Abstract Reliable fault diagnosis for wind turbine main shaft bearings faces dual challenges: existing deep learning models demand excessive computational resources industrial deployment, and weak fault-induced acoustic emission signals under low-speed/heavy-load conditions are prone to noise interference, thereby degrading diagnostic reliability. To address these challenges, we propose SEDSNet, a lightweight residual network bearing diagnosis. SEDSNet integrates depthwise separable...

To repair a failed block in the erasure-coded storage system, multiple related blocks have to be retrieved from other nodes across network. Such process can lead significant incast-type traffics and delays. The existing efforts mainly try schedule transmission of requested different avoid network congestion. At their cores, they utilize part involved hosts rely on or aggregate file others. While we notice that, programmability capability today's devices (i.e., routers switches) bring great...

Fault identification for closed-loop control systems is a future trend in the field of fault diagnosis. Due to inherent feedback adjustment mechanism, system generally very robust external disturbances and internal noises. Closed-loop often encourage faults propagate inside systems, which may lead consequence that amplitude becomes smaller characteristics difference more inapparent. Hence, it has been challenging achieve such systems. Traditional methods are not particularly designed thus...

Fault detection for closed-loop control systems is the future development in field of fault diagnosis. Since a system generally very robust to external disturbances, has been challenging hot research area. Traditional data-driven methods are not particularly designed and thus can be improved. In this paper, new method proposed, which based on parity space system. The main principle our transform residual into original restrict false or leak caused by estimation uncertain states. More...

Placing popular data at the network edge helps reduce retrieval latency, but it also brings challenges to limited storage space.Currently, using available yet not necessarily reliable resources is common sense for space expansion, while deploying deduplication strategies a general method better utilization.However, contradiction arises when jointly implementing with unreliable resources.On one hand, policy stipulates that any chunk can be stored exactly once; on other use of imposes should...

Two new mid-infrared (MIR) nonlinear optical (NLO) chalcogenide crystals, Ba10Zn7In6S26 (1) and Ba10Zn7Ga6Se26 (2) were discovered by traditional solid state reactions. Compounds 1 2 are isostructural, they represent a novel three-dimensional (3D) framework made apex-sharing T2 supertetrahedron 2D layers that stacked alternately another kind of supertetrahedra though sharing apexes along the c direction, with T1 tetrahedra inserting in tunnels Ba2+ cations locating both voids tunnels. Among...

SINS/CNS integrated navigation system is a high precision for ballistic missiles. Non-Gaussian noises often exist in the process, which can lead to divergence of filter method. A new Kalman using maximum correntropy theory introduced handle non-Gaussian noises. The robustness it analyzed briefly. To verify superiority algorithm, experiments loosely coupled were conducted. calculated results show that has higher accuracy used missile presence

In this paper, a fault diagnosis method is proposed based on component-wise expectation maximization algorithm and k-means algorithm, it applied to diagnosing the of satellite attitude determination control system. First, Gaussian mixture model its traditional parameter estimation are reviewed. The expectation-maximization used estimate parameters model, which can lower computational complexity estimation. Moreover, diagnosis, including detection isolation, carried out algorithm. Finally,...

To solve the problem that external systematic errors of optical camera cannot be fully estimated due to limited computing resources, a unified dimensionality reduction representation method for is proposed, and autonomous relative navigation realized. The translational misalignment are converted into three-dimensional rotation error, whose differential model can established through specific attitude control appropriate assumption. Then, error motion state jointly in an augmented Kalman...

Current industrial system develops more and complex intelligent, whose safety reliability relies on fault diagnosis technology. In the age of big-data, data-driven becomes state art, demand for diagnostic toolbox also increases. this paper, theories detection models, both static model dynamic, are revealed. And then a is designed based MATLAB/GUIDE. end, toolbox, as well theories, validated by simulation study three water tank system.

In the process of satellite attitude determination, satellites or sensors themselves often encounter a variety turbulence influences due to complexity space environments.Such can lead mutation and non-Gaussian noises for determination system.To solve these problems, in this paper, we construct unified error model influences, which is noise model, propose an improved filter method restrict system enhance accuracy robustness.The combined with jitters vibrations actual firstly designed.Then...

As one of the most critical issues for target track, α-Jerk model is an effective maneuvering track model. Colored noise always exists in measurement process tracking system which usually influences performance. A novel approach based on unscented particle filter derived. The (UPF), uses Kalman (UKF) to produce importance density function, a modified (PF). This article deal with colored combined state variable extension thought. Simulation results indicate that has higher estimate accuracy...

Due to the complexity of satellite communication channel and large amount data in imaging scenarios, conventional SNR estimation method has low efficiency. Also, because long distance time delay between ground wireless transmission system, estimate state information a certain degree hysteresis. Reservoir computing(RC) is recent neurologically inspired concept for processing dependent that lends itself particularly well hardware implementation by using device physics conduct processing. So,...

Cold data contributes a large portion of the big today and is usually stored in secondary storage. Various sketch structures are implemented to represent elements provide constant-time membership queries. Considering volume infrequent access cold data, we envision three rationales representation: 1) parallelized insertion, 2) accurate representation, 3) query/deletion. The existing sketches, however, fail achieve these simultaneously. To this end, propose novel named Parallelized cuckoo...

This letter introduces a high-power-added efficiency (PAE) and high second harmonic suppression (SHS) power amplifier monolithic microwave integrated circuit (MMIC) designed for operation in the frequency range from 8.5 GHz to 10.5 GHz. The design is based on 0.25 µm gallium nitride (GaN) electron mobility transistor (HEMT) process. To achieve gain, three-stage structure employed. Additionally, SHS topology incorporated output matching network enhance while maintaining efficiency....

The key of data-driven fault detection method lies in the full and effective understanding detected data, fitting for data is an means to realize parameterization model. In this paper, polynomial model autoregressive are used estimate predict non-stationary stationary respectively, so as achieve detection. estimation accuracy parameter analyzed. relationship between prediction duration, window, order given theoretically. Finally, numerical simulation results given, which can provide some...

During the target tracking process, some observation data may be missing due to equipment problems or operation errors, which affect filtering process of state and position determination accuracy. Therefore, needs effectively compensated. This paper provides a method compensate by using characteristics BP neural network learning system aiming at dynamic system. The is trained complete system, then predicted network. simulation results for both linear nonlinear show that indeed effective,...

Combined with data-driven and artificial intelligence, in this paper, we propose a two-layer fault detection mode based on time series modeling BP neural network. The method is used to realize the autonomy intelligence of system detection. Firstly, introduce non-stable characteristics measurement data construct non-stationary model. Secondly, use parameter model determine node number input layer, network parameters will be adaptively determined. Thus, training can test structure, then...

This paper investigates the optimal transmission scheduling problem in remote state estimation systems over an unreliable wireless channel where evolves as a Markov chain. However, due to inaccurate observations of state, is modeled hidden We propose prediction algorithm based on Viterbi estimate state. To save sensor's energy, we consider sensor transmissions while balancing between performance and energy expenditure. By jointly considering formulate decision process. prove existence policy...

Practical applications of microaerial vehicle face significant challenges including imprecise localization, limited on‐board energy, and motion uncertainty. This paper focuses on the latter two issues. The core proposed energy‐optimal path planning algorithm is an energy consumption model deriving from real measurements a specific quadrotor utilizing 2D Gaussian distribution function to simulate uncertainty random drift. Based these models, we formulate optimal traversing 3D map with minimum...