Raw Sensor Signals Research Articles

An adaptive Kalman filter approach for cardiorespiratory signal extraction and fusion of non-contacting sensors.

BackgroundExtracting cardiorespiratory signals from non-invasive and non-contacting sensor arrangements, i.e. magnetic induction sensors, is a challenging task. The respiratory and cardiac signals are mixed on top of a large and time-varying offset and are likely to be disturbed by measurement noise. Basic filtering techniques fail to extract relevant information for monitoring purposes.MethodsWe present a real-time filtering system based on an adaptive Kalman filter approach that separates signal offsets, respiratory and heart signals from three different sensor channels. It continuously estimates respiration and heart rates, which are fed back into the system model to enhance performance. Sensor and system noise covariance matrices are automatically adapted to the aimed application, thus improving the signal separation capabilities. We apply the filtering to two different subjects with different heart rates and sensor properties and compare the results to the non-adaptive version of the same Kalman filter. Also, the performance, depending on the initialization of the filters, is analyzed using three different configurations ranging from best to worst case.ResultsExtracted data are compared with reference heart rates derived from a standard pulse-photoplethysmographic sensor and respiration rates from a flowmeter. In the worst case for one of the subjects the adaptive filter obtains mean errors (standard deviations) of -0.2 min −1 (0.3 min −1) and -0.7 bpm (1.7 bpm) (compared to -0.2 min −1 (0.4 min −1) and 42.0 bpm (6.1 bpm) for the non-adaptive filter) for respiration and heart rate, respectively. In bad conditions the heart rate is only correctly measurable when the Kalman matrices are adapted to the target sensor signals. Also, the reduced mean error between the extracted offset and the raw sensor signal shows that adapting the Kalman filter continuously improves the ability to separate the desired signals from the raw sensor data. The average total computational time needed for the Kalman filters is under 25% of the total signal length rendering it possible to perform the filtering in real-time.ConclusionsIt is possible to measure in real-time heart and breathing rates using an adaptive Kalman filter approach. Adapting the Kalman filter matrices improves the estimation results and makes the filter universally deployable when measuring cardiorespiratory signals.

Distributed adaptive estimation of covariance matrix eigenvectors in wireless sensor networks with application to distributed PCA

We describe a distributed adaptive algorithm to estimate the eigenvectors corresponding to the Q largest or smallest eigenvalues of the network-wide sensor signal covariance matrix in a wireless sensor network. The proposed algorithm recursively updates the eigenvector estimates without explicitly constructing the full covariance matrix that defines them, i.e., without centralizing all the raw sensor signal observations. By only sharing fused Q-dimensional observations, each node obtains estimates of (a) the node-specific entries of the Q covariance matrix eigenvectors, and (b) Q-dimensional projections of the full set of sensor signal observations onto the Q eigenvectors. We also explain how the latter can be used for, e.g., compression and reconstruction of the sensor signal observations based on principal component analysis (PCA), in which each node acts as a data sink. We describe a version of the algorithm for fully-connected networks, as well as for partially-connected networks. In the latter case, we assume that the network has been pruned to a tree topology to avoid cycles in the network. We provide convergence proofs, as well as numerical simulations to demonstrate the convergence and optimality of the algorithm.

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.

Analysis of Residual Wavelet Scalogram for Machinery Fault Diagnosis

Wavelet analysis is a very useful tool for machinery faults diagnosis. However, actual application of wavelet analysis for machinery fault diagnosis in the field is still relatively rare. This is partly due to the fact that visual interpretation of wavelet results is often difficult and very challenging. This paper investigates an effective method to present wavelet analysis results in order to simplify the interpretation of wavelet analysis result for machinery faults diagnosis. Analysis of residual wavelet scalogram was proposed in this study as a mean to display and extract key faults signatures from raw sensor signals. Simulated signals were generated to test the feasibility of the proposed method. Test results showed that the proposed wavelet method provides a simple and more effective way to diagnose machinery faults.

A computationally efficient Kalman filter based estimator for updating look-up tables applied to NOx estimation in diesel engines

Nox estimation in diesel engines is an up-to-date problem but still some issues need to be solved. Raw sensor signals are not fast enough for real-time use while control-oriented models suffer from drift and aging. A control-oriented gray box model based on engine maps and calibrated off-line is used as benchmark model for Nox estimation. Calibration effort is important and engine data-dependent. This motivates the use of adaptive look-up tables. In addition to, look-up tables are often used in automotive control systems and there is a need for systematic methods that can estimate or update them on-line. For that purpose, Kalman filter (KF) based methods are explored as having the interesting property of tracking estimation error in a covariance matrix. Nevertheless, when coping with large systems, the computational burden is high, in terms of time and memory, compromising its implementation in commercial electronic control units. However look-up table estimation has a structure, that is here exploited to develop a memory and computationally efficient approximation to the KF, named Simplified Kalman filter (SKF). Convergence and robustness is evaluated in simulation and compared to both a full KF and a minimal steady-state version, that neglects the variance information. SKF is used for the online calibration of an adaptive model for Nox estimation in dynamic engine cycles. Prediction results are compared with the ones of the benchmark model and of the other methods. Furthermore, actual online estimation of Nox is solved by means of the proposed adaptive structure. Results on dynamic tests with a diesel engine and the computational study demonstrate the feasibility and capabilities of the method for an implementation in engine control units.

A measurement-domain adaptive beamforming approach for ultrasound instrument based on distributed compressed sensing: Initial development

High efficient acquisition of the sensor array signals and accurate reconstruction of the backscattering medium are important issues in ultrasound imaging instrument. This paper presents a novel measurement-domain adaptive beamforming approach (MABF) based on distributed compressed sensing (DCS) which seeks to simultaneously measure signals that are each individually sparse in some domain(s) and also mutually correlated with much few measurements under the Nyquist rate. Instead of sampling conventional backscattering signals at the Nyquist rate, few linear projections of the returned signal with random vectors are taken as measurements, which can reduce the amount of samples per channel greatly and makes the real-time transmission of sensor array data possible. Then high resolution ultrasound image is reconstructed from the few measurements of DCS directly by the proposed MABF algorithm without recovering the raw sensor signals with complex convex optimization algorithm. The simulated results show the effectiveness of the proposed method.

Ventilation Estimation Using a Multi-Sensor Integrated Measurement System

The multi-sensor Integrated Measurement System (IMS) is an indigenous activity monitor developed with grant support from the NIH's Exposure Biology Program, which is part of the Genes and Environment Initiative. The IMS has a piezoelectric respiration sensor secured at the level of the abdomen. This sensor's raw signal is used to estimate respiration variables to quantify the internal exposure (inhalation) to environmental pollutants. PURPOSE: To examine preliminary calibration analyses of the IMS respiration sensor for predicting breathing rate (BR) and volume (BV). METHODS: Fifty adults (32.6 ± 9.9 yrs) wore the IMS and a portable metabolic unit (criterion) while performing one of two activity routines (7 activities each and 7 min/activity). These consisted of both ambulatory and simulated free-living activities. The raw sensor signal was 'filtered' to eliminate tissue artifact and then used to estimate BR and BV. BR was estimated by first classifying individual signal frequencies (Hz) into specific ranges. The range with the highest correlation to the original signal was used to estimate BR using post-spectral analyses. BV was estimated using multiple linear regression analysis with estimated breathing rate and the 10th, 25th, 50th, 75th, and the 90th percentiles of the sensor signal for each minute as the predictor variables. RESULTS: The overall BR estimate had a root mean square error of 4.1 breaths/min. BR during light (computer work and filing papers) and moderate (cycling, walking over-ground and on a treadmill at different speeds and grades) activities were slightly overestimated while that for vigorous intensity activities (treadmill running, tennis, and basketball) were underestimated. BV estimates (18.9 to 43.9 L/min) had a mean absolute percent error of 26.6% and a root mean square error of 8.6 L/min. BV during light and vigorous activities were over and underestimated, respectively. CONCLUSION: Preliminary analyses suggest that the IMS respiration sensor has the potential to provide reasonable estimates of BR and BV. Although prediction errors were large for BV, the coefficient of determination for the mean predictions was high (R2= 0.96). This suggests that the errors are more systematic than random in nature and may be corrected by improving the prediction model. Supported by: NIH UO1 CA130783

Imprecise expectations for imprecise linear filtering

In most sensor measure based applications, the raw sensor signal has to be processed by an appropriate filter to increase the signal-to-noise ratio or simply to recover the signal to be measured. In both cases, the filter output is obtained by convoluting the sensor signal with a supposedly known appropriate impulse response. However, in many real life situations, this impulse response cannot be precisely specified. The filtered value can thus be considered as biased by this arbitrary choice of one impulse response among all possible impulse responses considered in this specific context. In this paper, we propose a new approach to perform filtering that aims at computing an interval valued signal containing all outputs of filtering processes involving a coherent family of conventional linear filters. This approach is based on a very straightforward extension of the expectation operator involving appropriate concave capacities.

Piezoelectric-paint-based two-dimensional phased sensor arrays for structural healthmonitoring of thin panels

A damage detection method based on an innovative 2D phased sensor array made ofpiezoelectric paint is proposed for in situ damage detection of a thin isotropic panel usingguided Lamb waves. A design analysis of candidate 2D arrays based on spiral,cruciform and circular element layouts is performed. In this study, a 2D phased sensorarray with a spiral configuration is fabricated using a piezoelectric composite(piezopaint) patch and used for detecting damages in an aluminum panel. Steered arrayresponses are generated from the raw sensor signals using a directional filteringalgorithm based on phased array signal processing. The fundamental flexural(or transverse), A0 mode, of the guided Lamb waves is used though the sensingand analysis technique is not limited to the mode used in this work. To enhancethe proposed analysis technique, empirical mode decomposition (EMD) and aHilbert–Huang transform (HHT) are applied. A new damage detection algorithmincluding threshold setting and damage index (DI) calculation is developed andimplemented for detecting damages in the form of holes and a simulated crack.The characteristic damage indices consistently increase as damage size grows.

Fluid level measurement in dynamic environments using a single ultrasonic sensor and Support Vector Machine (SVM)

A fluid level measurement system to accurately determine fluid levels in dynamic environments has been described. The measurement system is based on a single ultrasonic sensor and Support Vector Machine (SVM) based signal processing and classification scheme. For exemplification of the measurement system in dynamic environments, the novel measurement system is experimented and verified on a fuel tank of a running vehicle. The effects of slosh and temperature variations on the acoustic sensor based measurement system are reduced using the novel approach. The novel approach is based on ν-SVM classification method with the Radial Basis Function (RBF) to compensate for the measurement error induced by the sloshing effects in the tank due to the motion of the moving vehicle. In this approach, raw sensor signals are differentiated after smoothing with some selected pre-processing filters, namely, Moving Mean, Moving Median, and Wavelet filter. The derivative signal is then transformed into Frequency Domain to reduce the size of input features before performing the signal classification with SVM. Field trials were performed on actual vehicle under normal driving conditions at various fuel volumes ranging from 5 L to 50 L to acquire sample data from the ultrasonic sensor for the training of SVM model. Further drive trials were conducted to obtain data to verify the SVM results. A comparison of the accuracy of the predicted fluid level obtained using SVM and the pre-processing filters is provided. It is demonstrated that the ν-SVM model using the RBF kernel function and the Moving Median filter has produced the most accurate outcome compared with the other signal filtration methods in terms of fluid level measurement.

PIPELINE DEFECT PREDICTION USING SUPPORT VECTOR MACHINES

Oil and gas pipeline condition monitoring is a potentially challenging process due to varying temperature conditions, harshness of the flowing commodity and unpredictable terrains. Pipeline breakdown can potentially cost millions of dollars worth of loss, not to mention the serious environmental damage caused by the leaking commodity. The proposed techniques, although implemented on a lab scale experimental rig, ultimately aim at providing a continuous monitoring system using an array of different sensors strategically positioned on the surface of the pipeline. The sensors used are piezoelectric ultrasonic sensors. The raw sensor signal will be first processed using the discrete wavelet transform (DWT) as a feature extractor and then classified using the powerful learning machine called the support vector machine (SVM). Preliminary tests show that the sensors can detect the presence of wall thinning in a steel pipe by classifying the attenuation and frequency changes of the propagating lamb waves. The SVM algorithm was able to classify the signals as abnormal in the presence of wall thinning.

Acquisition of Flexible Image Recognition by Coupling of Reinforcement Learning and a Neural Network

The authors have proposed a very simple autonomous learning system consisting of one neural network (NN), whose inputs are raw sensor signals and whose outputs are directly passed to actuators as control signals, and which is trained by using reinforcement learning (RL). However, the current opinion seems that such simple learning systems do not actually work on complicated tasks in the real world. In this paper, with a view to developing higher functions in robots, the authors bring up the necessity to introduce autonomous learning of a massively parallel and cohesively flexible system with massive inputs based on the consideration about the brain architecture and the sequential property of our consciousness. The authors also bring up the necessity to place more importance on optimization of the total system under a uniform criterion than understandability for humans. Thus, the authors attempt to stress the importance of their proposed system when considering the future research on robot intelligence. The experimental result in a real- world-like environment shows that image recognition from as many as 6240 visual signals can be acquired through RL under various backgrounds and light conditions without providing any knowledge about image processing or the target object. It works even for camera image inputs that were not experienced in learning. In the hidden layer, template-like representation, division of roles between hidden neurons, and representation to detect the target uninfluenced by light condition or background were observed after learning. The autonomous acquisition of such useful representations or functions makes us feel the potential towards avoidance of the frame problem and the development of higher functions.

Application of empirical mode decomposition based energy ratio to vortex flowmeter state diagnosis

To improve the measurement performance, a method for diagnosing the state of vortex flowmeter under various flow conditions was presented. The raw sensor signal of the vortex flowmeter was adaptively decomposed into intrinsic mode functions using the empirical mode decomposition approach. Based on the empirical mode decomposition results, the energy of each intrinsic mode function was extracted, and the vortex energy ratio was proposed to analyze how the perturbation in the flow affected the measurement performance of the vortex flowmeter. The relationship between the vortex energy ratio of the signal and the flow condition was established. The results show that the vortex energy ratio is sensitive to the flow condition and ideal for the characterization of the vortex flowmeter signal. Moreover, the vortex energy ratio under normal flow condition is greater than 80%, which can be adopted as an indicator to diagnose the state of a vortex flowmeter.

On-line validation of the measurement uncertainty of a sensor using wavelet transforms

The measurement uncertainty of a sensor as a measure of its accuracy is normally derived through static analysis. Conventionally, uncertainty analysis is conducted during the design and development stage of a sensor to select appropriate components and measurement techniques, as well as interpreting its experimental data. The demand for quality assurance by measurement is now increasing. It is therefore desirable to develop a means of assessing the measurement uncertainty of a sensor during its operation on a process plant. A new method using wavelet transforms for the validation of the measurement uncertainty is proposed. Analytical results show that the process variable being measured by the sensor can be separated from a noisy raw sensor signal using its wavelet transforms, provided that the process variable is represented in terms of a limited-degree polynomial function. Unlike the conventional approach to uncertainty calculation, which requires ‘average’ or ‘typical’ values substituted for parameters which may vary, the proposed method uses only the latest output of the sensor regardless of the variations of the parameters, and thus can be applied on an online continuous basis. Experimental results obtained from a differential-pressure flow sensor on a water flow test rig confirm the effectiveness of the proposed method.

The Shape From Motion Approach to Rapid and Precise Force/Torque Sensor Calibration

We present a new technique for multi-axis force/torque sensor calibration called shape from motion. The novel aspect of this technique is that it does not require explicit knowledge of the redundant applied load vectors, yet it retains the noise rejection of a highly redundant data set and the rigor of least squares. The result is a much faster, slightly more accurate calibration procedure. A constant-magnitude force (produced by a mass in a gravity field) is randomly moved through the sensing space while raw data is continuously gathered. Using only the raw sensor signals, the motion of the force vector (the “motion”) and the calibration matrix (the “shape”) are simultaneously extracted by singular value decomposition. We have applied this technique to several types of force/torque sensors and present experimental results for a 2-DOF fingertip and a 6-DOF wrist sensor with comparisons to the standard least squares approach.

Using multivariate statistical methods to detect fires

Fire detectors must accurately detect fires, but they should not respond to false alarms. Contemporary smoke detectors sometimes cannot discriminate between smoke and odor sources. These detectors can also be slow in responding to smoldering fire sources. In this paper, a statistical approach for detecting fires based on fusing sensor signals from multiple sensors is presented. The multivariate statistical approach, called principal component analysis, is used to compress the sensor information down to a small number of variables that can be interpreted more easily than the raw sensor signals themselves. Experimental results presented here show that the proposed approach is more accurate than a conventional smoke alarm, particularly for early detection of smoldering fires. However, this new approach does not overcome the problem of false alarms. In spite of this current limitation, the method discussed holds great promise for future fire detection applications.