Search

Deep learning (DL)-based defect detection provides various robust automated solutions for maintaining the daily operations of electric multiple units (EMUs). However, a sophisticated detection model typically requires extensive, high-quality data. Therefore, effective detection of certain low-frequency defects can be challenging because it is impractical to collect a sufficient amount of data. To address the lack of targeted training data, a novel synthetic dataset generation method for defect detection of EMUs is proposed based on generative adversarial networks (GANs) and a 3-D CRH380A model. By using synthetic data, defect detection accuracy is significantly improved, even when the scope of real data is limited. The synthetic dataset is employed in this article to train three DL detection models (YOLO v5, Mask R-CNN, and TOOD), and comprehensive experiments are conducted to verify the proposed DefectGAN. The experimental results illustrate that the mixed dataset (synthetic and real data) can significantly improve defect detection accuracy for EMUs with limited data. With a few dozen real images, only mixing the synthetic images generated by the proposed DefectGAN into training datasets can increase the ${F}1$ score from 0 to 0.8. This work will significantly improve the accuracy of defect detection of existing trouble of moving EMU detection system (TEDS) during data scarcity, facilitate automated operation and maintenance of EMUs, and reduce overall operating costs.

The role of on-orbit computing for satellites is transitioning from being a backup measure to becoming a primary key function. However, the limited computing resources available on satellites make it difficult to deploy advanced models with large parameters. Additionally, satellite on-orbit computing requires high speed and accuracy, posing significant challenges for developing suitable models. To overcome these challenges, we propose an efficient vision transformer, RepSViT, for satellite on-orbit computing. The RepSViT introduces spiking neural networks (SNNs) with high biological plausibility, event-driven property, and low power consumption into the field of remote sensing image processing and satellite on-orbit computing for the first time and incorporates structural reparameterization. Specifically, we design a dynamic dilated spiking convolution ( $\text{D}^{2}$ SC) based on SNNs to improve the feature extraction capability and efficiency of RepSViT. We also develop a spiking guided attention module (SGAM) to make RepSViT pay more attention to object-related features with lower computational costs. Furthermore, we design an efficient coupled fine-coarse-grained block (ECFC) to enhance the model’s capability in extracting coarse and fine-grained features. To ensure effective feature extraction, inference speed, and reduced computational costs, we design a reparameterized feed-forward network (RepFFN). RepSViT achieves an inference latency of 8.33 ms and a recognition accuracy of 95% on an embedded GPU, utilizing 3.77 million parameters and consuming 0.6 GFLOPs computational costs.

In this paper, the finite-time dynamic event-triggered consensus of second-order multi-agent systems with mismatched disturbances is studied. First, a nonlinear finite-time disturbance observer is designed to estimate the mismatched and matched disturbance of each follower. Second, the nonlinear dynamic integral sliding surface is developed through the mismatched and matched disturbance estimated by the observer. Then, combined with the dynamic event trigger strategy and the designed integral sliding surface, a new distributed control protocol is proposed, ensuring the finite-time consensus of the system. An analysis of Zeno behavior has found that the event-triggered interval has a positive lower bound. Finally, the effectiveness of the proposed control algorithm is verified by numerical simulation. [ABSTRACT FROM AUTHOR]

This article studies the local exponential synchronization synthesis problem of directed Lur’e networks with coupling time-varying delay under the distributed impulsive control subject to actuator saturation. First, by utilizing proof by contradiction, impulsive comparison principle, and latest improved convex hull representation of saturation function, some delay-independent sufficient criteria for local exponential synchronization are presented in the form of bilinear matrix inequalities. Meanwhile, a novel method with less conservatism is developed to estimate the domain of attraction, which is radically different from the traditional method by means of contractive invariant set. Second, optimization problems constrained by the transformed linear matrix inequalities are established to acquire the maximum estimates of both the domain of attraction and average impulsive interval (AII), which are conveniently solved by the YALMIP toolbox in MATLAB software. Finally, a numerical simulation is rendered to demonstrate the effectiveness and advantages of the proposed theoretical results.

This study examines the leader-following consensus problem of a class of second-order nonlinear multi-agent systems, where the velocity information is supposed to be unmeasurable. Under the setting, this paper presents a novel aperiodically intermittent output feedback control protocol such that all followers reach consensus with the leader, in which a distributed state observer is built for each follower to observe the velocity state. Based on the Lyapunov stability theory and some matrix analysis techniques, a couple of sufficient conditions for the leader-following consensus of the nonlinear multi-agent system under study are obtained even though the velocity state is unavailable. Finally, the effectiveness of the theoretical results is verified by numerical simulation. [ABSTRACT FROM AUTHOR]

A new catalyst (Fe2(SO4)3-NaCl) was employed for the selective conversion of corncob hemicellulose to xylose via hydrothermal treatment. The optimal conditions included Fe2(SO4)3-NaCl treatment for 2 h at 140 °C to achieve the selective conversion of hemicellulose (92.4 wt%) and a xylose yield of 106.8 wt%, which is based on the hemicellulose content in corncobs. These conditions led to a cellulose and lignin dissolution of 7.9 and 21.2 wt%, respectively. Fe2(SO4)3was found to play an important role, rather than NaCl, in the disruption of interactions among the three major components, while the conversion of lignin and cellulose was inhibited by NaCl and Fe2(SO4)3, respectively. The co-addition of Fe2(SO4)3and NaCl promoted the selective dissolution of hemicellulose, leaving most of the cellulose and lignin unreacted in the solid residue. Additionally, Fe2(SO4)3inhibited the formation of formic and acetic acids. A second hydrothermal treatment on the hemicellulose-free residue with the same catalytic system at 140 °C for 2 h resulted in 42.2 wt% cellulose conversion and a fructose yield of 79.7 wt% based on converted cellulose. Therefore, a potential route for the comprehensive utilization of corncobs is proposed.

Detecting obstacles accurately is a vital part of realizing autonomous driving. But utilizing single sensors, for example, RGB camera or LiDAR itself, is difficult to meet the needs of autonomous driving under the condition of sparse environmental sensing data. In contrast, equipments that can collect data in large quantities are expensive and difficult to be utilized in practical applications. Hence, under the premise of ensuring low application costs, the difficult task is to obtain high-precision three-dimensional information of obstacles in real-time from sparse data. To overcome this problem, a real-time detection technique that combines LiDAR data and image information for obstacles detection is proposed in this article. It can detect vehicles close to the ART in sparse data in real-time. The sparse point cloud is de-noised initially and the ground is separated from the initial cloud data. The remained cloud, which includes surrounding vehicles, are processed by the clustering algorithm according to the structural relationship and scanning mechanism of the LiDAR. Then the processed LiDAR data is fused with the classification information generated from images based on a neural network to locate suspicious obstacles. The ratio values between the central values of vehicles and image width are obtained in the image coordinate system. Then the category information and ratio values are mapped on the processed LiDAR clusters during ART operation. The method is tested and verified in the real field. The evaluation results show that the proposed method can achieve an average accuracy of 85.5% for obstacle classification with a cost-effective sensing suit.

This paper constructs and settles a charging facility location problem with the link capacity constraint over a mixed traffic network. The reason for studying this problem is that link capacity constraint is mostly insufficient or missing in the studies of traditional user equilibrium models, thereby resulting in the ambiguous of the definition of road traffic network status. Adding capacity constraints to the road network is a compromise to enhance the reality of the traditional equilibrium model. In this paper, we provide a two-layer model for evaluating the efficiency of the charging facilities under the condition of considering the link capacity constraint. The upper level model in the proposed bi-level model is a nonlinear integer programming formulation, which aims to maximize the captured link flows of the battery electric vehicles. Moreover, the lower level model is a typical traffic equilibrium assignment model except that it contains the link capacity constraint and driving distance constraint of the electric vehicles over the mixed road network. Based on the Frank-Wolfe algorithm, a modified algorithm framework is adopted for solving the constructed problem, and finally, a numerical example is presented to verify the proposed model and solution algorithm.

This study addresses the problem of robust finite-time connectivity preserving consensus for second-order multi-agent systems (MASs) with a limited communication range. Considering that the communication ability of agents in practical applications is limited, this study introduces an innovative approach to the consensus protocol, which involves the incorporation of a potential function aimed at ensuring sustained communication connection within a MAS. In addition, a finite-time fault-tolerant consensus protocol for a second-order MAS with actuator additive faults and external disturbances is designed by combining sliding mode control (SMC) and adaptive control. Following the homogeneous theorem and the finite-time consensus theory, this study concludes that, under specific conditions, a MAS can achieve finite-time consensus. Moreover, an advanced control protocol named a super-twisting sliding mode control protocol is introduced to mitigate the undesirable chattering phenomenon in the SMC. Finally, the effectiveness and superiority of the proposed method are verified by numerical examples. • The finite-time connectivity preserving consensus control problem of second-order MAS is studied. • An adaptive SMC is designed to solve parameter uncertainty and controller failures. • STSMC effectively suppresses the chattering and improves the control performance. [ABSTRACT FROM AUTHOR]

The limited regulation strategies of the regeneration microenvironment significantly hinder bone defect repair effectiveness. One potential solution is using biomaterials capable of releasing bioactive ions and biomolecules. However, most existing biomaterials lack real‐time control features, failing to meet high regulation requirements. Herein, a new Strontium (Sr) and epigallocatechin‐3‐gallate (EGCG) based metal–phenolic network with polydopamine (PMPNs) modification is prepared. This material reinforces a biomimetic scaffold made of extracellular matrix (ECM) and hydroxyapatite nanowires (nHAW). The PMPNs@ECM/nHAW scaffold demonstrates exceptional scavenging of free radicals and reactive oxygen species (ROS), promoting HUVECs cell migration and angiogenesis, inducing stem cell osteogenic differentiation, and displaying high biocompatibility. Additionally, the PMPNs exhibit excellent photothermal properties, further enhancing the scaffold's bioactivities. In vivo studies confirm that PMPNs@ECM/nHAW with near‐infrared (NIR) stimulation significantly promotes angiogenesis and osteogenesis, effectively regulating the microenvironment and facilitating bone tissue repair. This research not only provides a biomimetic scaffold for bone regeneration but also introduces a novel strategy for designing advanced biomaterials. The combination of real‐time photothermal intervention and long‐term chemical intervention, achieved through the release of bioactive molecules/ions, represents a promising direction for future biomaterial development. Metal–phenolic network with polydopamine (PMPNs) modification endows the biomimetic scaffold made of extracellular matrix (ECM) and hydroxyapatite nanowires (nHAW) with the ability to scavenge free radical, reactive oxygen species (ROS), and photothermal properties. Furthermore, the enhanced release of Sr ions from PMPNs@ECM/nHAW scaffolds due to the responsiveness of PMPNs to ROS provides the biomimetic scaffold with much enhanced angiogenic and osteogenic properties.

In this paper, the chain structure of magnetorheological fluid (MRF) magnetic particles was studied and analyzed, the mechanical model of MRF with different diameter ferromagnetic particles was established, silicone oil-based MRF with different particle volume fractions was prepared, the shear properties of the MRF were tested, and the theoretical and experimental data were compared. The experimental results show that the shear stress is stable with the increase of shear strain rate under the action of the magnetic field, and it has a shear thinning effect. The shear stress increases linearly with the increase of particle volume fraction. The shear stress increases with the increase of magnetic induction intensity. After data analysis and in the case of control variables, the average error of improved theoretical data and experimental data is lower than that of previous theoretical data and experimental data, which verifies that the improved theory (mechanical model) has a certain accuracy.

Tea is the most consumed beverage worldwide, and l-theanine in tea leaves significantly affects their flavor and market quality. We have developed and validated a fast and reliable gas chromatographic method with flame ionization detection (GC-FID) to quantify l-theanine after its extraction from Camellia sinensis(tea plant) and derivatization. The procedure was completed in 40 min, from extraction to chromatographic analysis, with a recovery rate of more than 93% and allowing a high sample throughput. The GC-FID intraday precision was within 0.57–2.28%, while the interday precision ranged from 1.57 to 13.48%. The intraday accuracy ranged from −6.84 to 5.26%, while the interday accuracy ranged from −1.08 to 3.12%. The limit of detection was 2.28 μg/mL, and the limit of quantification was 6.47 μg/mL. The GC-FID method was validated by high-performance liquid chromatography with UV detection (HPLC–UV) and was used to investigate the biosynthesis and regulation of l-theanine in tea plants. We found that plants fed with ethylamine significantly increased l-theanine concentrations in roots, while exogenous supplementation of glutamic acid, carbamide, and glutamine did not significantly affect the l-theanine level in roots. Our results also indicated that roots were not indispensable for the biosynthesis of l-theanine, which was detected in undifferentiated embryonic calluses in concentrations (g/100 g dry weight) as high as in leaves of whole plants (1.67 and 1.57%, respectively) and without any exogenous theanine precursor supplementation.

This article is concerned with the distributed event/self-triggered synchronization problem for general linear multiagent systems with partial input saturation. Both the event-based and self-triggered laws are designed using the local sampled, possibly saturated, state, which ensures the bounded synchronization of the multiagent systems and exclusion of the Zeno-behavior. The continuous communication between agents is avoided under these triggering protocols. Different from the existing related works, we show the fully distributed design for multiagent systems, where the synchronization criteria, the designed input laws, and the proposed triggering protocols do not depend on any global information of the communication topology. In addition, the computational load of multiagent systems is reduced significantly.

In this paper, the event-triggered guaranteed cost consensus problem is considered for second-order multi-agent systems whose system states are unmeasurable. First, an observer-based distributed event-triggered condition is presented. The triggering time series for each agent is different, and the controller with a time-varying delay is updated only at the triggered times, which can reduce the communication time effectively. Second, using the event-triggered condition and the Lyapunov method, several sufficient conditions are obtained to achieve the guaranteed cost output feedback consensus. Moreover, for each event-triggered interval, a positive lower bound can be calculated, leading to the avoidance of Zeno behaviors. Third, appropriate controller and observer gains can be designed using the linear matrix inequality technique, which is verified through numerical simulations. [ABSTRACT FROM AUTHOR]

In this study, a class of finite-time consensus of multiple transmission control protocol/active queue management (TCP/AQM) networks is investigated on the basis of a design idea of multi-agent systems, and for the first time, to our knowledge, a novel congestion control concept with neural networks is proposed. First, the problem statement and design goal of the consensus of multiple TCP/AQM networks is given. Then, a pre-assigned finite-time function is introduced to ensure that the tracking error approaches a pre-defined area within finite time. Furthermore, by combining a barrier Lyapunov function and backstepping technique, a neural network-based distributed adaptive finite-time control protocol for the output consensus of multiple TCP/AQM networks is presented, which can effectively generate the desired controls and ensure that the convergent time of all errors has nothing to do with the initial condition and design parameters. In addition, all signals in the closed-loop system are bounded. Finally, an example is given to further illustrate the effectiveness of the theoretical finding presented. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the asymptotic output tracking control problem of probabilistic Boolean control networks. Firstly, based on the largest control invariant subset and the limit theory of Markov chains, the conditions are proposed to judge whether the output trajectory of a probabilistic Boolean control network can asymptotically track a constant reference signal via the state feedback control and output feedback control, respectively. Next, based on an augmented system, similar criteria are obtained to judge whether the output of a reference Boolean network is asymptotically trackable by the output trajectory of a probabilistic Boolean control network under the state feedback control and output feedback control, respectively. The corresponding controller design methods are given. Two examples are shown to illustrate the effectiveness of the obtained results. [ABSTRACT FROM AUTHOR]

Abstract This paper is concerned with local H ∞ synchronization control of uncertain complex networks, in which there are nonlinear coupling functions between network nodes, and the coupling terms as well as network nodes are subject to time-varying delays. In order to make a complex network achieve H ∞ synchronization, a non-fragile state feedback control scheme is used. By employing the Lyapunov functional method, a sufficient condition on the existence of H ∞ non-fragile controllers is derived. This condition can not only guarantee the asymptotic stability of each network node, but also ensure a prescribed robust H ∞ performance level for the synchronization error system under study. The effectiveness of the proposed method is finally verified through a numerical example. [ABSTRACT FROM AUTHOR]

Previously reported haemostatic reference intervals in normal pregnancy displayed considerable contradictions to establish convince gestational age-related haemostatic reference values. 30 clinical reports were recruited to collect and assemble existing clinical reports from the database D-dimer levels increased progressively with gestational ages and exceeded conventional value of 1 mg/L after 29–36 weeks, and reached a peak at 24 h postpartum with mean value of 6.44 mg/L [95% confidence interval (CI): 5.84 to 7.05] and returned to 0.79 mg/L (95% CI: 0.43 to 1.16) at 1–8 weeks postpartum. Analogously, the level of fibrinogen gradually increased throughout the pregnancy, and peaked at 48–72 h after birth, with mean value of 9.05 g/L (95% CI: 2.22 to 15.89) and then returned to 3.62 g/L (95% CI: 3.03 to 4.20) at 1–8 weeks postpartum. However, in the middle trimester, asynchronously prothromb in fragments 1 + 2 (F1+2) level elevated and reached a peak at 28–36 weeks with mean value of 3.05 nmol/L (95% CI: 2.41 to 3.70), and then decreased in the later trimester, and reached 1.92 nmol/L (95% CI: 0.58 to 3.27) at 48–72 h post-partum, close to normal levels. Previously reported gestational age-related haemostatic reference intervals in pregnancy could not be used as a standard.

Stabilization for multi-group coupled models stochastic by delay feedback control and nonlinear impulsive control are considered in this paper. Using graph theory and Lyapunov method, some sufficient conditions are acquired by some control methods. Those criteria are easier to verify and no need to solve any linear matrix inequalities. These results can be designed more easily in practice. At last, the effectiveness and advantage of the theoretical results are verified by an example.

Solvent-thermal conversion of biomass was promising for obtaining value-added chemicals. However, little was known about the interactions between solvents and biomass in the process, which hindered the effective utilization of biomass. The effects of γ-valerolactone (GVL) and H2O on enhancing pubescensdegradation via the cleavage of inter- and intramolecular linkages were studied. At 160 °C, H2O selectively promoted the cleavage of the intermolecular linkages by forming hydrogen bonds, making mainly contributions to hemicellulose dissolution, while GVL and H2O promoted lignin dissolution by forming hydrogen bonds with −OCH3group of lignin. H2O promoted the cleavage of β-(1,4)-glycosidic bonds in hemicellulose derived oligomers to xylose, while the oxygen in the ring of GVL might interact with hydroxyl groups of xylose unit to enhance the dehydration of xylose to furfural, whereas GVL with H2O promoted the depolymerization of lignin to oligomers mainly including β-O-4′ and β–β′ linkages connecting to G and S units.

The interactions between lignin oligomers and solvents determine the behaviors of lignin oligomers self-assembling into uniform lignin nanoparticles (LNPs). Herein, several alcohol solvents, which readily interact with the lignin oligomers, were adopted to study their effects during solvent shifting process for LNPs’ production. The lignin oligomers with widely distributed molecular weight and abundant guaiacyl units were extracted from wood waste (mainly consists of pine wood), exerting outstanding self-assembly capability. Uniform and spherical LNPs were generated in H2O-n-propanol cosolvent, whereas irregular LNPs were obtained in H2O-methanol cosolvent. The unsatisfactory self-assembly performance of the lignin oligomers in H2O-methanol cosolvent could be attributed to two aspects. On one hand, for the initial dissolution state, the distinguishing Hansen solubility parameter and polarity between methanol solvent and lignin oligomers resulted in the poor dispersion of the lignin oligomers. On the other hand, strong hydrogen bonds between methanol solvent and lignin oligomers during solvent shifting process, hindered the interactions among the lignin oligomers for self-assembly.

This paper investigates the finite-time H ∞ cluster consensus for a class of nonlinear multi-agent systems with uncertainties and external disturbances. First, a new aperiodic intermittent controller is designed, which considers both intermittent control and cluster consensus control. Second, the problem of finite-time H ∞ cluster control when the system encounters external disturbances is investigated. With the help of the Lyapunov stability criterion and linear matrix inequality analysis techniques(LMI), sufficient conditions for the system to achieve finite-time H ∞ cluster consensus control under the action of intermittent control strategy are obtained. Finally, the feasibility of the theoretical analysis is verified by numerical simulation. • The uncertainties, nonlinear dynamics, and external disturbances are considered. • The finite-time H ∞ consensus control of MAS under intermittent control is studied. • A new aperiodic intermittent control protocol is proposed. • Cluster consensus control enables the system to achieve multiple control objectives. [ABSTRACT FROM AUTHOR]

Maternal sepsis results in poor outcomes such as fetal or maternal death. The incidence and mortality rates of maternal sepsis vary in different places because of differences in economic development, race and medical conditions. Identifying the clinical features and determining possible mechanisms for avoiding morbidity and preventing poor outcomes would benefit committed patients. Therefore, this was an epidemiological study at a maternity transfer center in Southeast China that aimed to identify local disease features of maternal sepsis. To investigate the incidence and risk factors associated with maternal sepsis and its progression to severe sepsis in a large population-based birth cohort. This local epidemiological study was conducted in at a tertiary care center in Guangzhou, China, from 2015 to 2019. A total of 74,969 pregnant women experiencing childbirth were included in this study; Of these, 74 patients with maternal sepsis were diagnosed according to the sepsis criterion, and 118 patients without sepsis in the same period were selected randomly as the control group to study possible reasons for postpartum sepsis. This retrospective analysis covered the entire period from the first trimester to puerperium. Clinical data were collected using the hospital’s electronic medical record system. Multivariate logistic regression was used to analyze risk factors for maternal sepsis. The incidences of maternal sepsis, the maternal mortality, and the fetal mortality were 0.099%, 0.004%, and 0.007%, respectively. Septic shock was associated with a higher severity of illness. All poor outcomes (maternal or fetal death) occurred during pregnancy. Postpartum sepsis had the longest onset period, and was associated with premature rupture of fetal membranes and preeclampsia. Sepsis is an important cause of both maternal and fetal mortality. Herein, we describe an epidemiological study that evaluated the incidence, development, and prognosis of local maternal sepsis. Furthermore, the characteristics of maternal sepsis are likely due to unknown pathological mechanisms, and patients would benefit from identifying more effective treatments for maternal sepsis.

•Screening for c.800C>T is necessary for Chinese pregnant women with hypercholanemia.•Hypercholanemia occured early in NTCPD pregnant individuals.•Hypercholanaemia from NTCPD in pregnancy should be distinguished from ICP.

Traditional engine waveform analysis in time-domain fails to perform an accurate fault diagnosis when the fault waveform is very close to the normal waveform in time domain. A novel engine waveform analysis method is presented. In this paper, the aim is to perform fault diagnosis efficiently under such circumstances. This method is proposed by combining a new technique, called sensitive frequency band (SFB) selection, with the developed Hilbert–Huang transform (HHT). This can alleviate "mode mixing" by removing noise from the engine waveforms and reveal the time–frequency characteristics for a signal by deriving its time–frequency spectrum (TFS) distribution. The method is then applied to analyze the engine injector-pulse-width waveforms, and it works well for signal noise reduction and fault diagnosis.

The problem of observer-based decentralized event-triggered H ∞ control for networked systems is investigated in this paper. The system measurements and control inputs are grouped into several nodes. An event-triggered transmission scheme is introduced to choose those necessary signals to be transmitted from the plant to the observer and from the controller to the actuator, where signals from control input nodes and from measurement nodes are released not necessarily synchronously. Under the decentralized event-triggered transmission scheme, H ∞ performance analysis for the resultant closed-loop system is made, and an algorithm to designing suitable observers and controllers is presented as well. Moreover, a sufficient condition is derived such that the inter-event time is strictly greater than zero. Finally, two examples are given to show the effectiveness of the method proposed in this paper. [ABSTRACT FROM AUTHOR]