Calibrating chemical multisensory devices for real world applications: An in-depth comparison of quantitative machine learning approaches

Abstract

Chemical multisensor devices need calibration algorithms to estimate gas concentrations. Their possible adoption as indicative air quality measurements devices poses new challenges due to the need to operate in continuous monitoring modes in uncontrolled environments. Several issues, including slow dynamics, continue to affect their real world performances. At the same time, the need for estimating pollutant concentrations on board the devices, especially for wearables and IoT deployments, is becoming highly desirable. In this framework, several calibration approaches have been proposed and tested on a variety of proprietary devices and datasets; still, no thorough comparison is available to researchers. This work attempts a benchmarking of the most promising calibration algorithms according to recent literature with a focus on machine learning approaches. We test the techniques against absolute and dynamic performances, generalization capabilities and computational/storage needs using three different datasets sharing continuous monitoring operation methodology. Our results can guide researchers and engineers in the choice of optimal strategy. They show that non-linear multivariate techniques yield reproducible results, outperforming linear approaches. Specifically, the Support Vector Regression method consistently shows good performances in all the considered scenarios. We highlight the enhanced suitability of shallow neural networks in a trade-off between performance and computational/storage needs. We confirm, on a much wider basis, the advantages of dynamic approaches with respect to static ones that only rely on instantaneous sensor array response. The latter have been shown to be best choice whenever prompt and precise response is needed.

Review of the Performance of Low-Cost Sensors for Air Quality Monitoring

“…In the majority of cases, these tested LCS consisted of electrochemical sensors. Only for NO 2 did we observe that supervised learning techniques (ANN, RF, SVM) performed slightly better than MLRs when looking at the R 2 of comparison tests in the field, except for SVR, which is in slight contradiction with other studies [101]. However, the number of records is much higher for MLR than for supervised learning techniques.…”

“…The use of supervised learning techniques (ANN, RF, or SVR) either did not improve performance for CO or gave similar results as MLR for NO. This is in slight contradiction with other studies on the performance of supervised techniques [100,101]. In the majority of cases, these tested LCS consisted of electrochemical sensors.…”

Evaluation and calibration of a low-cost particle sensor in ambient conditions using machine-learning methods

“…In addition, the study concluded that the performances of the low-cost sensors were significantly impacted by temperature and relative humidity (RH). Recurrent NN architectures were also tested for calibrating some gas sensors (De Vito et al, 2018;Esposito et al, 2016). The results showed that the dynamic approaches performed better than traditional static calibration approaches.…”

A Comparative Study of Calibration Methods for Low-Cost Ozone Sensors in IoT Platforms

“…Whenever the sensor response has a linear behavior with respect to the reference data, multiple linear regression (MLR) [11], [12], [14], [19] is used for calibrating the sensors. Nevertheless, when the response is non-linear, models such as K-nearest neighbors (KNN) [19], [20], Gaussian processes [20] and more recently supportvector regression (SVR) [14], [15], [21], random forest (RF) [13], [14], [20] and artificial neural networks (ANN) [11], [15], [20] have been used to calibrate low-cost sensors. Most of these works are focused on studying and analyzing the quality of different commercial low-cost sensors and the performance of electro-chemical sensors.…”

“…For this reason, much research has focused on the interaction of environmental conditions such as temperature and relative humidity [3], [4], [7], [8], [9] or on the interactions of other pollutants [10], [11] with respect to one pollutant sensor. In addition, there is recently a greater interest in comparing and studying [11], [12], [13], [14], [15] how signal processing techniques behave for calibrating different air pollution lowcost sensors in IoT platforms. Many of these investigations focus on comparing what is the error obtained using several linear and non-linear machine learning algorithms.…”