Single-phase pulse-width modulation (PWM) rectifier is commonly used in a high-speed railway electric traction system. Occurrence of unexpected failure in the sensors of the detection system may lead ...to feedback values deviation and system degradation, which can be extremely detrimental to the operation safety of the electric locomotive. This paper presents a fast and reliable fault diagnosis and fault resilient control strategy for catenary current and dc-link voltage sensor faults in the control system for an electric traction single-phase PWM rectifier. Sliding mode observers are designed to produce analytical redundancy. In order to avoid unobservable states and fluctuation introduced by discrete dynamics in observer design, a novel description for switching variables is presented in system modeling. Normalized residuals are generated using measured and observed values. The fault diagnosis unit proposed can detect and isolate three types of sensor faults online by comparing residuals with thresholds. System reconfiguration is realized by substituting the observed values for the information of faulty sensors. Simulation and experimental results are demonstrated to validate the effectiveness of the strategy.
Cybersecurity event detection aims to detect and classify the occurrence of cybersecurity events from a large amount of data. Previous approaches to event detection have used trigger word detection ...as the entry point for the task. However, constructing trigger words requires selecting words that describe the type of event occurrence from a large number of event sentences, which makes annotation of the training corpus time-consuming and expensive. To solve this issue, based on the latest achievements of contrastive learning in sentence embedding, we propose a novel method called
T
rigger-free
C
ybersecurity
E
vent
D
etection Based on
C
ontrastive
L
earning (TCEDCL), which incorporates semantics into the representation model to detect events without triggers. To demonstrate the feasibility of our proposed TCEDCL method, we collected and constructed a dataset from numerous cybersecurity news and blog posts. Extensive experiments have shown that our proposed method performs exceptionally well in detecting cybersecurity events and enables the extension of new types of event detection.
In this paper, an interactive multitask learning method for Chinese text sentiment classification is proposed. Here, the classic BiLSTM + attention + CRF model is used to obtain full use of the ...interaction relationship between tasks, and it simultaneously solves the two tasks of emotional dictionary expansion and sentiment classification. The proposed method divides text sentiment classification and emotional dictionary expansion into primary task and subtask, and it adopts the Enhanced Language Representation with Informative Entities (ERNIE) model as the text representation learning model for the primary task. Then, through the maximum pooling layer and the fully connected layer, the text sentiment classification task is completed. Meanwhile, the classical BiLSTM + attention + CRF model is used to extract emotional words from the text in the subtask. In addition, the multitask information interaction mechanism is used, and the prediction information on the autonomous subtask is fed back into the potential representation of the two tasks. After iterative training, the performance of the two tasks is further optimized. Micro-blogs with COVID-19 are used here as the subject to form the experimental data set. The results demonstrate the superiority of the proposed method over other approaches, and they further verify the superiority of ERNIE over BERT, RoBERTa and XLNet for the sentiment classification of Chinese text.
The capabilities of accurate prediction of cyberattacks have long been desired as detection methods cannot avoid the damages caused by occurrences of cyberattack. Attack prediction still remains an ...open issue especially to specify the upcoming steps of an attack with the quickly evolving intelligent techniques at the attackers’ side. This study proposes a composite learning approach (namely CL-AP2), which fulfills this task in two phases of “attack portraying” and “attack prediction”: (1) (Attack Portraying) CL-AP2 generates a Temporal Attack Knowledge Graph (TAKG) from real-time system logs providing full knowledge that formulates time-aware entities related to attacks and the relations amongst them; Over the TAKG, a Tactic-based Cyber Kill Chain (TCKC) model highlights the attacker’s portrait via evaluation of behaviors in the past, i.e., presenting the tactical path and attack steps taken by the attacker; (2) (Attack Prediction) The Soft Actor–Critic algorithm applies to identify the most possible attack trajectory confined in the attack portrait; The transformer model finally derives the specific attack technique to be taken next.
Experiments have been performed versus the state-of-the-art counterparts over a public dataset and results indicate that: (1) CL-AP2 can effectively reveal the tactical path taken by the attacker and form a complete portrait of the attack; and (2) CL-AP2 excels in predicting attack techniques to be taken by attackers and providing the defense guidance against the predicted attacks.
Self-balancing two-wheeled robots (TWRs) have attracted significant attentions over the last decade around the world. However, the traditional controller performs worse when TWR is driven under ...severe traveling conditions like steep slopes, speed bumps and other uneven terrains on practical pavement, which leads to potential security issues. This paper takes uneven terrains into consideration in system modeling and presents a hierarchical fast terminal sliding mode approach in controller design. The sliding mode controller system is proven asymptotically stable in terms of Lyapunov stability theorem. According to intensive simulation results, the proposed control approach is effective to hold TWR balanced and track desired velocity by numerous analyses. Simultaneously, the control system can also reject disturbances from external environment and reduce shuddering through uneven terrains.
Inflammatory response is a kind of nonspecific immune response, with the central link of vascular response, which is mainly manifested by changes in neutrophils and vascular endothelial cells. In ...recent years, the in vivo and in vitro role of intravenous anesthetic propofol in inhibiting inflammatory response has been attracting more and more attention, but the anti-inflammatory mechanisms of propofol for mononuclear cells still remain undefined. In this study, proteomics analysis was applied to investigate protein expression profile changes in serum mononuclear cells following intervention of rats with endotoxemia using propofol. After two-dimensional electrophoresis and mass spectrometric identification, it has been found that the protein Annexin A1 was up-regulated in the propofol intervention group. Annexin A1 is a glucocorticoid-dependent anti-inflammatory protein. After detection using ELISA and Western blot assays, it has also been found that propofol can not only promote the expression of Annexin A1, but also inhibit the phosphorylation level of p38 and release of inflammatory factors (IL-1β, IL-6 and TNF-α) in rats with endotoxemia. In order to further determine the role of up-regulated expression of Annexin A1 in anti-inflammation of propofol, this gene was silenced in vitro in human THP-1 cells, to detect the phosphorylation status of p38 and release of inflammatory factors. The results show that Annexin A1 can negatively regulate phosphorylation of p38 and release of IL-1β, IL-6 and TNF-α in THP-1 cells following propofol intervention and lipopolysaccharide (LPS) stimulation. Our results clearly indicate that propofol can up-regulate Annexin A1 to inhibit the phosphorylation level of p38 and release of IL-1β, IL-6 and TNF-α, so as to inhibit inflammatory response. Therefore, it can be speculated that Annexin A1 might be the key signaling protein in the in vivo and in vitro anti-inflammatory mechanisms of propofol.
It has been a hot issue and also a difficulty to study the fault tolerance of multilevel converters. The topology changes to eight-switch three-phase inverter (ESTPI) if one leg of three-level NPC ...inverter is failed. An optimal modulation algorithm of ESTPI is proposed in this paper. Due to the loss of one leg, the DC-link voltage utilization of ESTPI is reduced by half. Hence, an overmodulation algorithm based on superposition principle is proposed to improve DC-link voltage utilization, without increasing DC-link voltage. The proposed method is easier to implement compared with conventional overmodulation strategies for less computation. Theoretical proof and experimental results show that the proposed overmodulation algorithm is effective, and the DC-link voltage utilization can be increased maximally by 10.3%.
Federated learning is a new machine learning framework that allows mutually distrusting clients to reap the benefits from the joint training model without explicitly disclosing their private ...datasets. However, the high communication cost between the cloud server and clients has become the main challenge due to the limited network bandwidth. Moreover, the model parameters it shares may be utilized to perform model inversion attacks. Aimed at these problems, a new scheme for highly efficient federated learning with strong privacy preservation in cloud computing is presented. We design a lightweight encryption protocol to provide provably privacy preservation while maintaining desirable model utility. Additionally, an efficient optimization strategy is employed to enhance the training efficiency. Under the defined threat model, we prove the proposed scheme is secure against the honest-but-curious server and extreme collusion. We evaluate the effectiveness of our scheme and compare it with existing related works on MNIST and UCI Human Activity Recognition Dataset. Results show that our scheme reduces the execution time by 20% and transmitted ciphertext size by 85% on average while achieving similar accuracy as the compared secure multiparty computation (SMC) based methods.
The four-switch three-phase inverter (FSTPI) is the fault-tolerant topology of the six-switch three-phase inverter (SSTPI), and the maximum output voltage is limited in the linear modulation region. ...Therefore, an overmodulation strategy for FSTPI is presented, based on virtual voltage vector and superposition principle, in order to improve the utilization of the DC voltage, enhance the output capacity of the inverter and extend the operation range of induction motor. Compared with the traditional overmodulation strategy, the proposed strategy can make the inverter transit smoothly from the linear operation to the six-step operation with simple online computation, which is convenient for digital realization and can possess balanced output voltages. Accordingly, the relationship between SVPWM and carrier-based PWM in the overmodulation region for FSTPI is derived. The simulation results illustrate the validity and effectiveness of the theoretical analysis.
PDF
