Janusz Smulko

Investigation of noise-induced instabilities in quantitative biological spectroscopy and its implications for noninvasive glucose monitoring.

Over the past decade, optical spectroscopy has been employed in combination with multivariate chemometric models to investigate a wide variety of diseases and pathological conditions, primarily due to its excellent chemical specificity and lack of sample preparation requirements. Despite promising results in several proof-of-concept studies, its translation to the clinical setting has often been hindered by inadequate accuracy of the conventional spectroscopic models. To address this issue and the possibility of curved (nonlinear) effects in the relationship between the concentrations of the analyte of interest and the mixture spectra (due to fluctuations in sample and environmental conditions), support vector machine-based least-squares nonlinear regression (LS-SVR) has been recently proposed. In this paper, we investigate the robustness of this methodology to noise-induced instabilities and present an analytical formula for estimating modeling precision as a function of measurement noise and model parameters. This formalism can be readily used to evaluate uncertainty in information extracted from spectroscopic measurements, particularly important for rapid-acquisition biomedical applications. Subsequently, using field data (Raman spectra) acquired from a glucose clamping study on an animal model subject, we perform the first systematic investigation of the relative effect of additive interference components (namely, noise in prediction spectra, calibration spectra, and calibration concentrations) on the prediction error of nonlinear spectroscopic models. Our results show that the LS-SVR method gives more accurate results and is substantially more robust to additive noise when compared with conventional regression methods such as partial least-squares regression (PLS), when careful selection of the LS-SVR model parameters are performed. We anticipate that these results will be useful for uncertainty estimation in similar biomedical applications where the precision of measurements and its response to noise in the data set is as important, if not more so, than the generic accuracy level.

Detecting harmful gases using fluctuation-enhanced sensing with Taguchi sensors

Sensing techniques are often required to not only be versatile and portable, but also to be able to enhance sensor information. This paper describes and demonstrates a new approach to chemical signal analysis that we call fluctuation-enhanced sensing. It utilizes the entire bandwidth of the sensor signal in contrast to more conventional approaches that rely on the dc response. The new principle holds prospects for significantly reducing the necessary number of sensors in artificial noses and tongues, and it can provide improved sensitivity.

Advanced Agent Identification With Fluctuation-Enhanced Sensing

Conventional agent sensing methods normally use the steady state sensor values for agent classification. Many sensing elements (Hines , 1999; Ryan, 2004; Young, 2003;Qian, 2004; Qian, 2006; Carmel, 2003) are needed in order to correctly classify multiple agents in mixtures. Fluctuation-enhanced sensing (FES) looks beyond the steady-state values and extracts agent information from spectra and bispectra. As a result, it is possible to use a single sensor to perform multiple agent classification. This paper summarizes the application of some advanced algorithms that can classify and estimate concentrations of different chemical agents. Our tool involves two steps. First, spectral and bispectral features will be extracted from the sensor signals. The features contain unique agent characteristics. Second, the features are fed into a hyperspectral signal processing algorithm for agent classification and concentration estimation. The basic idea here is to use the spectral/bispectral shape information to perform agent classification. Extensive simulations have been performed by using simulated nanosensor data, as well as actual experimental data using commercial sensor (Taguchi). It was observed that our algorithms are able to accurately classify different agents, and also can estimate the concentration of the agents. Bispectra contain more information than spectra at the expense of high-computational costs. Specific nanostructured sensor model data yielded excellent performance because the agent responses are additive with this type of sensor. Moreover, for measured conventional sensor outputs, our algorithms also showed reasonable performance in terms of agent classification.

ON THE STATISTICAL ANALYSIS OF NOISE IN CHEMICAL SENSORS AND ITS APPLICATION FOR SENSING

Resistance noise data from a single gas sensor can be utilized to identify gas mixtures. We calculated the power spectral density, higher order probability densities and the bispectrum function of the recorded noise samples; these functions are sen sitive to different natural vapors and can be employed to select a proper detection criterion for gas composites and odors.

HAZARDOUS GASES DETECTION BY FLUCTUATION-ENHANCED GAS SENSING

Gas sensing can be enhanced by observing resistance fluctuations in Taguchi Gas Sensors. Unfortunately, fluctuation measurements need advanced measurement setups. This paper presents a possibility of fluctuation-enhanced gas sensing that can be utilized in a simplified and cheap gas sensing system. A detailed study has been conducted for two hazardous gases: ammonia (NH3) and hydrogen sulfide (H2S) using TGS 826 and TGS 825 resistance sensors. The research was focused on practical application of gas detection by resistance fluctuations for the selected harmful gases. Repeatability of noise measurements was investigated to establish rules for most reliable gas detection by a single gas sensor.

Pitting corrosion in steel and electrochemical noise intensity

Electrochemical noise data in the presence of pitting corrosion were analyzed. A correlation between the intensity of the observed noise and mass loss of steel electrodes was recognized. The registered noise was decomposed into a set of band limited components using wavelet transform. It has been observed that the standard deviation of the chosen component was more strictly correlated with mass loss of electrodes than the standard deviation of the other components. The frequency band of the chosen component was adequate to the band where energy of transients, typical for pitting corrosion dominated. The measurement results were obtained only for the limited number of electrodes due to a very long time of noise observation.

Nonlinearity of electrochemical noise caused by pitting corrosion

Measurements of statistical quantities other than spectrum or amplitude distribution can provide additional information about mechanisms of electrochemical noise generation. The third-order cumulant and bispectrum are used to determine whether the nonlinear components exist in the observed noise. The electrochemical current noise, recorded under conditions when pitting corrosion dominated, is analyzed. Transients, characteristic for metastable pitting, are observed. The recognized nonlinear components in the recorded current fluctuations characterize processes of the electrode film repassivation during metastable pitting. It is suggested that the intensity and form of the nonlinear components are related to the intensity of metal dissolution into the electrolyte.

Detection of random transients caused by pitting corrosion

The corrosion processes can be estimated by measurements of electrochemical noise. Noise can be observed as current and voltage fluctuations in a three-electrode setup. The presence of pitting corrosion is manifested as transients in current and voltage fluctuations. For an estimation of the pitting corrosion presence in the recorded current fluctuations, the detection of the characteristic transients has been performed. For the determination of transient occurrence, a detection algorithm of the local changes in a spectrogram of the current fluctuations has been used. The spectrogram has been derived using short-time Fourier transform (STFT). The detected transients, characteristic of pitting corrosion, have been filtered. A time-varying filtering has been performed using an iterative algorithm, presented in the paper.

New approaches for improving selectivity and sensitivity of resistive gas sensors: a review

Purpose – This paper aims to present the methods of improving selectivity and sensitivity of resistance gas sensors. Design/methodology/approach – This paper compares various methods of improving gas sensing by temperature modulation, UV irradiation or fluctuation-enhanced sensing. The authors analyze low-frequency resistance fluctuations in commercial Taguchi gas sensors and the recently developed tungsten trioxide (WO3) gas-sensing layers, exhibiting a photo-catalytic effect. Findings – The efficiency of using low-frequency fluctuations to improve gas detection selectivity and sensitivity was confirmed by numerous experimental studies in commercial and prototype gas sensors. Research limitations/implications – A more advanced measurement setup is required to record noise data but it will reduce the number of gas sensors necessary for identifying the investigated gas mixtures. Practical implications – Fluctuation-enhanced sensing can reduce the energy consumption of gas detection systems and assures bett...

METHODS OF ELECTROCHEMICAL NOISE ANALYSIS FOR INVESTIGATION OF CORROSION PROCESSES

Electrochemical corrosion processes can be investigated by observation of charge flows between the electrolyte and the corroding metal. Usually, the charge flows are observed as spontaneous current and voltage fluctuations (electrochemical noise) in a three-electrode setup. Different types of corrosion processes can be recognized by electrochemical noise analysis. Uniform corrosion rate can be evaluated by estimation of polarization resistance between the metal and electrolyte. Local corrosion events (breakdowns of the passive layer) that produce characteristic transients observed in noise can be detected as well. Different methods of electrochemical noise analysis are presented in a brief review. The limitations and advantages of the proposed methods for corrosion monitoring and research are underlined. The experimental results are also discussed.