SENSOR SCIENCE

Research Projects and publications

Multi-Task Learning for Multi-Dimensional Regression: Application to Luminescence Sensing

Submitted to Applied Sciences (MDPI) (September 2019)

Author(s): Umberto Michelucci, TOELT LLC (Switzerland), Francesca Venturini, Zürcher Hochschule für Angewandte Wissenschaften (Switzerland) and TOELT LLC (Switzerland);

Share & Cite This Article       

Cite: Michelucci, U.; Venturini, F. Multi-Task Learning for Multi-Dimensional Regression: Application to Luminescence Sensing. Preprints 2019, 2019100009 (doi: 10.20944/preprints201910.0009.v1).

Abstract

Luminescence sensors are based on the determination of emitted intensity or decay time when a luminophore is in contact with its The classical approach to non-linear regression in physics, is to take a mathematical model describing the functional dependence of the dependent variable from a set of independent variables, and then, using non-linear fitting algorithms, extract the parameters used in the modeling. Particularly challenging are real systems, characterised by several additional influencing factors related to specific components, like electronics or optical parts. In such cases, to make the model reproduce the data, empirically determined terms are built-in the models to compensate for the impossibility of modeling things that are, by construction, impossible to model. A new approach to solve this issue is to use neural networks, particularly feed-forward architectures with a sufficient number of hidden layers and an appropriate number of output neurons, each responsible for predicting the desired variables. Unfortunately, feed-forward neural networks (FFNNs) usually perform less efficiently when applied to multi-dimensional regression problems, that is when they are required to predict simultaneously multiple variables that depend from the input dataset in fundamentally different ways. To address this problem, we propose multi-task learning (MTL) architectures. These are characterized by multiple branches of task-specific layers, which have as input the output of a common set of layers. To demonstrate the power of this approach for multi-dimensional regression, the method is applied to luminescence sensing. Here the MTL architecture allows predicting multiple parameters, the oxygen concentration and the temperature, from a single set of measurements.

Optical oxygen sensing with artificial intelligence

SPIE Photonics West Conference Proceedings / Talk

Author(s): Francesca Venturini, Michael Baumgartner, Zürcher Hochschule für Angewandte Wissenschaften (Switzerland); Umberto Michelucci, TOELT LLC (Switzerland)

Share & Cite This Article       

Proc. SPIE 10937, Optical oxygen sensing with artificial intelligence, 1093714 (February 2019).

Abstract

Luminescence sensors are based on the determination of emitted intensity or decay time when a luminophore is in contact with its environment. Typically, since the absolute values of the measured quantities depend on the specifc sensing element and scheme used, a sensor needs an analytical model to describe the dependence of the quantity to be determined, for example the oxygen concentration concentration, from sensed quantity, for example the decay time. Additionally, since the details of this dependence are device specifc, a sensor needs to be calibrated at known reference conditions. This work explores an entirely new artifcial intelligence approach and demonstrates the feasibility of oxygen sensing through machine learning. After training the neural network on synthetic data, it was tested on measured data to verify the prediction of the model. The results show a mean deviation of the predicted from the measured concentration of 0.5 % air, which is comparable to many commercial and low-cost sensors. The accuracy of the model predictions is limited by the ability of the generated data to describe the measured data, opening up future possibilities for signifcant improvement by performing the training on experimental data. In this work the approach is tested at different temperatures, showing its applicability in the entire range relevant for biological applications.

Optical oxygen sensing with artificial intelligence

Published in Sensors (MDPI)

PDF available: https://www.mdpi.com/1424-8220/19/4/777/pdf

Share & Cite This Article       

Michelucci, U.; Baumgartner, M.; Venturini, F. Optical Oxygen Sensing with Artificial Intelligence. Sensors 2019, 19, 777.

Abstract

Luminescence-based sensors for measuring oxygen concentration are widely used both in industry and research due to the practical advantages and sensitivity of this type of sensing. The measuring principle is the luminescence quenching by oxygen molecules, which results in a change of the luminescence decay time and intensity. In the standard approach, this change is related to an oxygen concentration using the Stern–Volmer equation. This equation, which in most of the cases is non-linear, is parametrized through device-specific constants. Therefore, to determine these parameters every sensor needs to be precisely calibrated at one or more known concentrations. This work explores an entirely new artificial intelligence approach and demonstrates the feasibility of oxygen sensing through machine learning. The specifically developed neural network learns very efficiently to relate the input quantities to the oxygen concentration. The results show a mean deviation of the predicted from the measured concentration of 0.5 % air, comparable to many commercial and low-cost sensors. Since the network was trained using synthetically generated data, the accuracy of the model predictions is limited by the ability of the generated data to describe the measured data, opening up future possibilities for significant improvement by using a large number of experimental measurements for training. The approach described in this work demonstrates the applicability of artificial intelligence to sensing technology and paves the road for the next generation of sensors.

Novel Algorithm for Calibration-Free Absorption Spectroscopy Sensor

Umberto Michelucci and Francesca Venturini

Published: 4 December 2017

PDF Full Text Download: https://www.mdpi.com/2504-3900/1/8/833/pdf

Share & Cite This Article       

MDPI and ACS Style

Venturini, F.; Michelucci, U. Novel Algorithm for Calibration-Free Absorption Spectroscopy Sensor. Proceedings20171, 833.

Excerpt

Due to the enormous progress in availability and performance of laser light sources and electro-optical components, tunable laser diode absorption spectroscopy is currently more and more used for quantitative assessments of gas in several fields, such as medical breath analysis, atmospheric environmental monitoring, chemical analysis, industrial process control, and high-resolution molecular spectroscopy. […]

Novel Semi-Parametric Algorithm for Interference-Immune Tunable Absorption Spectroscopy Gas Sensing

Umberto Michelucci and Francesca Venturini

Received: 10 September 2017 / Revised: 1 October 2017 / Accepted: 4 October 2017 / Published: 7 October 2017

PDF Full Text Download: https://www.mdpi.com/1424-8220/17/10/2281/pdf

Share & Cite This Article       

MDPI and ACS Style

Michelucci, U.; Venturini, F. Novel Semi-Parametric Algorithm for Interference-Immune Tunable Absorption Spectroscopy Gas Sensing. Sensors 2017, 17, 2281.

Abstract

One of the most common limits to gas sensor performance is the presence of unwanted interference fringes arising, for example, from multiple reflections between surfaces in the optical path. Additionally, since the amplitude and the frequency of these interferences depend on the distance and alignment of the optical elements, they are affected by temperature changes and mechanical disturbances, giving rise to a drift of the signal. In this work, we present a novel semi-parametric algorithm that allows the extraction of a signal, like the spectroscopic absorption line of a gas molecule, from a background containing arbitrary disturbances, without having to make any assumption on the functional form of these disturbances. The algorithm is applied first to simulated data and then to oxygen absorption measurements in the presence of strong fringes.To the best of the authors’ knowledge, the algorithm enables an unprecedented accuracy particularly if the fringes have a free spectral range and amplitude comparable to those of the signal to be detected. The described method presents the advantage of being based purely on post processing, and to be of extremely straightforward implementation if the functional form of the Fourier transform of the signal is known. Therefore, it has the potential to enable interference-immune absorption spectroscopy. Finally, its relevance goes beyond absorption spectroscopy for gas sensing, since it can be applied to any kind of spectroscopic data.