In this paper, we propose a novel method for fault diagnosis in micro-electromechanical system (MEMS) inertial sensors using a bidirectional long short-term memory (BLSTM)-based Hilbert–Huang transform (HHT) and a convolutional neural network (CNN). First, the method for fault diagnosis of inertial sensors is formulated into an HHT-based deep learning problem. Second, we present a new BLSTM-based empirical mode decomposition (EMD) method for converting one-dimensional inertial data into two-dimensional Hilbert spectra. Finally, a CNN is used to perform fault classification tasks that use time–frequency HHT spectrums as input. According to our experimental results, significantly improved performance can be achieved, on average, for the proposed BLSTM-based EMD algorithm in terms of EMD computational efficiency compared with state-of-the-art algorithms. In addition, the proposed fault diagnosis method achieves high accuracy in fault classification.

In this study, a comb-type capacitive accelerometer based on a silicon carbide (SiC) microstructure is presented and investigated by the finite element method (FEM). It has the advantages of low weight, small volume, and low cross-coupling. Compared with silicon(111) accelerometers with the same structure, it has a higher natural frequency. When the accelerometer vibrates, its resistive force consists of two main components: a viscous damping and an elastic damping force. It was found that viscous damping dominates at low frequency, and elastic damping dominates at high frequency. The second-order linear system of the accelerometer was analyzed in the time-frequency domain, and its dynamic characteristics were best when the gap between the capacitive plates was 1.23 μm. The range of this accelerometer was 0–100 g, which is 1.64 times that of a silicon(111) accelerometer with the same structure. In addition, the accelerometer could work normally at temperatures of up to 1200 °C, which is much higher than the working temperatures of silicon devices. Therefore, the proposed accelerometer showed superior performance compared to conventional silicon-based sensors for inertial measurements.

Target continuous tracking problem is an important research topic in the field of computer vision. To conquer difficulties of target loss under complex situation, size changes and illumination changes, an improved tracking module is designed via combining tracking module in TLD with KCF (Kernelized Correlation Filters). The comparison between TLD and the proposed algorithm shows that the performance of the latter is improved in tracking accuracy. Under complex situation of target occlusion, scale change, illumination change, etc., algorithm presented can output the tracking results stably, which is more robust and more suitable for long-term target tracking.

In this study, we investigated the viscosity, squeeze-film damping, and a SiC-based capacitive micro-accelerometer in rarefied air. A specific expression for the effective viscosity coefficient of the air was derived, and when the air pressure drops from the standard atmospheric pressure, the viscosity of the air will decrease accordingly. Decreases in the air pressure and the viscosity of the air lead to the change in the squeeze-film air damping in the micro-accelerometer, and both the viscous damping force and the elastic damping force of the air film between the moving electrode plate and the fixed electrode plate will also decrease. The damping coefficient and relative damping ratio of the micro-accelerometer in rarefied air were calculated, which was also confirmed by simulations. The changes of the damping coefficient and the relative damping ratio of the system will directly affect the dynamic characteristics of the micro-accelerometer. When the air pressure in the working environment is below the standard atmospheric pressure, the micro-accelerometer will be in an underdamping state. With the decrease in the air pressure, the working bandwidth of the micro-accelerometer will decrease significantly, and the resonant phenomenon may appear. However, the decrease in the air pressure will not have a notable impact on the response time of the micro-accelerometer. Therefore, this work provides a theoretical basis for the study of the performance characteristics of a SiC-based capacitive accelerometer in rarefied air.

This paper bring forward a kind of pendulum micromechanical angular accelerometer with force feedback, which might be proposed by us, first. The sensor adopts a pendulum with mass joined to the anchors by a pair torsion-spring beams as sensing device, differential-capacitor device for detecting the angular acceleration about the X axis, an electrostatic force device for feedback loop. The sensor's structure and basic operating principle were discussed. The theoretical relation of angular acceleration and output of sensor was derived in detail based on the geometrical and dynamical models. The simulation result shows that the scale factor of sensor is 0.388917 V / r / s and verifies the accuracy of the theories model. It can be used in robust motor control of industrial robots and aerocraft.