An adaptive Kalman filtering algorithm based on back-propagation (BP) neural network applied for simultaneously detection of exhaled CO and N2O

Zhou, Sheng; Liu, Ningwu; Shen, Chongyang; Zhang, Lei; He, Tianbo; Yu, Benli; Li, Jingsong

doi:10.1016/j.saa.2019.117332

Cited by 38 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To reduce external interferences, several methods had been proposed to improve the readout system, such as the adaptive filter [ 18 , 19 ], least squares [ 20 , 21 ], minimum mean square error (MMSE) [ 22 ], maximum a posteriori probability (MAP) [ 23 ], iterative method [ 24 ], Wiener filter [ 25 ], and Kalman filter [ 26 ]. Regarding the natural properties and application of pressure sensors, the Kalman filter has a superior performance for the readout system with a single variable of input, a linearly response output, a reduction of noise and a high efficiency of prediction [ 26 , 27 , 28 ]. Therefore, a newly developed readout system integrated with a microprocessor, impedance converter, and algorithm design of the Kalman filter is first proposed and illustrated in the current study for a pressure sensor array of 14 × 18 pixels on a textile-based mattress for clinic interfacial pressure monitoring.…”

Section: Introductionmentioning

confidence: 99%

Flexible Textile-Based Pressure Sensing System Applied in the Operating Room for Pressure Injury Monitoring of Cardiac Operation Patients

Shih¹,

Wang

Chao

et al. 2020

Sensors

View full text Add to dashboard Cite

Pressure injury is the most important issue facing paralysis patients and the elderly, especially in long-term care or nursing. A new interfacial pressure sensing system combined with a flexible textile-based pressure sensor array and a real-time readout system improved by the Kalman filter is proposed to monitor interfacial pressure progress in the cardiac operation. With the design of the Kalman filter and parameter optimization, noise immunity can be improved by approximately 72%. Additionally, cardiac operation patients were selected to test this developed system for the direct correlation between pressure injury and interfacial pressure for the first time. The pressure progress of the operation time was recorded and presented with the visible data by time- and 2-dimension-dependent characteristics. In the data for 47 cardiac operation patients, an extreme body mass index (BMI) and significantly increased pressure after 2 h are the top 2 factors associated with the occurrence of pressure injury. This methodology can be used to prevent high interfacial pressure in high-risk patients before and during operation. It can be suggested that this system, integrated with air mattresses, can improve the quality of care and reduce the burden of the workforce and medical cost, especially for pressure injury.

show abstract

Section: Introductionmentioning

confidence: 99%

Flexible Textile-Based Pressure Sensing System Applied in the Operating Room for Pressure Injury Monitoring of Cardiac Operation Patients

Shih¹,

Wang

Chao

et al. 2020

Sensors

View full text Add to dashboard Cite

show abstract

“…Meanwhile, to keep the correctness of the implementation of the BP-KF, we no longer adjust it but use the original architecture form. 26 Similarly, the S-G filter is also optimized on the simulated training set to determine the optimal window size and the order of the fitting polynomial. 23 In addition, all filters are evaluated and verified on a simulated test set and experimental set.…”

Section: T H Imentioning

confidence: 99%

“…The filtering algorithms often used by researchers include Savitzky−Golay (SG) filtering algorithm, 22,23 Kalman filtering (KF) algorithm, 24,25 and the dual-optimized back-propagation adaptive Kalman filtering (BP-KF) algorithm. 26,27 In addition, in chemometrics, the orthogonal signal correction (OSC) and its variants are also used as preprocessing tools to remove systematic noise such as baseline variation and multiplicative scatter effects, in particular, the direct orthogonal signal correction (DOSC) is widely used. 28 The core idea of SG filtering is to filter the data in the window by weighting.…”

Section: Introductionmentioning

confidence: 99%

“…29 Although BP-KF is designed to correct the erratic state values in nonlinear systems through neural networks, it is still challenging to evaluate the accurate noise covariance in dynamic detection, and the present model does not provide a good solution. 26 The basic idea of OSC is to filter out the information not related to the property Y to be measured in the original spectral matrix X by means of mathematical orthogonality, but the extremely low signal-tonoise ratio (SNR) or nonorthogonal noise will lead to degradation of the filtering effect. On the other hand, the capability of artificial neural networks to expressively represent complex data inspired us to readdress its application in solving filtering problems.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adaptively Optimized Gas Analysis Model with Deep Learning for Near-Infrared Methane Sensors

et al. 2022

View full text Add to dashboard Cite

Noise significantly limits the accuracy and stability of retrieving gas concentration with the traditional direct absorption spectroscopy (DAS). Here, we developed an adaptively optimized gas analysis model (AOGAM) composed of a neural sequence filter (NSF) and a neural concentration retriever (NCR) based on deep learning algorithms for extraction of methane absorption information from the noisy transmission spectra and obtaining the corresponding concentrations from the denoised spectra. The model was trained on two data sets, including a computationally generated one and the experimental one. We have applied this model for retrieving methane concentration from its transmission spectra in the near-infrared (NIR) region. The NSF was implemented through an encoder–decoder structure enhanced by the attention mechanism, improving robustness under noisy conditions. Further, the NCR was employed based on a combination of a principal component analysis (PCA) layer, which focuses the algorithm on the most significant spectral components, and a fully connected layer for solving the nonlinear inversion problem of the determination of methane concentration from the denoised spectra without manual computation. Evaluation results show that the proposed NSF outperforms widely used digital filters as well as the state-of-the-art filtering algorithms, improving the signal-to-noise ratio by 7.3 dB, and the concentrations determined with the NCR are more accurate than those determined with the traditional DAS method. With the AOGAM enhancement, the optimized methane sensor features precision and stability in real-time measurements and achieves the minimum detectable column density of 1.40 ppm·m (1σ). The promising results of the present study demonstrate that the combination of deep learning and absorption spectroscopy provides a more effective, accurate, and stable solution for a gas monitoring system.

show abstract

“…Common MLFFNNs include perceptron networks, BP networks, and RBF networks. The BP network used in this paper is a typical application of back-propagation learning algorithm in MLFFNN [41]. The MLFFNN model is mainly includes of the input layer, the hidden layer, and an output layer.…”

Section: Case Studymentioning

confidence: 99%

Time-Series Prediction of Iron and Silicon Content in Aluminium Electrolysis Based on Machine Learning

Chen

Liu

et al. 2021

IEEE Access

View full text Add to dashboard Cite

An adaptive Kalman filtering algorithm based on back-propagation (BP) neural network applied for simultaneously detection of exhaled CO and N2O

Cited by 38 publications

References 27 publications

Flexible Textile-Based Pressure Sensing System Applied in the Operating Room for Pressure Injury Monitoring of Cardiac Operation Patients

Flexible Textile-Based Pressure Sensing System Applied in the Operating Room for Pressure Injury Monitoring of Cardiac Operation Patients

Adaptively Optimized Gas Analysis Model with Deep Learning for Near-Infrared Methane Sensors

Time-Series Prediction of Iron and Silicon Content in Aluminium Electrolysis Based on Machine Learning

Contact Info

Product

Resources

About