Improve Calibration Performance Research Articles

Calibration and Inter-Unit Consistency Assessment of an Electrochemical Sensor System Using Machine Learning.

This paper addresses the challenges of calibrating low-cost electrochemical sensor systems for air quality monitoring. The proliferation of pollutants in the atmosphere necessitates efficient monitoring systems, and low-cost sensors offer a promising solution. However, issues such as drift, cross-sensitivity, and inter-unit consistency have raised concerns about their accuracy and reliability. The study explores the following three calibration methods for converting sensor signals to concentration measurements: utilizing manufacturer-provided equations, incorporating machine learning (ML) algorithms, and directly applying ML to voltage signals. Experiments were performed in three urban sites in Greece. High-end instrumentation provided the reference concentrations for training and evaluation of the model. The results reveal that utilizing voltage signals instead of the manufacturer's calibration equations diminishes variability among identical sensors. Moreover, the latter approach enhances calibration efficiency for CO, NO, NO2, and O3 sensors while incorporating voltage signals from all sensors in the ML algorithm, taking advantage of cross-sensitivity to improve calibration performance. The Random Forest ML algorithm is a promising solution for calibrating similar devices for use in urban areas.

Motion-Based Extrinsic Sensor-to-Sensor Calibration: Effect of Reference Frame Selection for New and Existing Methods

This paper studies the effect of reference frame selection in sensor-to-sensor extrinsic calibration when formulated as a motion-based hand-eye calibration problem. As the sensor trajectories typically contain some composition of noise, the aim is to determine which selection strategies work best under which noise conditions. Different reference selection options are tested under varying noise conditions in simulations, and the findings are validated with real data from the KITTI dataset. The study is conducted for four state-of-the-art methods, as well as two proposed cost functions for nonlinear optimization. One of the proposed cost functions incorporates outlier rejection to improve calibration performance and was shown to significantly improve performance in the presence of outliers, and either match or outperform the other algorithms in other noise conditions. However, the performance gain from reference frame selection was deemed larger than that from algorithm selection. In addition, we show that with realistic noise, the reference frame selection method commonly used in the literature, is inferior to other tested options, and that relative error metrics are not reliable for telling which method achieves best calibration performance.

Machine learning models for mitral valve replacement: A comparative analysis with the Society of Thoracic Surgeons risk score.

Current Society of Thoracic Surgeons (STS) risk models for predicting outcomes of mitral valve surgery (MVS) assume a linear and cumulative impact of variables. We evaluated postoperative MVS outcomes and designed mortality and morbidity risk calculators to supplement the STS risk score. Data from the STS Adult Cardiac Surgery Database for MVS was used from 2008 to 2017. The data included 383,550 procedures and 89 variables. Machine learning (ML) algorithms were employed to train models to predict postoperative outcomes for MVS patients. Each model's discrimination and calibration performance were validated using unseen data against the STS risk score. Comprehensive mortality and morbidity risk assessment scores were derived from a training set of 287,662 observations. The area under the curve (AUC) for mortality ranged from 0.77 to 0.83, leading to a 3% increase in predictive accuracy compared to the STS score. Logistic Regression and eXtreme Gradient Boosting achieved the highest AUC for prolonged ventilation (0.82) and deep sternal wound infection (0.78 and 0.77) respectively. EXtreme Gradient Boosting performed the best with an AUC of 0.815 for renal failure. For permanent stroke prediction all models performed similarly with an AUC around 0.67. The ML models led to improved calibration performance for mortality, prolonged ventilation, and renal failure, especially in cases of reconstruction/repair and replacement surgery. The proposed risk models complement existing STS models in predicting mortality, prolonged ventilation, and renal failure, allowing healthcare providers to more accurately assess a patient's risk of morbidity and mortality when undergoing MVS.

Partial least squares regression residual extreme learning machine (PLSRR-ELM) calibration algorithm applied in fast determination of gasoline octane number with near-infrared spectroscopy

Based on near-infrared (NIR) spectroscopy, a new quantitative calibration algorithm, called “Partial Least Squares Regression Residual Extreme Learning Machine (PLSRR-ELM)”, was proposed for fast determination of research octane number (RON) for blended gasoline. In this algorithm, partial least square (PLS) cooperates with non-linear extreme learning machine (ELM) to separate the relationship information suitable for each other from the raw relationship information (between NIR spectrum and corresponding property) with the unknown degree of non-linearity, with aim of calibrating them respectively. Since the advantages of both PLS and ELM are fully utilized, it is expected that PLSRR-ELM can address the relationship information more effectively and leads to improved calibration performance over PLS and ELM alone. The calibration performance of PLSRR-ELM was evaluated by a set of on-line gasoline blending sample data from a refinery. As a result, it showed an enhanced prediction performance, e.g., about 13% or 11% decrease in the root mean squared error of test (RMSE-T) over PLS or ELM alone, respectively. In method comparison, the model performance of PLRR-ELM exceeds all other methods including PLS, Poly-PLS, KPLS, ELM, and ANN, demonstrating its superiority for fast prediction of gasoline RON.

Missing Data Imputation on IoT Sensor Networks: Implications for on-Site Sensor Calibration

IoT sensors are becoming increasingly important supplement to traditional monitoring systems, particularly for in-situ based monitoring. Data collected using IoT sensors are often plagued with missing values occurring as a result of sensor faults, network failures, drifts and other operational issues. Missing data can have substantial impact on in-field sensor calibration methods. The goal of this research is to achieve effective calibration of sensors in the context of such missing data. To this end, two objectives are presented in this paper. 1) Identify and examine effective imputation strategy for missing data in IoT sensors. 2) Determine sensor calibration performance using calibration techniques on data set with imputed values. Specifically, this paper examines the performance of Variational Autoencoder (VAE), Neural Network with Random Weights (NNRW), Multiple Imputation by Chain Equations (MICE), Random Forest-based Imputation (missForest) and K-Nearest Neighbour (KNN) for imputation of missing values on IoT sensors. Furthermore, the performance of sensor calibration via different supervised algorithms trained on the imputed dataset were evaluated. The analysis showed VAE technique to outperform the other methods in imputing the missing values at different proportions of missingness on two real-world datasets. Experimental results also showed improved calibration performance with imputed dataset.

Polarimetric calibration of SAR images using reflection symmetric targets with low helix scattering

Polarimetric synthetic aperture radar is an important technology for Earth observation. Generally, polarimetric calibration (PolCAL) should be performed before any quantitative applications. The reflection symmetric targets can relieve the dependence of PolCAL on calibrators. In this work, we use a helix-ratio feature to extract these targets. Through analysis, we find that the helix-ratio feature has a high tolerance to polarimetric distortions and is suitable for extracting samples in un-calibrated images for PolCAL. An adaptive segmentation algorithm is also developed to extract reflection symmetric targets automatically. Experiments with C- and X-band airborne images show that the extraction of reflection symmetric targets can improve the calibration accuracy of crosstalk. In addition, the proposed method can maintain the polarimetric signature of the targets in the calibration and can improve calibration performance when compared with other reflection symmetric target extraction methods.

Milk infrared spectra from multiple instruments improve performance of prediction models

Milk data predicted from infrared spectra are the basis of selection programmes for dairy species, but several factors influence the robustness of prediction models, e.g., (i) variability of collected samples, (ii) number and quality of reference data, and (iii) influence traits on milk spectral fingerprint. Use of single or multiple instruments spectra from the same set of samples to develop mid infrared calibration models was investigated. Fitting statistics for milk coagulation properties were low to fair. Better results were achieved for titratable acidity and cheese yield; coefficients of determination in cross validation ranged from 0.52 to 0.77, depending on the trait analysed and training set. Merging spectral data from two instruments did not improve calibration performance but provided the best fitting statistics in external validation for most of the traits on different validation sets. Merging spectral data from different instruments is a cost-effective method to improve calibration performance and robustness.

Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis.

Pre-eclampsia is a leading cause of maternal and perinatal mortality and morbidity. Early identification of women at risk is needed to plan management. To assess the performance of existing pre-eclampsia prediction models and to develop and validate models for pre-eclampsia using individual participant data meta-analysis. We also estimated the prognostic value of individual markers. This was an individual participant data meta-analysis of cohort studies. Source data from secondary and tertiary care. We identified predictors from systematic reviews, and prioritised for importance in an international survey. Early-onset (delivery at < 34 weeks' gestation), late-onset (delivery at ≥ 34 weeks' gestation) and any-onset pre-eclampsia. We externally validated existing prediction models in UK cohorts and reported their performance in terms of discrimination and calibration. We developed and validated 12 new models based on clinical characteristics, clinical characteristics and biochemical markers, and clinical characteristics and ultrasound markers in the first and second trimesters. We summarised the data set-specific performance of each model using a random-effects meta-analysis. Discrimination was considered promising for C-statistics of ≥ 0.7, and calibration was considered good if the slope was near 1 and calibration-in-the-large was near 0. Heterogeneity was quantified using I2 and τ2. A decision curve analysis was undertaken to determine the clinical utility (net benefit) of the models. We reported the unadjusted prognostic value of individual predictors for pre-eclampsia as odds ratios with 95% confidence and prediction intervals. The International Prediction of Pregnancy Complications network comprised 78 studies (3,570,993 singleton pregnancies) identified from systematic reviews of tests to predict pre-eclampsia. Twenty-four of the 131 published prediction models could be validated in 11 UK cohorts. Summary C-statistics were between 0.6 and 0.7 for most models, and calibration was generally poor owing to large between-study heterogeneity, suggesting model overfitting. The clinical utility of the models varied between showing net harm to showing minimal or no net benefit. The average discrimination for IPPIC models ranged between 0.68 and 0.83. This was highest for the second-trimester clinical characteristics and biochemical markers model to predict early-onset pre-eclampsia, and lowest for the first-trimester clinical characteristics models to predict any pre-eclampsia. Calibration performance was heterogeneous across studies. Net benefit was observed for International Prediction of Pregnancy Complications first and second-trimester clinical characteristics and clinical characteristics and biochemical markers models predicting any pre-eclampsia, when validated in singleton nulliparous women managed in the UK NHS. History of hypertension, parity, smoking, mode of conception, placental growth factor and uterine artery pulsatility index had the strongest unadjusted associations with pre-eclampsia. Variations in study population characteristics, type of predictors reported, too few events in some validation cohorts and the type of measurements contributed to heterogeneity in performance of the International Prediction of Pregnancy Complications models. Some published models were not validated because model predictors were unavailable in the individual participant data. For models that could be validated, predictive performance was generally poor across data sets. Although the International Prediction of Pregnancy Complications models show good predictive performance on average, and in the singleton nulliparous population, heterogeneity in calibration performance is likely across settings. Recalibration of model parameters within populations may improve calibration performance. Additional strong predictors need to be identified to improve model performance and consistency. Validation, including examination of calibration heterogeneity, is required for the models we could not validate. This study is registered as PROSPERO CRD42015029349. This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 24, No. 72. See the NIHR Journals Library website for further project information.

Challenges Calibrating Hydrology for Groundwater-Fed Wetlands: a Headwater Wetland Case Study.

This study aims to adapt the Soil and Watershed Assessment Tool (SWAT), a ubiquitously used watershed model, for ground-water dominated surface waterbodies by accounting for recharge from the aquifers. Using measured flow to a headwater slope wetland in Alabama's coastal plain region as a case study, we present challenges and relatively simple approaches in using the SWAT model to predict flows from the draining watershed and relatively simple approaches to model groundwater upwelling. SWAT-simulated flow at the study watershed was limited by precipitation, and consequently, simulated flows were several times smaller in magnitude than observed flows. Thus, our first approach involved a separate stormflow and baseflow calibration which included the use of a regression relationship between observed and simulated baseflow (E NASH = 0.67). Our next approach involved adapting SWAT to simulate upwelling groundwater discharge instead of deep aquifer losses by constraining the range of deep losses, β deep parameter, to negative values (E NASH = 0.75). Finally, we also investigated the use of artificial neural networks (ANN) in conjunction with SWAT to further improve calibration performance. This approach used SWAT-calibrated flow, evapotranspiration, and precipitation as inputs to ANN (E NASH = 0.88). The methods investigated in this study can be used to navigate similar flow calibration challenges in other groundwater dominant watersheds which can be very useful tool for managers and modelers alike.

A meta-model-based optimization approach for fast and reliable calibration of building energy models

Building energy model calibration with optimization aims to bridge the gap between simulated energy consumption and measurement, thus aiding building retrofit and operation. However, the difficulty of the optimization in calibration including both optimization hyperparameter settings and problem complexity (multi-modal and under-determined) make the calibration with optimization approach difficult to be applied in practice with full reliability. Meanwhile, current calibration with optimization treats building calibration as a purely mathematical problem while neglecting the importance of engineering judgment in the calibration practice. In this paper, we introduced meta-models into the calibration with optimization approach with an auto-correction mechanism to improve calibration performance with respect to time and reliability. To better illustrate the approach, we presented a case study with validation. The proposed method was demonstrated to alleviate difficulty of optimization while improving calibration time and reliability in the study. Comparing two types of meta-models, we found that using the GP (Gaussian Process) achieved better performance with less computation time and higher accuracy compared to the MLR (Multiple Linear Regression). To efficiently train emulators, we can start with generating only a small amount of white-box simulation results. It is also important to generate enough initial starts to ensure robustness of calibration.

Non-kinematic calibration of industrial robots using a rigid–flexible coupling error model and a full pose measurement method

Abstract This study addresses the non-negligible flexibility problem of industrial robots and describes a rigid–flexible coupling error model for robot non-kinematic calibration. Simulation confirms that compliance errors are major sources of inaccuracies and affect robot accuracy with geometric errors. To reduce the impact of measurement noise and improve calibration performance, an enhanced approach is proposed for full pose measurement and identification optimisation. This approach is based on the self-adaption particle swarm optimisation (SAPSO) algorithm with a dual-index (i.e. observability index O1 and identification accuracy index Omin) and a modified Levenberg–Marquardt algorithm, which considers external constraints (i.e. structural interference and angular limitation). Simulation results illustrate that the 36 error parameters of the proposed error model can be identified with improved accuracy and stability using the proposed approach. Finally, experimental results obtained with a Staubli TX60L robot with a FARO laser tracker are introduced to demonstrate the practical effectiveness of our approach in robot accuracy improvement.

New insights to improve resolution and reliability of Raman spectral analysis using higher-density multiscale regression

When paired with sophisticated multivariate calibration methods, spectroscopy techniques stand out conspicuously for high-throughput analysis of complex samples. However, despite the identification capability of multivariate analysis, fluctuant spectral interference and complicated chemometrics procedure often combine to limit conventional spectroscopy techniques. In this regard, we proposed a novel strategy, higher-density multiscale regression (HDMR), to adaptively process Raman spectra for quantitative analysis. HDMR firstly splits the Raman spectra into frequency components at different scales using higher-density discrete wavelet transform (HDDWT), yielding a larger number of subbands to benefit a good separation of spectral bands. Parallel member models constructed with these decomposed components are then fused into a final prediction through partial least square (PLS) reweighting strategy. With HDMR, the pre-treatment process and multivariate calibration are integrated into a unified calibration model for Raman spectra at hand, regardless of its structural characteristics. This would definitely avoid of information leakage in multivariate calibration, thus promoting the application of Raman spectroscopy technique as a general tool for analytical chemistry. To validate the performance of the HDMR, two industrial Raman spectral data sets were investigated to yield challenges representative of those encountered in Raman spectroscopy. This proposed strategy has improved calibration performance through the reweighting way. Satisfactory calibration results suggest that HDMR provides a universal tool for reliable Raman spectral analysis, which may well extend to other spectral data sets.

Assessment of Interval PLS (iPLS) Calibration for the Determination of Peroxide Value in Edible Oils

AbstractPartial least squares (PLS) is a commonly used (and sometimes misused) chemometric technique for calibrating Fourier transform infrared spectroscopy, and allows the analysis of a variety of quality parameters associated with edible oils. Peroxide value (PV) is a typical parameter of interest; however, developing a robust, optimal, and reliable calibration method can be a daunting task. This paper examines and compares the use of interval PLS as a tool to develop a PLS PV calibration method for a single‐bounce attenuated total reflectance accessory relative to full spectrum PLS and experienced PLS, making use of correlation, variance, and pure component spectra. Using mixtures of fresh and oxidized oil covering a PV range of 1–20 meq/kg, backward interval PLS could systematically produce quality calibrations without the need to resort to experienced PLS. The experienced PLS requires a degree of spectral knowledge as well as diligent and tedious spectral examination, including largely unstructured iterative calibrations and cross‐validations to improve calibration performance. The backward interval PLS is also better than the full spectrum PLS in terms of model performance. In addition, the general model developed could account for the errors caused by oil types.

A New Calibration Method of Three Axis Magnetometer With Nonlinearity Suppression

Nonlinearity is a prominent limitation to the calibration performance of vector magnetometers. A new calibration model is proposed to suppress the nonlinearity of magnetometers and improve calibration performance, in which the nonlinearity coefficients of scale factors are considered. The experimental system mainly consists of a three-axis fluxgate magnetometer (MAG3300), a 2D nonmagnetic rotation equipment, and a proton magnetometer (CZM-3), in which the nonmagnetic rotation equipment is used to change the position of three-axis fluxgate magnetometer, and the scalar value of magnetic field is obtained with the proton magnetometer and considered to be the true value. The principle of this new calibration method is analyzed, and the calibration procedures are introduced. Experimental results show that after nonlinearity suppression, the root-mean-square error of calibration can be reduced by three times. It suggests an effective way to improve the calibration performance of three-axis magnetometers.

Full Correction of Scattering Effects by Using the Radiative Transfer Theory for Improved Quantitative Analysis of Absorbing Species in Suspensions

Sample-to-sample photon path length variations that arise due to multiple scattering can be removed by decoupling absorption and scattering effects by using the radiative transfer theory, with a suitable set of measurements. For samples where particles both scatter and absorb light, the extracted bulk absorption spectrum is not completely free from nonlinear particle effects, since it is related to the absorption cross-section of particles that changes nonlinearly with particle size and shape. For the quantitative analysis of absorbing-only (i.e., nonscattering) species present in a matrix that contains a particulate species that absorbs and scatters light, a method to eliminate particle effects completely is proposed here, which utilizes the particle size information contained in the bulk scattering coefficient extracted by using the Mie theory to carry out an additional correction step to remove particle effects from bulk absorption spectra. This should result in spectra that are equivalent to spectra collected with only the liquid species in the mixture. Such an approach has the potential to significantly reduce the number of calibration samples as well as improve calibration performance. The proposed method was tested with both simulated and experimental data from a four-component model system.

Calibration of Infrared Milk Analyzers: Modified Milk Versus Producer Milk

Mid-infrared (MIR) milk analyzers are traditionally calibrated using sets of preserved raw individual producer milk samples. The goal of this study was to determine if the use of sets of preserved pasteurized modified milks improved calibration performance of MIR milk analyzers compared with calibration sets of producer milks. The preserved pasteurized modified milk sets exhibited more consistent day-to-day and set-to-set calibration slope and intercept values for all components compared with the preserved raw producer milk calibration sets. Pasteurized modified milk calibration samples achieved smaller confidence interval (CI) around the regression line (i.e., calibration uncertainty). Use of modified milk calibration sets with a larger component range, more even distribution of component concentrations within the ranges, and the lower correlation of fat and protein concentrations than producer milk calibration sets produced a smaller 95% CI for the regression line due to the elimination of moderate and high leverage samples. The CI for the producer calibration sets were about 2 to 12 times greater than the CI for the modified milk calibration sets, depending on the component. Modified milk calibration samples have the potential to produce MIR milk analyzer calibrations that will perform better in validation checks than producer milk-based calibrations by reducing the mean difference and standard deviation of the difference between instrument values and reference chemistry.

Comparison of near-infrared and mid-infrared spectroscopy for the determination of distillation property of kerosene

Near-infrared (NIR) and mid-infrared (MIR) spectroscopy have been compared and evaluated for the determination of the distillation property of kerosene with the use of partial least squares (PLS) regression. Since kerosene is a complex mixture of similar hydrocarbons, both spectroscopic methods will be best evaluated with this complex sample matrix. PLS calibration models for each percent recovery temperature have been developed by using both NIR and MIR spectra without spectral pretreatment. Both methods have shown good correlation with the corresponding reference method, however NIR provided better calibration performance over MIR. To rationalize the improved calibration performance of NIR, spectra of the same kerosene sample were continuously collected and the corresponding spectral reproducibility was evaluated. The greater spectral reproducibility including signal-to-noise ratio of NIR led to the improved calibration performance, even though MIR spectroscopy provided more qualitative spectral information. The reproducibility of measurement, signal-to-noise ratio, and richness of qualitative information should be simultaneously considered for proper selection of a spectroscopic method for quantitative analysis.

Comparison of Near-Infrared and Raman Spectroscopy for the Determination of Chemical and Physical Properties of Naphtha

Near-infrared (NIR) and Fourier transform (FT) Raman spectroscopy have been compared and evaluated for the quantitative analysis of naphtha with the use of partial least-squares (PLS) regression. Naphtha has been chosen for this purpose since it is a complex mixture of similar hydrocarbons; therefore, both spectroscopic methods can be evaluated with a complex sample of different concentration ranges and also a physical property. Six different chemical compositions [total paraffin, total naphthene (cycloalkane), total aromatic, C6 paraffin, benzene, and cyclopentane] and specific gravity (as a physical parameter) have been selected to evaluate both spectroscopic methods. PLS calibration models for each property have been developed by using both NIR and Raman spectra without spectral pretreatment. Both methods showed good correlation with the corresponding reference methods, but NIR provided the better calibration performance over Raman. The superior signal-to-noise ratio as well as spectral reproducibility of NIR led to the improved calibration performance even though Raman spectroscopy provided richer spectral information. The signal-to-noise ratio, reproducibility of measurement, and richness of spectral information should be simultaneously considered for proper selection of a spectroscopic method, especially for quantitative analysis.