Dual Oxygen and Temperature Luminescence Learning Sensor with Parallel Inference

Abstract

A well-known approach to the optical measure of oxygen is based on the quenching of luminescence by molecular oxygen. The main challenge for this measuring method is the determination of an accurate mathematical model for the sensor response. The reason is the dependence of the sensor signal from multiple parameters (like oxygen concentration and temperature), which are cross interfering in a sensor-specific way. The common solution is to measure the different parameters separately, for example, with different sensors. Then, an approximate model is developed where these effects are parametrized ad hoc. In this work, we describe a new approach for the development of a learning sensor with parallel inference that overcomes all these difficulties. With this approach we show how to generate automatically and autonomously a very large dataset of measurements and how to use it for the training of the proposed neural-network-based signal processing. Furthermore, we demonstrate how the sensor exploits the cross-sensitivity of multiple parameters to extract them from a single set of optical measurements without any a priori mathematical model with unprecedented accuracy. Finally, we propose a completely new metric to characterize the performance of neural-network-based sensors, the Error Limited Accuracy. In general, the methods described here are not limited to oxygen and temperature sensing. They can be similarly applied for the sensing with multiple luminophores, whenever the underlying mathematical model is not known or too complex.

New Autonomous Intelligent Sensor Design Approach for Multiple Parameter Inference

“…In this paper the authors have described a new generation of sensors that does not need any a-priori mathematical model. In [11,29] they also describe an AIS prototype that can extract multiple physical quantities (oxygen concentration and temperatures) at the same time from one single measurement. This prototype can extract two physical quantities without knowing in advance how they are related to each other.…”

“…When colliding with molecular oxygen, the energy of the excited luminophore is reduced due to radiationless deactivation [27]. The AIS built by the authors enables accurate dual-sensing by measuring one single quantity using Multi Task Learning (MTL) neural networks [28] that can predict both oxygen concentration and temperature [11,29]. MTL architectures have been used because they can learn correlated tasks [30][31][32][33][34][35].…”

“…The work to build the first prototype of an AIS, as reported in the two main author's papers [11,29], demonstrate for the first time that a new generation of sensors (AISs) are possible and can satisfy most of the needs of a future generation of sensors.…”

“…Machine learning algorithms can be used to let the sensor find the necessary mathematical formulas autonomously. Neural networks, as the authors proved in one particular case [11,29], are an extremely efficient class of algorithms in this regard.…”

Implementation of multi-task learning neural network architectures for robust industrial optical sensing

“…The automated acquisition procedure is explained in detail in a previous work. 2 The availability of a large set of experimental data was an essential prerequisite for the training and test of the ANN. Additionally, it allowed to test the the classical approach based on an analytical model when applied to measurements under varying conditions of oxygen concentration, temperature, and modulation frequency, as described in Section 3.1.…”