side inductors can be greatly canceled to improve the performance of LCL filters on basis of the integration of CM inductor to suppress CM EMI. The detailed design guidance of the proposed integration method is presented. A single-phase decoupled magnetic integration filter is fabricated and tested, and the experimental results verify the effectiveness and feasibility of the proposed method.

This work studies how to estimate the mean-field density of large-scale systems in a distributed manner. Such problems are motivated by the recent swarm control technique that uses mean-field approximations to represent the collective effect of the swarm, wherein the mean-field density (especially its gradient) is usually used in feedback control design. In the first part, we formulate the density estimation problem as a filtering problem of the associated mean-field partial differential equation (PDE), for which we employ kernel density estimation (KDE) to construct noisy observations and use filtering theory of PDE systems to design an optimal (centralized) density filter. It turns out that the covariance operator of observation noise depends on the unknown density. Hence, we use approximations for the covariance operator to obtain a suboptimal density filter, and prove that both the density estimates and their gradient are convergent and remain close to the optimal one using the notion of input-to-state stability (ISS). In the second part, we continue to study how to decentralize the density filter such that each agent can estimate the mean-field density based on only its own position and local information exchange with neighbors. We prove that the local density filter is also convergent and remains close to the centralized one in the sense of ISS. Simulation results suggest that the centralized suboptimal density filter is able to generate convergent density estimates, and the local density filter is able to converge and remain close to the centralized filter., arXiv admin note: text overlap with arXiv:2009.05366

With the aid of neural networks, this article develops two data-driven designs of fault detection (FD) for dynamic systems. The first neural network is constructed for generating residual signals in the so-called finite impulse response (FIR) filter-based form, and the second one is designed for recursively generating residual signals. By theoretical analysis, we show that two proposed neural networks via self-organizing learning can find their optimal architectures, respectively, corresponding to FIR filter and recursive observer for FD purposes. Additional contributions of this study lie in that we establish bridges that link model- and neural-network-based methods for detecting faults in dynamic systems. An experiment on a three-tank system is adopted to illustrate the effectiveness of two proposed neural network-aided FD algorithms.

In this paper, an event-triggered output feedback adaptive fuzzy control scheme is developed for a class of uncertain discrete-time multi-input multi-output (MIMO) nonlinear systems with immeasurable states and unknown control gains. Due to the existence of unmeasured states, a set of fuzzy filters are designed, then a filter-based event-triggered adaptive fuzzy control scheme is developed for discrete-time MIMO nonlinear systems. To solve the problem of state estimation caused by the coupling of control input and unknown control gains, a fuzzy filters-based state observer is developed by combining filter states. And then, a parameterized state observe is constructed to effectively estimate immeasurable state signals by the combination of the fuzzy filter states and the gradient-based fuzzy parameter updating law. Based on filter states and estimation states, an event-based adaptive fuzzy control scheme is proposed by novel intermediate errors and backstepping. To overcome the causal problem caused by the low-triangular structure, the variable substitution and a predictor are, respectively, employed to forecast the future state and reference signals. A series of stability analyses illustrates that the proposed scheme achieves immeasurable state estimations, guarantees the ultimately uniformly boundedness of the closed-loop system, and obtains the good tracking performance, while reducing communication occupancy. Finally, simulation studies on a numerical and a practical example are conducted to demonstrate the effectiveness of the proposed scheme.

The high-frequency (HF) signal injection methods have been widely used for the low-speed sensorless control of permanent magnet motor under vector control framework. However, for the fault tolerant permanent magnet synchronous motor (FTPMSM) system under fault condition, the fault tolerant control is usually based on independent control of each phase current, rather than the vector control, which prevents the existing sensorless control from being directly applied. To cope with this issue, this paper proposes a novel sensorless control by HF signal injection in any non-fault two phase windings for the FTPMSM system with no coordinate transformation, which can ensure the low- and zero-speed sensorless control performance both in fault-free and fault tolerant operation conditions. This sensorless control scheme is based on independent control framework of each phase current, which consists of the real-time inductance observer, the non-orthogonal phase-locked loop (NPLL), the notch filter and the polarity discrimination of magnet pole. The real-time inductance observer is proposed to estimate the differential mode inductances of two non-fault phase. The NPLL is proposed to achieve the rotor position and speed estimation based on the obtained two phase inductances. To enhance the sensorless control performance, a new notch filter is proposed to filter out the injected HF response component in phase current. The proposed sensorless FTPMSM system has preeminent low-speed control performance and robustness in non-fault and fault tolerant operation process, which is also verified via experiments on a six-phase FTPMSM platform.

Fundamental harmonic analysis (FHA) model is widely used for wireless power transfer (WPT) due to its physical insight and compact form (for example, 8RL/2) of diode rectifier equivalent resistance. Because this compact form is valid for continuous conduction mode (CCM), and resonant tank parameters such as inductance and capacitance are independent of equivalent resistance when designing these parameters based on double side LCC (DS-LCC) FHA model, the selected optimal parameters may not satisfy the CCM prerequisite. This will make actual system characteristics deviate from the FHA model predictions, especially the output voltage and power. Therefore, the original optimal design is invalid and designer needs to execute a circular design, namely designing resonant tank parameters based on FHA model, then judging or simulating whether satisfy CCM operation, redesigning parameters if not satisfy CCM. In order to validate the original optimal design even if CCM isn`t established and make design process straightforward, an effective remedy which is adding a series LC filter in secondary side is proposed, and a computation method about the required minimum filter parameter is provided. Effectiveness of proposed remedy is verified by simulation and experiment. Besides, computation method is packaged as a program to help researchers design rapidly.

One of the main application scenario that the upcoming 5G systems are expected to support is Internet of Things (IoT) traffic backhauling. With the exponential increase of devices, it follows the critical challenge of handling the huge data volume that they produce. Terrestrial Networks may not be sufficient for that, or not always available. Satellites then come into play, and they can provide the ubiquitous coverage targeted by 5G. On the other side, aggregation can be an efficient strategy to reduce the amount of traffic, and use network resources in an efficient manner. Building on that, in this work we consider an hybrid IoT-satellite network architecture for collecting MQTT traffic. We design an advanced filter namely MQTT-MFA that performs MQTT topics aggregation in close proximity of MQTT Publishers before data is sent over the satellite link. The developed MQTT-MFA implementation leverages on MQTT bridge, and it is compliant with Mosquitto, an open source MQTT broker. Extensive simulation results have demonstrated that MQTT-MFA requires only 10% of the TCP/IP overhead generated in a baseline architecture (where the bridge is not set up) and 1/4 number of bytes to transmit the same amount of information, when compression of the data payload is also enabled.

This paper proposes an adaptive self-tuned current controller design for an LCL-filtered LC-tuned single-phase shunt hybrid active power filter (HAPF), which offers adaptability and flexibility of the control system to filter parameters changes. System modelling and online identification of parameters are presented. Then, a systematic self-tuned controller (STC) design is proposed to calculate the controller parameters and obtain the final control law adaptively. The exact HAPF parameters design is described and the overall closed-loop stability evaluation with considering digital control delay is discussed in this paper. To confirm the theoretical achievements, simulations and experimental tests on a real prototype system are performed. The results are in good accordance and demonstrate the effectiveness and superior performance of the suggested control technique. Furthermore, a comprehensive comparison between the proposed control technique and other controllers is provided to show the superiority of the proposed system from different aspects.

Object tracking is a fundamental task for the visual perception system on the intelligent unmanned aerial vehicle (UAV). The high efficiency of correlation filter (CF) based trackers has advanced the widespread development of online UAV object tracking. This kind of method can effectively train a filter to discriminate the target from the background. However, most CF-based methods require a fixed label function over all the previous samples, leading to over-fitting and filter degradation, especially in complex drone scenarios. To address this problem, a novel adaptive response reasoning approach is proposed for CF learning. It can leverage temporal information in filter training and significantly promote the robustness of the tracker. Specifically, the proposed response reasoning method goes beyond the standard response consistency requirement and constructs an auxiliary label of the current sample. Besides, it helps learn a generic relationship between the previous and current filters, thereby realizing self-regulated filter updating and enhancing the discriminability of the filter. Extensive experiments on four well-known challenging UAV tracking benchmarks with 278 videos sequences show that the presented method yields superior results to 40 state-of-the-art trackers with real-time performance on a single CPU, which is suitable for UAV online tracking missions.

Different from typical circuit breakers, breakers used to switch AC filters in convertor stations have to operate frequently according to the real-time operation mode and daily load changes. Therefore, the performance of filter breakers decreases very quickly, which leads to frequent accidents. Both preinsertion resistors and controlled switching devices can suppress switching transients. In this paper, a comparative analysis of whether a controlled switching device and preinsertion resistor should be adopted is carried out using PSCAD. The transient is calculated as the combined result of the switching angle and preinsertion resistor value, and the optimal configurations are proposed. After determining the resistor value and whether a controlled switching device should be included, the reference voltage of the arrester should be reevaluated.

Merging at highway on-ramps while interacting with other human-driven vehicles is challenging for autonomous vehicles (AVs). An efficient route to this challenge requires exploring and exploiting knowledge of the interaction process from demonstrations by humans. However, it is unclear what information (or environmental states) is utilized by the human driver to guide their behavior throughout the whole merging process. This paper provides quantitative analysis and evaluation of the merging behavior at highway on-ramps with congested traffic in a volume of time and space. Two types of social interaction scenarios are considered based on the social preferences of surrounding vehicles: courteous and rude. The significant levels of environmental states for characterizing the interactive merging process are empirically analyzed based on the real-world INTERACTION dataset. Experimental results reveal two fundamental mechanisms in the merging process: 1) Human drivers select different states to make sequential decisions at different moments of task execution, and 2) the social preference of surrounding vehicles can impact variable selection for making decisions. It implies that efficient decision-making design should filter out irrelevant information while considering social preference to achieve comparable human-level performance. These essential findings shed light on developing new decision-making approaches for AVs., Comment: 12 pages, 8 figures

The three-phase dual active bridge (3p-DAB) is a popular topology suitable for high power DC-DC isolated conversion. This converter features bi-directional power flow, zero-voltage switching capability, and reduced transformer and filter volume. These features meet the requirements of several applications, such as solid-state transformers. But when operating under light load the resulting converter efficiency is highly degraded by the loss of soft-switching capability. This paper introduces the use of a Wye-extended-Delta (Y-e) transformer winding in the 3p-DAB converter, which increases the efficiency under light load, keeping the simplicity of phase-shift control. The framework developed allows a complete analysis of the converter, including the determination of the power and rms current, zero-voltage switching boundaries and the impact of the selected phase-shifting angle in the transformer design. Additionally, this paper presents a negative sequence modulation scheme, allowing maximum efficiency under light loads to be achieved for both direct and reverse power flows. The performance of the 3p-DAB with Y-e transformers is compared with previously reported winding options for 3p-DAB. The experimental results show a converter efficiency increase, up to 10\% for direct power flow and up to 9\% for reverse power flow, when operating at light load (10\% of nominal power).

The air ground integrated networks can leverage unmanned aerial vehicle (UAV) communications to tackle the ever-increasing and unbalanced traffic load in future communication systems. This paper investigates the UAV enabled traffic offloading problem in air ground integrated networks with mixed user traffic. The problem jointly maximizes the system load balance and the total UAV reward, which can be formulated under a two-layer network graph model. In the cellular network graph, the association between the delay-sensitive users and the access points (APs) as well as the association between the UAVs and the APs are formulated. In the UAV network graph, the association between the delay-insensitive users and the UAVs is formulated. By observing the coupling relationship of the decision variables, we decouple the problem into three sub-problems and solve the first two sub-problems with reduced complexity. Then, we devise a Deep Neural Network (DNN) empowered genetic algorithm to solve the last sub-problem. The DNN can be leveraged to filter out the non-optimal solutions in the initialization operator of the genetic algorithm for improving the efficiency. Performance comparisons are provided between the proposed traffic offloading scheme and the existing ones, which validate the advantages of the DNN empowered genetic algorithm regarding its convergence, accuracy, and robustness.

During the past four decades the use of nonwoven geotextiles as filter materials in geotechnical works has experienced a marked increase. In this study, the water permeability coefficient of nonwoven geotextiles after 23 years of exploitation in drainage system was determined to evaluate the influence of physical clogging on filtration performance of geotextiles. The obtained results showed that the geotextiles were clogged by soil particles. What is more, the objective in this study was to establish a gradient ratio test method for estimating a clogging phenomenon of geotextile filter adequately. A series of gradient ratio tests on three kinds of needle-punched nonwoven geotextile filters whose number of constrictions differs was performed. A geotextile filter was put under a silty sand layer. Based on the test results, it was confirmed that the water permeability coefficient became small as the gradient ratio increased, depending on the number of constrictions of the geotextile. It was also found that the modified gradient ratio GR4 is more sensitive than GR8 and GR25. Based on interpretation of the experimental results, the authors recommend the acceptance limit of GR8 equal to 3.7 and of GR4 equal to 4.8.

Modern industrial plants generally consist of multiple manufacturing units, and the local correlation within each unit can be used to effectively alleviate the effect of spurious correlation and meticulously reflect the operation status of the process system. Therefore, the local correlation, which is called spatial information here, should also be taken into consideration when developing the monitoring model. In this study, a cascaded monitoring network (MoniNet) method is proposed to develop the monitoring model with concurrent analytics of temporal and spatial information. By implementing convolutional operation to each variable, the temporal information that reveals dynamic correlation of process data and spatial information that reflects local characteristics within individual operation unit can be extracted simultaneously. For each convolutional feature, a submodel is developed and then all the submodels are integrated to generate a final monitoring model. Based on the developed model, the operation status of the newly collected sample can be identified by comparing the calculated statistics with their corresponding control limits. Similar to the convolutional neural network (CNN), the MoniNet can also expand its receptive field and capture deeper information by adding more convolutional layers. Besides, the filter selection and submodel development in MoniNet can be replaced to generalize the proposed network to many existing monitoring strategies. The performance of the proposed method is validated using two real industrial processes. The illustration results show that the proposed method can effectively detect process anomalies by concurrent analytics of temporal and spatial information.

Federated learning enables collaborative deep learning over multiple clients without sharing their local data, and it becomes increasingly popular due to the good balance between data privacy and model usability. Generally, it faces the heavy communication overhead when a large number of clients are involved and the low convergence rate incurred by non-IID data. However, few existing solutions cannot simultaneously address the communication and statistic challenges. In this paper, we propose a computation and communication efficient federated learning via adaptive sampling. By capturing different data distribution among clients, we utilize the concept of self-paced learning to adaptively adjust thresholds to filter training data for each client and also to select suitable clients to be involved in each global learning round. We prove its correctness through theoretical analysis and also evaluate its performance through experimental evaluations on real-world datasets. Detailed experimental results show that it can effectively reduce the communication cost while achieving the good trade-off between accuracy and efficiency.

Based on the interval type-2 fuzzy (IT2F) approach, this paper investigates the fault detection filter design problem for a class of nonhomogeneous higher-level Markov jump systems with uncertain transition probabilities. Considering that the mode information of the system cannot be obtained synchronously by the filter, the hidden Markov model can be seen as a detector to handle this asynchronous problem, and the parameter uncertainty can be processed by the IT2F approach with the lower and upper membership functions. Then, the asynchronous IT2F filter is designed to deal with the fault detection problem. Furthermore, the Gaussian transition probability density function is introduced to describe the uncertainty transition probabilities of the system and the filter. Based on Lyapunov theory, the existence of the designed asynchronous IT2F filter and the dissipativity of the filter error system can be well ensured. The simulation study on a quarter-car suspension system verifies that the designed asynchronous IT2F filter can detect faults without error alarms.

The design of electromagnetic interference (EMI) filter with smaller size and lightweight for a high efficiency and high power density 100 kW SiC five-level T-type (5LT2) photovoltaic (PV) inverter becomes more challenging due to the increased EMI noise resulted by high switching frequency, fast switching transients, and the removal of LCL filter. This paper proposes an unbalanced Wheatstone bridge and a two-stage notch filter to attenuate the increased common mode (CM) and differential mode (DM) noise respectively. The conventional Wheatstone bridge will suffer size and weight penalty for 100 kW PV converter due to the high parasitic capacitance of the 100 kW PV panel. An unbalanced Wheatstone bridge with a one-stage LC filter is therefore developed to address this issue and reduce the total CM inductance. The proposed two-stage notch filter is designed to achieve high attenuation, solve its parameter variation issue and help reduce the size and weight of DM filter. The total weight of the designed EMI filter for this 100 kW PV inverter is 3.54 kg. The measured conducted EMI spectrum of the 100 kW SiC PV inverter with and without proposed EMI filter are provided to validate the effectiveness of the EMI filter design.

Accurate and real-time prediction of network traffic can not only help system operators allocate resources rationally according to their actual business needs but also help them assess the performance of a network and analyze its health status. In recent years, neural networks have been proved suitable to predict time series data, represented by the model of a long short-term memory (LSTM) neural network and a temporal convolutional network (TCN). This article proposes a novel hybrid prediction method named SG and TCN-based LSTM (ST-LSTM) for such network traffic prediction, which synergistically combines the power of the Savitzky-Golay (SG) filter, the TCN, as well as the LSTM. ST-LSTM employs a three-phase end-to-end methodology serving time series prediction. It first eliminates noise in raw data using the SG filter, then extracts short-term features from sequences applying the TCN, and then captures the long-term dependence in the data exploiting the LSTM. Experimental results over real-world datasets demonstrate that the proposed ST-LSTM outperforms state-of-the-art algorithms in terms of prediction accuracy.

In this paper, a novel zero-voltage switching (ZVS) full-bridge converter is proposed to overcome the disadvantages of traditional phase-shifted full-bridge (PSFB) converter. In the proposed converter, a half-bridge (HB) converter with center-tap rectifier (CTR) is integrated into the traditional PSFB converter by sharing the lagging-leg switches and output filter. The HB converter operates with full duty cycle to ensure continuous energy transmission and the output voltage is regulated by adjusting the phase-shifted time of FB converter. The proposed converter has many advantages such as reduced filter requirement, wide ZVS range and low primary circulating current. In addition, the rectifier in the proposed converter is composed of a full-bridge rectifier (FBR) section and a CTR section. This feature makes the proposed converter well appropriate for the medium output voltage applications. Key operation principle and characteristics are fully discussed in this paper. The effectiveness of proposed converter is verified using a 1.2-kW prototype circuit with 280 V input voltage and 60-120 V output voltage.

Fast scale stability of Dual active bridge (DAB) converter with input filter under constant power load (CPL) is analyzed in this paper. Since the commonly used reduced-order average model is not effective to accurately describe the dynamic behavior of DAB converter, a discrete-time model is established. The bifurcation points of the system are predicted by using bifurcation diagrams. In addition, Poincar map theory and the Jacobian matrix of the system are used for bifurcation analysis. The eigenvalue locis of Jacobian matrix are carried. Serval Hopf bifurcations and a saddle-node bifurcation, which are predicted by the eigenvalue locis, have been found in the converter. The influence of system parameters on the stability of the system is explained in detail. It is found that high output voltage bandwidth, input filter and light-load are the main causes of instability in this system. To stabilize the converter, a proper feedback control is presented. In addition, stable boundaries are given to help to avoid instability. All the findings have been verified by simulation and experimental results.

A vast amount of activation values of DNNs are zeros due to ReLU (Rectified Linear Unit), which is one of the most common activation functions used in modern neural networks. Since ReLU outputs zero for all negative inputs, the inputs to ReLU do not need to be determined exactly as long as they are negative. However, many accelerators usually do not consider such aspects of DNNs, losing a huge amount of opportunities for speedups and energy savings. To exploit such opportunities, we propose early negative detection (END), a computation pruning technique that detects the negative results at an early stage. The key to the early negative detection is the adoption of inverted twos complement representation for filter parameters. This ensures that as soon as the intermediate results become negative, the final results are guaranteed to be negative. Upon detection, the remaining computation can be skipped and the following ReLU output can be simply set to zero. We also propose a DNN accelerator architecture (ComPreEND) that takes advantage of such skipping. ComPreEND with END significantly improves both the energy efficiency and the performance according to the evaluation.

We consider a robust state space filtering problem in the case that the transition probability density is unknown and possibly degenerate. The resulting robust filter has a Kalman-like structure and solves a minimax game: the nature selects the least favorable model in a prescribed ambiguity set which also contains non-Gaussian probability densities, while the other player designs the optimum filter for the least favorable model. It turns out that the resulting robust filter is characterized by a Riccati-like iteration evolving on the cone of the positive semidefinite matrices. Moreover, we study the convergence of such iteration in the case that the nominal model is with constant parameters on the basis of the contraction analysis in the same spirit of Bougerol. Finally, some numerical examples show that the proposed filter outperforms the standard Kalman filter.

In recent years, the number of traffic accident deaths due to distracted driving has been increasing dramatically. Fortunately, distracted driving can be detected by the rapidly developing deep learning technology. Nevertheless, considering that real-time detection is necessary, three contradictory requirements for an optimized network must be addressed: a small number of parameters, high accuracy, and high speed. We propose a new D-HCNN model based on a decreasing filter size with only 0.76M parameters, a much smaller number of parameters than that used by models in many other studies. D-HCNN uses HOG feature images, L2 weight regularization, dropout and batch normalization to improve the performance. We discuss the advantages and principles of D-HCNN in detail and conduct experimental evaluations on two public datasets, AUC Distracted Driver (AUCD2) and State Farm Distracted Driver Detection (SFD3). The accuracy on AUCD2 and SFD3 is 95.59% and 99.87%, respectively, higher than the accuracy achieved by many other state-of-the-art methods.

As a sort of model-free tracking approach, discriminative correlation filter (DCF)-based trackers have shown prominent performance in unmanned aerial vehicle (UAV) tracking. Nevertheless, typical DCFs acquire all samples oriented to filter training merely from the current frame by cyclic shift operation in the spatial domain but ignore the consistency between samples across the timeline. The lack of temporal cues restricts the performance of DCFs under object appearance variations arising from object/UAV motion, scale variations, and viewpoint changes. Besides, many existing methods commonly neglect the channel discrepancy in object position estimation and generally treat all channels equally, thus limiting the further promotion of the tracking discriminability. This work proposes a novel tracking approach based on a multi-regularized correlation filter, i.e., MRCF tracker. By regularizing the deviation of responses and the reliability of channels, the tracker enables smooth response variations and adaptive channel weight distributions simultaneously, leading to favorable adaption to object appearance variations and enhancement of discriminability. Exhaustive experiments on five authoritative UAV-specific benchmarks validate the competitiveness and efficiency of MRCF against top-ranked trackers. Furthermore, we apply our proposed tracker to monocular UAV self-localization under air-ground robot coordination. Evaluations indicate the practicability of the presented method in UAV localization applications.

This work presents three single-phase isolated ac-ac converters based on the dual active bridge (DAB) converter and interleaved transformers. The structures allow obtaining filter elements with reduced dimensions, as well as bidirectional power flow and high efficiency over a wide load range. Multilevel voltage waveforms exist in the input and output sides of the symmetrical converters, resulting in reduced harmonic content. The possible arrangements include a symmetrical half-bridge converter (SHBC), a symmetrical full-bridge converter (SFBC), and a symmetrical interleaved full-bridge converter (SIFBC). The introduced topologies can be regarded as prominent solutions for applications involving solid-state transformers (SSTs), dynamic voltage restorers (DVRs), and uninterruptible power supplies (UPSs). A thorough mathematical analysis of the SHBC is carried out and experimental results are presented to investigate the converter performance in steady-state and transient conditions.

This study investigates the long-term instability problem of periodic feedback loop-based control schemes such as repetitive control (RC), periodic adaptive disturbance observer (PADOB), and iterative learning control (ILC). The long-term instability problem is one wherein a closed-loop system becomes unstable because a tracking error diverges with the increase in the number of iterations. A novel robust periodic adaptive disturbance observer is proposed in this paper to guarantee the long-term stability of the closed-loop system. The overall system eventually becomes stable because the proposed control scheme using an IIR filter with an optimized lead time can stabilize the error propagation in the iteration domain. Compared to other methods such as zero-phase and IIR filters, the proposed control scheme is very practical because it allows selecting the cut-off frequency of the filter intuitively to stabilize the control system. The optimized lead time in the IIR filter is calculated considering the phase response of the IIR filter; it can improve adaptation speed using a high adaptation gain of the control system. Various simulation and experimental tests based on high-load motion systems that require high-precision positioning are performed to verify the theoretical analysis of the long-term instability problem and the effectiveness of the proposed control scheme.

Epipolar geometry is a fundamental constraint used in computer vision systems to estimate parameters using correspondence. The most common way to describe epipolar geometry is by means of a 3x3 matrix called the fundamental matrix, and such matrices are used to store the precise geometric information relating a pair of stereo images. Its efficient estimation substantially improves initialization in visual simultaneous localization and mapping (vSLAM), which uses correspondence-based epipolar geometry to determine the trajectory of the camera and a three-dimensional scene. Conventional robust methods for epipolar geometry estimation can become computationally inefficient and inaccurate when there are low-quality correspondences. Because semantic information can be more stable than pixel intensities/descriptors, a novel Faster Region-based Convolutional Network (R-CNN) learning-based approach called the semantic filter is proposed in this paper to address these problems. The semantic filter is first trained on different semantic patches, which are described in terms of their different outlier distributions, providing different semantic labels for image contexts. Then, the patches with low-level semantic labels are filtered out. Finally, precise and robust correspondences can be determined by matches using the high-level semantic contexts, making the correspondence-based calculation more accurate. For dynamic outdoor scenes, the results of extensive experiments show that our semantic filter can help vSLAM localize accurately and robustly on a map from different viewpoints. In a completely static scenario, our semantic filter can remove the low-quality correspondences, enabling the mobile robot to operate well.

Motor drive systems supplied by a fuel-cell/battery are especially demanding when it comes to the design of the inverter. Besides a high performance (high efficiency η and power density ρ), the inverter has to cope with the wide DC voltage variation of the fuel-cell/battery that supplies the motor drive. A promising three-phase inverter topology, denoted as Y-VSI, is presented in this paper. The Y-VSI is a modular three-phase inverter, and comprises three identical phase-modules connected to a common star “Y” point. Each phase-module is equivalent to a buck-boost DC/DC converter, which allows the AC output voltages to be higher or lower than the DC input voltage. Thereby, the Y-VSI effectively copes with the wide variation of the fuel-cell/battery voltage. Each phase-module can be operated in a similar fashion to a conventional DC/DC converter, independently of the remaining two phases. Accordingly, a straightforward and simple operation/control of the Y-VSI is possible. In addition, the Y-VSI features an integrated output filter. This allows for continuous/sinusoidal motor voltage waveforms, eliminating the need of an additional filter between the inverter and the motor. This paper details the operating principle of the Y-VSI, and comparatively evaluates two modulation strategies. In order to validate the proposed concepts, a Y-VSI hardware prototype is assembled within the context of a high-speed motor drive. In the investigated drive system, a fuel-cell supplies the Y-VSI, which in return controls a 280 krpm 1 kW electric compressor. The Y-VSI hardware prototype is compared against a state-of-the-art hardware prototype, which features two energy conversion stages. It is shown that the Y-VSI is ∆η = +2.3% more efficient and at the same time ∆ρ = +10% more power dense compared to the conventional inverter solution.

This paper presents a systematic and an efficient synthesis technique for designing single-band and dual-band waveguide filters with transmission zeros. The proposed EM-based synthesis technique directly results in a filter design with an RF performance that is in excellent agreement with the ideal filter performance, thus significantly reducing the post fine optimization effort. Furthermore, the proposed technique also reduces the EM simulation resources required for the filter synthesis. The synthesis technique is lucidly explained by designing a symmetrical 6th order waveguide filter with 4 transmission zeros, a 9th order filter with 3 asymmetric transmission zeros and a 6th order dual-band filter. To the best of author’s knowledge, this is the first filter synthesis technique, which can be adopted to design waveguide filters including transmission zeros with minimum or no post fine optimization.

We show theoretically and experimentally how a flat-top second-order response can be achieved with a self-coupled single add-drop ring resonator based on two couplers with different splitting ratios. The resulting device is a 1x1 filter, reflecting light back in the input waveguide at resonating wavelengths in the passbands, and transmitting light in the output waveguide at all other non-resonating wavelengths. Different implementations of the filter have been designed and fabricated on a micron-scale silicon photonics platform. They are based on compact Euler bends - either U-bends or L-bends - and Multi-Mode Interferometers as splitters for the ring resonators. Different finesse values have been achieved by using either 50:50 MMIs in conjunction with 85:15 MMIs or 85:15 MMIs in conjunction with 95:05 double MMIs. Unlike ordinary lowest order directional couplers, the MMIs couple most of the power in the cross-port which make them particularly suitable for the topology of the self-coupled ring, which would otherwise require a waveguide crossing. Experimental results are presented, showing good agreement with simulations. The proposed devices can find applications as wavelength-selective reflectors for relatively broad-band lasers or used as 2x2 add-drop filters when two exact replicas of the device are placed on the arms of a Mach-Zehnder interferometer., 9 pages, 17 references, 12 figures

To enable chemical speciation, monitoring networks collect particulate matter (PM) on different filter media, each subjected to one or more analytical techniques to quantify PM composition present in the atmosphere. In this work, we propose an alternate approach that uses one filter type (teflon or polytetrafluoroethylene, PTFE, commonly used for aerosol sampling) and one analytical method, Fourier transform infrared (FT-IR) spectroscopy to measure almost all of the major constituents in the aerosol. In the proposed method, measurements using the typical multi-filter, multi-analytical techniques are retained at a limited number of sites and used as calibration standards. At all remaining sites, only sampling on PTFE and analysis by FT-IR is performed. This method takes advantage of the sensitivity of the mid-IR domain to various organic and inorganic functional groups and offers a fast and inexpensive way of exploring sample composition. As a proof of concept, multiple years of samples collected within the Interagency Monitoring of PROtected Visual Environment network (IMPROVE) are explored with the aim of retaining high quality predictions for a broad range of atmospheric compounds including mass, organic (OC), elemental (EC), and total (TC) carbon, sulfate, nitrate, and crustal elements. Findings suggest that models based on only 21 sites, covering spatial and seasonal trends in atmospheric composition, are stable over a 3 year period within the IMPROVE network with acceptable prediction accuracy (R2 > 0.9, median bias less than 3 %) for most constituents. The major limitation is measuring nitrate as it is known to volatilize off of PTFE filters. Incorporating additional sites at low cost, partially replacing existing, more time- and cost-intensive techniques, or using the FT-IR data for quality control or substitute for missing data, are among the potential benefits of the one-filter, one-method approach.

Introduction The aim of this study was to evaluate mean glandular dose (MGD) and image quality in low energy imaging from contrast-enhanced spectral mammography (CESM) when using double-layer filtration. Methodology A dedicated phantom was used to quantitatively estimate the MGD and image quality. The target slab of the phantom consisted of three iodine coins having a concentration of 1.0 mgI/cm3, 2.0 mgI/cm3, 4.0 mgI/cm3, a 100% adipose equivalent coin and a 100% glandular equivalent coin. The phantom was exposed using a semiautomated function at 28 k, 30 kV and 32 kV. MGD. Contrast to noise ratio (CNR) and figure of merit (FOM) were estimated for Mo/Rh, Mo/Rh + Cu, Mo/Rh + Al and Mo/Rh + Cd combinations using three breast equivalent compositions. Results MGD was reduced up to a maximum of 1.03 mGy from 1.17 mGy for 100% adipose tissue. 1.18 mGy from 1.34 mGy for 50% glandular tissue and 1.39 mGy from 1.72 mGy for the 100% glandular phantom when using double-layer filtration. All of the above-mentioned results were obtained for the 50 mm phantom using 32 kV. CNR and FOM values were not significantly reduced with a double-layer filter when compared to a single-layer filter. Conclusion The present study concluded that Mo/Rh + Cu is the best combination to reduce the MGD significantly when compared to Mo/Rh + Al or Mo/Rh + Cd. Mo/Rh + Cu also achieved optimal image quality when compared to the Mo/Rh single filter combination. Implications of practice The use of a double-layer filter in low energy imaging of CESM results in a significant reduction in MGD without degrading the quality of the image.

Solid-state transformers (SSTs) have emerged as an interesting solution for replacing the traditional low-frequency transformers in power distribution systems owing to the capacity of performing ancillary services and providing the integration of multiple energy sources. This work proposes a four-port, single-phase, multi-directional ac-ac converter for SST applications. Considering the integration of conversion stages, it is possible to achieve higher efficiency and lower component count when compared with the typical three-stage topologies reported in the literature. The introduced structure is based on the interleaving concept associated with the well-known dual active bridge (DAB) converter. The medium-voltage (MV) side employs a multilevel neutral-point-clamped (NPC) converter. The voltages across the ac ports present five levels, whereas the operating frequency of the filter elements is twice the switching frequency. Experimental results are presented and thoroughly discussed to validate the theoretical assumptions.

Most DC power distribution architectures are organized as a typical cascaded converter system that is composed of source and load converters, and may experience a system oscillation due to the negative impedance effect of constant power loads. The load-side virtual impedance method is a possible solution to such stability problem, which however makes the dynamic response of the load converter slow, and thus increases the requirement of the filter capacitance. In this paper, an auxiliary circuit that is connected to the output port of the load converter is proposed to enhance the transient performance of the power converter system, while the system can be stabilized by the virtual impedance method. The control and parameter design of the proposed auxiliary circuit is introduced. The capacitance requirement of the system for an expected transient deviation of the output voltage is significantly reduced by using the proposed auxiliary circuit. An example design is given to verify the feasibility of the proposed method with simulation analysis and experimental results.

In this article, the fault detection (FD) filter design problem is addressed for discrete-time memristive neural networks with time delays. When constructing the system model, an event-triggered communication mechanism is investigated to reduce the communication burden and a fault weighting matrix function is adopted to improve the accuracy of the FD filter. Then, based on the Lyapunov functional theory, an augmented Lyapunov functional is constructed. By utilizing the summation inequality approach and the improved reciprocally convex combination method, an FD filter that guarantees the asymptotic stability and the prescribed H∞ performance level of the residual system is designed. Finally, numerical simulations are provided to illustrate the effectiveness of the presented results.

The paper deals with the performance checking of an optimally designed LCL-filter connected along the grid-side converter (GSC) in a doubly fed induction generator (DFIG) based wind energy conversion system (WECS) in attenuating harmonics. A repetitive approach is required for designing the parameters of the LCL-filter. The reasons being the explicitness of design parameters and design requirements of the proposed filter like IEEE-519 Std. for harmonic current attenuation, maximum voltage difference allowed in between two sides of the filter to limit switching losses, and limit of reactive power compensation. The proposed filter is compared with the conventional LCL-filter. Based on the proposed method, a minimal inductance and optimal capacitance-based filter is obtained in this paper which is validated using a 2.2 kW DFIG based WECS. Comparative stability analysis of DFIG system with and without proposed filter is performed using small signal stability analysis. The system under study is simulated using MATLAB/Simulink and OPAL-RT 4510, simultaneously. Simulation results validates the effective use of proposed filter design method into DFIG system in terms of harmonic attenuation, total harmonic distortion and reactive power compensation.

When suffering the grid voltage distortion, the phase-locked loop (PLL) should equip with the fast response filter to precisely extract the fundamental voltage information for proper grid integration. To shorten the time delay induced by filters, this paper proposes a new three-samples filter (TSF) to fast eliminate the harmonics of sampled grid voltage. Theoretically, it can filter out arbitrary individual harmonic with short time delay which could be as small as only two sampling periods. Moreover, TSF can reject a group of harmonics simultaneously within the specifically set time interval. The implementation issues of TSF as in-loop filter or prefilter for both three- and single-phase applications are demonstrated in SRF-PLL. Simulation and experimental results verified its advanced performance.