Machine Learning Methods of Regression for Plasmonic Nanoantenna Glucose Sensing

The measurement and quantification of glucose concentrations is a field of major interest, whether motivated by potential clinical applications or as a prime example of biosensing in basic research. In recent years, optical sensing methods have emerged as promising glucose measurement techniques in the literature, with surface-enhanced infrared absorption (SEIRA) spectroscopy combining the sensitivity of plasmonic systems and the specificity of standard infrared spectroscopy. The challenge addressed in this paper is to determine the best method to estimate the glucose concentration in aqueous solutions in the presence of fructose from the measured reflectance spectra. This is referred to as the inverse problem of sensing and usually solved via linear regression. Here, instead, several advanced machine learning regression algorithms are proposed and compared, while the sensor data are subject to a pre-processing routine aiming to isolate key patterns from which to extract the relevant information. The most accurate and reliable predictions were finally made by a Gaussian process regression model which improves by more than 60% on previous approaches. Our findings give insight into the applicability of machine learning methods of regression for sensor calibration and explore the limitations of SEIRA glucose sensing.

Managing and Utilizing Big Data in Atmospheric Monitoring Systems for Underground Coal Mines

Underground coal mining Atmospheric Monitoring Systems (AMS) have been implemented for real-time or near real-time monitoring and evaluation of the mine atmosphere and related parameters such as gas concentration (e.g., CH4, CO, O2), fan performance (e.g., power, speed), barometric pressure, ambient temperature, humidity, etc. Depending on the sampling frequency, AMS can collect and manage a tremendous amount of data, which mine operators typically consult for everyday operations as well as long-term planning and more effective management of ventilation systems. The raw data collected by AMS need considerable pre-processing and filtering before they can be used for analysis. This paper discusses different challenges related to filtering raw AMS data in order to identify and remove values due to sensor breakdowns, sensor calibration periods, transient values due to operational considerations, etc., as well as to homogenize time series for different variables. The statistical challenges involve the removal of faulty values and outliers (due to systematic problems) and transient effects, gap-filling (by means of interpolation methods), and homogenization (setting a common time reference and time step) of the respective time series. The objective is to derive representative and synchronous time series values that can subsequently be used to estimate summary statistics of AMS and to infer correlations or nonlinear dependence between different data streams. Identification and modeling of statistical dependencies can be further exploited to develop predictive equations based on time series models.

Cross-Sensor Nighttime Lights Image Calibration for DMSP/OLS and SNPP/VIIRS with Residual U-Net

Remote sensing of nighttime lights (NTL) is widely used in socio-economic studies of economic growth, urbanization, stability of power grid, environmental light pollution, pandemics and military conflicts. Currently, NTL data are collected with two sensors: (1) Operational Line-scan System (OLS) onboard the satellites from the Defense Meteorology Satellite Program (DMSP) and (2) Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the Suomi NPP (SNPP) and NOAA-20 satellites from the Joint Polar Satellite System (JPSS). However, the nighttime images acquired by these two sensors are incompatible in spatial resolution and dynamic range. To address this problem, we propose a method for the cross-sensor calibration with residual U-net convolutional neural network (CNN). The CNN produces DMSP-like NTL composites from the VIIRS annual NTL composites. The pixel radiances predicted from VIIRS are highly correlated with NTL observed with OLS (0.96 < R2 < 0.99). The method can be used to extend long-term series of annual NTL after the end of DMSP mission or to cross-calibrate same year NTL from different satellites to study diurnal variations.

Calibration and Data Quality Assurance Technical Advancements for Quantitative Remote Sensing in the DRAGON 4 Project

Robust calibration and validation (Cal and Val) should guarantee the accuracy of the retrieved information, make the remote sensing data consistent and traceable, and maintain the sensor performance during the operational phase. The DRAGON program has set up many remote sensing research topics on various application domains. In order to promote the effectiveness of data modeling and interpretation, it is necessary to solve various challenges in Cal and Val for quantitative RS applications. This project in the DRAGON 4 program aims to promote the cooperation of the Cal and Val experts from European and Chinese institutes in Cal and Val activities, and several achievements have been obtained in the advanced on-orbit optical sensor calibration, as well as microwave remote sensor calibration and product generation. The outcomes of the project have benefited the related remote sensing modeling and product retrieval, and promoted the radiometric calibration network (RadCalNet) as an international operational network for calibration, intercalibration, and validation. Moreover, this project provided local governments with a more accurate OMI NO2 data in China, which were used to study the air quality control during APEC period, Parade period and G20 period. This will be of ongoing be value for monitoring atmospheric environmental quality and formulating pollution control strategies.

Automated Drilling Variances Detection Through Smart Alarms System

In a drive to enhance drilling operational awareness, the Real-Time Operations Center (RTOC) has developed a State-of-the-Art event detection algorithm that consistently highlights the deviations of critical parameters by actively comparing real-time values against comprehensive physical models and alerting the users through a dashboard. The process relies on different levels of frequency and severity in order to detect events at their onset and prevent developing into a situation that compromises the operations. The first pillar of the solution consists of deterministic modelling of the expected values for a series of parameters in order to provide the basis for comparison and diagnostics. The main parameters sought to be modelled consist of the Standpipe Pressure, the Rotary Torque and the Hook load, which respectively are generated through individual methods taking into consideration actual conditions as well as relevant contextual data to ensure accuracy. The second pillar of the solution consists of visual alerts, triggered and displayed on a dashboard based on frequency and severity levels, as percentage of deviation from accepted operational envelope. The solution has been initially implemented during drilling operations where different issues were expected to take place, finding that whenever such occurrences took place, the algorithms were able to signal potential events in most of the cases. Some challenges were observed mainly due to sensor calibration and behavior since the expected model values not necessarily match reality, including residual pressure when the pumps are off or when the string is set on slips but the hook load values still present some variance. Also, it has been observed during transient periods where flow and rotation are changed drastically, that the stabilization to a steady state present with high variance, which has demanded the introduction of further logics within the algorithms to account for these effects and avoid the generation of false indications of issues. The solution has given encouraging results thus far in signaling different dysfunctions on the drilling process without the need of immediate human interpretation of data, which has allowed to move forward in the digitalization of operations, not only by timely signaling the onset of issues, but as well by providing the basis to further develop real time diagnosis of the problems to accelerate their resolution. The conception of the event detection based on deterministic real time analysis of individual channels against robust physical models from the existing digital twin solution has proven an immediate asset for operations on its own. By providing clear signaling of issues, while providing a solid framework to ultimately develop a diagnostic solution to translate a potential event into a proactive approach to support decision making process.

Magnetic Field-Based Vehicle Positioning System in Long Tunnel Environment

Recently, long tunnels are becoming more prevalent in Korea, and exits are added at certain sections of the tunnels. Thus, a navigation system should correctly guide the user toward the exit; however, adequate guidance is not delivered because the global navigation satellite system (GNSS) signal is not received inside a tunnel. Therefore, we present an accurate position estimation system using a magnetic field for vehicles passing through a tunnel. The position can be accurately estimated using the magnetic sensor of a smartphone with an appropriate attitude estimation and magnetic sensor calibration. Position estimation was realized by attaching the smartphone on the dashboard during navigation and calibrating the sensors using position information from the GNSS and magnetic field database before entering the tunnel. This study used magnetic field sequence data to estimate vehicle positions inside a tunnel. Furthermore, subsequence dynamic time warping was applied to compare the magnetic field data stored in the buffer with the magnetic field database, and the feasibility and performance of the proposed system was reviewed through an experiment in an actual tunnel. The analysis of the position estimation results confirmed that the proposed system could appropriately deliver tunnel navigation.

Galvanic pH sensor for continuous monitoring of soil parameters in agriculture

The article discusses trends in agriculture development in digitalization direction, substantiates need for monitoring pH level throughout year. The main solutions existing on market at moment are given, as well as need to develop a pH sensor for USKD-Agro hardware and software complex. The article discusses in detail the ongoing physical and chemical processes, on them basis pH level is determined, circuitry solutions used, as well as pH sensor implementation. The article describes the sensor calibration process, methodology and test results.