Measured Sensor Signals Research Articles

Development of Highly Efficient Energy Harvester Based on Magnetic Field Emanating From a Household Power Line and Its Autonomous Interface Electronics

A magnetic energy harvester and its interface electronics are developed to operate a surface acoustic wave (SAW) oscillator sensor that requires at least 5–6 V dc (50–100 mW) to drive and transmit the measured sensor signals to the reader system wirelessly via antennas. The developed energy harvester system comprises a magnetic energy harvester with a coil wound around a magnetic core, an energy storage interface, and a power management interface. From a 220-V, 3-A household ac power line, a 12-V dc is harvested in the storage capacitor via a 5-min energy scavenging process to activate the SAW oscillator sensor. After 2 min of operation using the sensor system, the power management interface is completely shut down to recharge the storage capacitor and then restarted after sufficient charges are accumulated in the storage capacitor to reoperate the SAW oscillator sensor. This autonomous sequence is continuously repeated. The entire interface electronics is developed on a single printed circuit board (PCB) to reduce the energy loss by the board and to facilitate installation in the desired applications. A simulation is performed in COMSOL to determine the optimal energy harvester parameters and to predict the magnetic flux density that forms the induced current around the ferrite magnetic core based on the magnetic core shape. Experimental results show that the developed energy harvester system can harvest sufficient energy from the household power line in a short period to operate the complex RF SAW sensor and the sensor interface electronics. Furthermore, it can be used to operate the SAW oscillator sensor for the time required to obtain valid information to be transmitted wirelessly to the reader system.

Steering Actuator Fault Diagnosis for Autonomous Vehicle With an Adaptive Denoising Residual Network

A steering actuator is the core component of an autonomous vehicle. Because of noise interference, the extraction of useful fault features for an accurate fault diagnosis of steering actuator becomes an arduous task. This paper proposes an adaptive de-noising residual network (AD-ResNet) for fault diagnosis of steering actuator with reduced noise interference and improved accuracy. First, a high-complexity model based on long short-term memory is built to estimate redundant signals, and then fault features are generated by comparing redundant signals with the measured sensor signals. Second, an attention unit is designed to determines the unique set of thresholds for each fault feature. Finally, hard thresholding is embedded in a deep network as a nonlinear transform layer to reduce noise. Simulation experiments were performed at different signal-to-noise ratios, and the results confirmed that the proposed method could adaptively de-noise the high-noise signals. The proposed method was then validated against a fault dataset that was built using road experiments with an autonomous vehicle, and it achieved an accuracy, a sensitivity, a specificity, and an F1-score of 91.66%, 91.66%, 97.22%, and 0.9159, respectively, averaged over all fault states. The proposed method outperformed the comparison methods in the experiments.

Unsupervised fault detection and recovery for intelligent robotic rollators

Fault diagnosis is a key safety component in robotic assistive technologies. Although conventional model-based methods for sensor fault diagnosis in mobile robots have been well established, they face challenges due to model parameter changes and uncertainties. On the other hand, data-driven approaches becomes more appealing in order to take advantage from available historical data in the era of Big Data. To provide a new generic unsupervised solution to the fault detection and recovery, we explicitly include kinematic relations and temporal finite differences from measured sensor signals into training a multi-task deep neural network. To evaluate the proposed fault diagnosis and recovery framework, experiments have been conducted on a robotic rollator platform. Experiments under several conditions confirm that the proposed approach, which leverages machine learning-enhanced algorithms, exhibits reliable performance. Outperforming other baselines and state-of-the-art diagnosis algorithms, the framework presents a promising solution to sensor fault recovery challenges in assistive devices.

A Basic Study on Detection of Movement State in Stride by Artificial Neural Network for Estimating Stride Length of Hemiplegic Gait Using IMU.

In order to perform rehabilitation training for improving motor function, measurement of movements and evaluation of motor function become effective. In our research group, the method of estimating stride length during walking by using an inertial sensor attached to the foot was developed. However, since the method used thresholds to detect movement state in each stride for calculation of stride length, there was a difficulty in determination of threshold values for each subject and each stride with hemiplegic subjects in some cases. This study aimed at developing an automatic detection method of movement state in stride by artificial neural network (ANN) for hemiplegic gait. In this paper, three-layer ANN and four-layer ANN with feature extraction layers by autoencoder were tested. Teacher signals were obtained from measured sensor signals by the threshold-based method. The ANN with feature extraction layers was shown to be effective for detecting the movement state of healthy subjects and a hemiplegic subject. The movement state detected by ANN was also suggested to be effective in stride length estimation. It is expected to evaluate the ANN-based method using data measured with more hemiplegic subjects.

Convolutional neural network‐based wind‐induced response estimation model for tall buildings

AbstractThis study presents a convolutional neural network (CNN)‐based response estimation model for structural health monitoring (SHM) of tall buildings subject to wind loads. In this model, the wind‐induced responses are estimated by CNN trained with previously measured sensor signals; this enables the SHM system to operate stably even when a sensor fault or data loss occurs. In the presented model, top‐level wind‐induced displacement in the time and frequency domains, and wind data in the frequency domain are configured into the input map of the CNN to reflect the resisting capacity of a tall building, the change in the dynamic characteristics of the building due to wind loads, and the relationship between wind load and the building. To evaluate stress, which is used as a safety indicator for structural members in the building, the maximum and minimum strains of columns are set as the output layer of the CNN. The CNN is trained using measured wind and wind response data to predict the column strains during a future wind load. The presented model is validated using data from a wind tunnel test of a building model. The performance of the presented model is verified through strain estimation with data that were not used in the CNN training. To assess the validity of the presented input map configuration, the estimation performance is compared with a CNN that considered only the time domain responses as input. Furthermore, the effects of the variations in the configuration of the CNN on the wind response estimation performance are examined.

Simulation of a NOx Sensor for Model-Based Control of Exhaust Aftertreatment Systems

Optimal control of the exhaust aftertreatment components is an important aspect for minimizing pollutants. Nitrogen oxides are therefore monitored using amperometric gas sensors. However, they show marked cross-sensitivities to ammonia, which can be disadvantageous, e.g., if they are applied in NH3 selective catalytic reduction systems. In this study, a model that simulates the operation of a NOx sensor is developed. The three oxygen pumping cells and the upstream diffusion barriers were implemented in a 1D geometry, with the relevant electrode reactions taking place over its length. The diffusion barriers were calculated quasi-stationary. The dynamics of incoming gas components as a function of the exhaust mass flow were integrated with a first order linear time invariant system. The problem was calculated using the convection–diffusion equation. The reaction kinetics were determined based on stationary measurements in laboratory atmosphere, whereby initially only the reaction with respect to NO and NO2 was considered. Then, the cross-sensitivities to oxygen, water and ammonia were examined. The developed data set could then be used for transient boundary conditions. The simulated and the measured sensor signals in the case of engine test bench conditions with real driving cycles agree very well.

Determination Scheme for Accurate Defect Depth in Underground Pipeline Inspection by Using Magnetic Flux Leakage Sensors

The magnetic flux leakage-type nondestructive testing has been applied for the highly efficient inspection of defects in ferromagnetic materials such as underground gas pipelines. In the system, the magnetic field is applied to magnetize a steel pipe so as to induce the leakage signal in the vicinity of defects on the pipe. From the measured sensor signals, decomposing or estimating sizes and shapes of defects are necessary. For the maintenance, the depth estimation is the most important procedure for the management of a safety accident. However, the previous depth estimation contains high error rate inevitably because the depth signal implies the width and length error simultaneously. This paper focused on the enhanced algorithm for the depth sizing in various defects by using the decoupling algorithm to enhance the accuracy of defect depth. Estimated results show good agreements with measured ones.

Coulometric trace humidity measurement in technical gases.

Trace humidity was measured by using miniaturized planar coulometric sensors in technical gases such as hydrogen, nitrogen, helium, nitrous oxide, and synthetic air. Frost point temperatures in the gases ranged from -60 °C to -30 °C, which is equivalent to a vapour mole fraction from 10 μmol mol-1 to 376 μmol mol-1. In addition, the generated humidity was determined by using a precision dew point hygrometer as reference. Nonlinear calibration functions were calculated that correlated electric current (sensor signal) and reference humidity. Parameters of functions were tested with one-way analysis of variances (ANOVA) to prove if all used sensors had a similar behavior in the same gas during experiments. Results of ANOVA confirmed that averaged functions can be applied for trace humidity measurement in nitrogen, helium, nitrous oxide, and synthetic air. The calculated functions were negligibly different for nitrogen, helium, and synthetic air. In humidified nitrous oxide, a minor change of parameters was observed due to lower electrical currents. In total contrast to that, the measured sensor signals were significantly higher in humidified hydrogen and each sensor required its own calibration function. The reason was a recombination effect that favoured multiple measurements of water molecules. Nevertheless, it was possible to measure continuously trace humidity in all tested gases by using coulometric sensors with an expanded uncertainty below 2 K (k = 2).

Mover Position Detection for PMTLM Based on Linear Hall Sensors through EKF Processing.

Accurate mover position is vital for a permanent magnet tubular linear motor (PMTLM) control system. In this paper, two linear Hall sensors are utilized to detect the mover position. However, Hall sensor signals contain third-order harmonics, creating errors in mover position detection. To filter out the third-order harmonics, a signal processing method based on the extended Kalman filter (EKF) is presented. The limitation of conventional processing method is first analyzed, and then EKF is adopted to detect the mover position. In the EKF model, the amplitude of the fundamental component and the percentage of the harmonic component are taken as state variables, and they can be estimated based solely on the measured sensor signals. Then, the harmonic component can be calculated and eliminated. The proposed method has the advantages of faster convergence, better stability and higher accuracy. Finally, experimental results validate the effectiveness and superiority of the proposed method.

Advanced braking algorithm for robust longitudinal risk management

This paper describes an advanced braking algorithm for robust longitudinal risk management to prevent or mitigate rear-end collisions. Since the proposed safety system works in conjunction with a human driver, the longitudinal safety system must be acceptable to the driver. The key to achieving this is to ensure that control signals are given at the appropriate time. In order to construct an suitable control system, a new longitudinal safety index was developed by using the time to collision and the warning index, which are well-known safety parameters in longitudinal safety control. The clearance and the relative velocity constitute essential signals for safety monitoring in longitudinal safety control. These measured sensor signals include the uncertainty in the measurement noise. To design a robust safety monitoring and control mode decision, information consisting of fusion of the measurement signals was obtained. In addition, an expected error range of the proposed safety index from the measurement noise can be defined from the covariance matrix of the Kalman filter and the deviation of the function of the new longitudinal safety index. The control performance of the proposed algorithm was evaluated by analysis of the simulation results. From this analysis, it can be concluded that the proposed advanced braking algorithm can substantially enhance the longitudinal safety and can guarantee a robust performance with respect to the sensor uncertainty.

Preparation for commissioning of structural sensors of Wendelstein 7-X magnet system

The modular stellarator Wendelstein 7-X (W7-X) in Greifswald, Germany is currently under commissioning. The magnet system (MS) consists of five symmetric modules, each one containing two flip-symmetric half modules of five non-planar and two planar superconducting coils operated in the cryostat system (CS) at 4K. In order to validate the FE models that formed the basis of the design and the operational limits, the MS and CS were instrumented with more than 3000 sensors.The paper gives an overview of the numerical tools developed to enable safe operation of the MS based on comparisons between measured sensor signals and their FE model predictions. Output extraction from the FE models was standardized for all potentially critical components and all sensor locations. A graphical user interface (GUI) was programmed to compare critical results of different simulations and models.The GUI shows the design limit for each component. With local models, the safe design of each component was demonstrated against these design limits. The GUI also gives the degree of utilization of each component as the ratio between the calculated loads relative to the design limit. The highest degree of utilization over all components was used to define safe modes of operation and to set limits on allowed deviations of the measured signals relative to the projection before operation has to be interrupted.

Detection of the position, direction and speed of sliding contact with a multi-layer compliant tactile sensor fabricated using direct-print technology

A multi-layer resistance based compliant tactile sensor was fabricated using direct-print (DP) and soft molding processes. The sensor consists of two layers of embedded stretchable sensing elements sandwiched by three layers of a polyurethane rubber material. The sensing elements were created by the DP process using a photopolymer filled with multi-wall carbon nanotubes, which exhibit the property of piezoresistivity. The printed sensing elements were fully cured using ultraviolet light. The sensing elements within each layer of the sensor structure change in electrical resistance when external forces are applied. By processing the measured sensor signals, the fabricated sensor was able to detect the position of contact forces with a 3 mm spatial resolution, as well as their two-dimensional translation directions and speeds. Based on the results, it is concluded that the fabricated sensors are promising in robotic applications and the developed process and material can be a reliable and robust way to build highly stretchable tactile sensors.

Experimental control of natural perturbations in channel flow

AbstractA combined approach using system identification and feed-forward control design has been applied to experimental laminar channel flow in an effort to reduce the naturally occurring disturbance level. A simple blowing/suction strategy was capable of reducing the standard deviation of the measured sensor signal by 45 %, which markedly exceeds previously obtained results under comparable conditions. A comparable reduction could be verified over a significant streamwise extent, implying an improvement over previous, more localized disturbance control. The technique is effective, flexible, and robust, and the obtained results encourage further explorations of experimental control of convection-dominated flows.

Development of Automatic Wafer Centering System for Vacuum Transfer Robot Using for Semiconductor Manufacturing

The automatic wafer centering system having the function as wafer slip is measured and arm is controlled to compensate the wafer position error by slip is proposed. Two dimensional positions to compensate the wafer slip from measured sensor signal are calculated by using the compensation algorithm. Automatic wafer centering system is developed and verification test is conducted. In the verification test, developed automatic wafer centering system has measurement accuracy of under 0.2mm for radial and tangential direction.

딥스틱게이지형 소형 엔진오일열화감지센서의 시험결과 분석

This paper presents the test results of small dipstick-gage-type engine-oil-deterioration-detection sensor. The measured sensor signal characteristics for the capacitance and temperature are analyzed. The engine oil deterioration condition correlates with the electrical property of the dielectric constant that comprised with physical properties such as TAN (Total Acid Number), TBN (Total Base Number) and viscosity. Several problems encontered during the test of the sensor system are improved. The results of vehicle tests show that the capacitance signal is stable after the engine stops. Therefore, the sensor should start measuring the parameters for monitoring the engine oil condition after the engine stops. The engine is considered to be in a stopped state if the difference between the maximum and minimum values of the oil capacitance measured every 1 min is below 0.02 pF. The key test results in this paper will help in the development of an engine oil change warning algorithm.

Dynamic model for automotive side impact crashes

A rigid body model to represent a side impact crash is constructed using five degrees-of-freedom (dof) for the vehicle and three dof for each occupant in the vehicle. Nonlinear stiffness and damping elements and the presence of physical gaps between several components make the model highly nonlinear. The model is validated using experimental crash test data from a National Highway Traffic Safety Administration (NHTSA) database. To simplify the parameter identification process and reduce the number of parameters to be identified at each stage, a two-step process is adopted in which the vehicle is first assumed to be unaffected by the presence of the occupants, and its model parameters are identified. Subsequently, the parameters in the occupant models are identified.The active set method with a performance index that includes both the L2 and L∞ norms is used for parameter identification. A challenge is posed by the fact that the optimisation problem involved is non-convex. To overcome this challenge, a large set of random initial values of parameter estimates is generated and the optimisation method is applied with all these initial conditions. The values of parameters that provide the minimal performance index from the entire set of initial conditions are then chosen as the best parameter values. The optimal parameters values thus identified are shown to significantly improve the match between the model responses and the experimentally measured sensor signals from the NHTSA crash test.

An adaptive learning damage estimation method for structural health monitoring

Structural health monitoring is an important problem of interest in many civil infrastructure and aerospace applications. In the last few decades, many techniques have been investigated to address the detection, estimation, and classification of damage in structural components. One of the key challenges in the development of real-world damage identification systems, however, is variability due to changing environmental and operational conditions. Conventional statistical methods based on static modeling frameworks can prove to be inadequate in a dynamic and fast changing environment, especially when a sufficient amount of data is not available. In this paper, a novel adaptive learning structural damage estimation method is proposed in which the stochastic models are allowed to perpetually change with the time-varying conditions. The adaptive learning framework is based on the use of Dirichlet process (DP) mixture models, which provide the capability of automatically adjusting to structure within the data. Specifically, time–frequency features are extracted from periodically collected structural data (measured sensor signals), that are responses to ultrasonic excitation of the material. These are then modeled using a DP mixture model that allows for a growing, possibly infinite, number of mixture components or latent clusters. Combined with input from physically based damage growth models, the adaptively identified clusters are used in a state-space setting to effectively estimate damage states within the structure under varying external conditions. Additionally, a data selection methodology is implemented to enable judicious selection of informative measurements for maximum performance. The utility of the proposed algorithm is demonstrated by application to the estimation of fatigue-induced damage in an aluminum compact tension sample subjected to variable-amplitude cyclic loading.

Real-Time Identification of Incipient Surface Morphology Variations in Ultraprecision Machining Process

Real-time monitoring and control of surface morphology variations in their incipient stages are vital for assuring nanometric range finish in the ultraprecision machining (UPM) process. A real-time monitoring approach, based on predicting and updating the process states from sensor signals (using advanced neural networks (NNs) and Bayesian analysis) is reported for detecting the incipient surface variations in UPM. An ultraprecision diamond turning machine is instrumented with three miniature accelerometers, a three-axis piezoelectric dynamometer, and an acoustic emission (AE) sensor for process monitoring. The machine tool is used for face-turning aluminum 6061 discs to a surface finish (Ra) in the range of 15–25 nm. While the sensor signals (especially the vibration signal in the feed direction) are sensitive to surface variations, the extraneous noise from the environment, machine elements, and sensing system prevents direct use of raw signal patterns for early detection of surface variations. Also, nonlinear and time-varying nature of the process dynamics does not lend conventional statistical process monitoring techniques suitable for characterizing UPM-machined surfaces. Consequently, instead of just monitoring the raw sensor signal patterns, the nonlinear process dynamics wherefrom the signal evolves are more effectively captured using a recurrent predictor neural network (RPNN). The parameters of the RPNN (weights and biases) serve as the surrogates of the process states, which are updated in real-time, based on measured sensor signals using a Bayesian particle filter (PF) technique. We show that the PF-updated RPNN can effectively capture the complex signal evolution patterns. We use a mean-shift statistic, estimated from the PF-estimated surrogate states, to detect surface variation-induced changes in the process dynamics. Experimental investigations show that variations in surface characteristics can be detected within 15 ms of their inception using the present approach, as opposed to 30 ms or higher with the conventional statistical change detection methods tested.

Crack Damage Assessment Based on Gaussian Process Model

The material dispersion of structure led to a lager error in crack length evaluating, the assessment method utilizing Gaussian Process (GP) model was proposed to solve the problem. The fatigue crack was monitored by active Lamb monitoring technology, and the four damage indices were extracted from the measured sensor signal, and then inputted to the GP model to realize the online evaluating of crack length. The fatigue test of hole-edge crack was made in LY12-CZ Aluminum specimen, which was used in aerospace structures frequently, the results shows that the method could efficiently decrease the evaluation error of crack length, which caused by material dispersion of structure.

Real-Time Autocalibration of MEMS Accelerometers

In this paper, the self-calibration of micromechanical acceleration sensors is considered, specifically, based solely on user-generated movement data without the support of laboratory equipment or external sources. The autocalibration algorithm itself uses the fact that under static conditions, the squared norm of the measured sensor signal should match the magnitude of the gravity vector. The resulting nonlinear optimization problem is solved using robust statistical linearization instead of the common analytical linearization for computing bias and scale factors of the accelerometer. To control the forgetting rate of the calibration algorithm, artificial process noise models are developed and compared with conventional ones. The calibration methodology is tested using arbitrarily captured acceleration profiles of the human daily routine and shows that the developed algorithm can significantly reject any misconfiguration of the acceleration sensor.