Laplacian Scores-Based Feature Reduction in IoT Systems for Agricultural Monitoring and Decision-Making Support.

Internet of things (IoT) systems generate a large volume of data all the time. How to choose and transfer which data are essential for decision-making is a challenge. This is especially important for low-cost and low-power designs, for example Long-Range Wide-Area Network (LoRaWan)-based IoT systems, where data volume and frequency are constrained by the protocols. This paper presents an unsupervised learning approach using Laplacian scores to discover which types of sensors can be reduced, without compromising the decision-making. Here, a type of sensor is a feature. An IoT system is designed and implemented for a plant-monitoring scenario. We have collected data and carried out the Laplacian scores. The analytical results help choose the most important feature. A comparative study has shown that using fewer types of sensors, the accuracy of decision-making remains at a satisfactory level.

Integrating Radio-Over-Fiber Communication System and BOTDR Sensor System.

In this paper, we propose and experimentally demonstrate for the first time, the integration of a radio-over-fiber (RoF) communication system and a Brillouin optical time-domain reflectometry (BOTDR) distributed sensor system using a single optical fiber link. In this proof-of-concept integrated system, the communication system is composed of three modulation formats of quadrature phase-shift keying (QPSK), 16-quadrature amplitude modulation (16-QAM) and 64-QAM, which are modulated onto an orthogonal frequency division multiplexing (OFDM) signal. Whereas, the BOTDR sensor system is used for strain and/or temperature monitoring over the fiber distance with a spatial resolution of 5 m using a 25 km single-mode silica fiber. The error vector magnitude (EVM) is analyzed in three modulation formats in the presence of various BOTDR input pump powers. Using QPSK modulation, optimized 18 dBm sensing and 10 dBm data power, the measured EVM values with and without bandpass filter are 3.5% and 14.5%, respectively. The proposed system demonstrates a low temperature measurement error (±0.49 °C at the end of 25 km) and acceptable EVM values, which were within the 3GPP requirements. The proposed integrated system can be effectively applied for practical applications, which significantly reduces the fiber infrastructure cost by effective usage of a single optical fiber link.

Integrated label-free erbium-doped fiber laser biosensing system for detection of single cell Staphylococcus aureus.

A critical challenge to realize ultra-high sensitivity with optical fiber interferometers for label free biosensing is to achieve high quality factors (Q-factor) in liquid. In this work a high Q-factor of 105, which significantly improves the detection resolution is described based on a structure of single mode -core-only -single mode fiber (SCS) with its multimode (or Mach-Zehnder) interference effect as a filter that is integrated into an erbium-doped fiber laser (EDFL) system for excitation. In the case study, the section of core-only fiber is functionalized with porcine immunoglobulin G (IgG) antibodies, which could selectively bind to bacterial pathogen of Staphylococcus aureus (S. aureus). The developed microfiber-based biosensing platform called SCS-based EDFL biosensors can effectively detect concentrations of S. aureus from 10 to 105 CFU/mL, with a responsivity of 0.426 nm wavelength shift in the measured spectrum for S. aureus concentration of 10 CFU/mL. The limit of detection (LoD) is estimated as 7.3 CFU/mL based on the measurement of S. aureus with minimum concentration of 10 CFU/mL. In addition, when a lower concentration of 1 CFU/mL is applied to the biosensor, a wavelength shift of 0.12 nm is observed in 10% of samples (1/10), indicating actual LoD of 1 CFU/mL for the proposed biosensor. Attributed to its good sensitivity, stability, reproducibility and specificity, the proposed EDFL based biosensing platform has great potentials for diagnostics.

Acousto-Pi: An Opto-Acoustofluidic System Using Surface Acoustic Waves Controlled With Open-Source Electronics for Integrated In-Field Diagnostics.

Surface acoustic wave (SAW) devices are increasingly applied in life sciences, biology, and point-of-care applications due to their combined acoustofluidic sensing and actuating properties. Despite the advances in this field, there remain significant gaps in interfacing hardware and control strategies to facilitate system integration with high performance and low cost. In this work, we present a versatile and digitally controlled acoustofluidic platform by demonstrating key functions for biological assays such as droplet transportation and mixing using a closed-loop feedback control with image recognition. Moreover, we integrate optical detection by demonstrating in situ fluorescence sensing capabilities with a standard camera and digital filters, bypassing the need for expensive and complex optical setups. The Acousto-Pi setup is based on open-source Raspberry Pi hardware and 3-D printed housing, and the SAW devices are fabricated with piezoelectric thin films on a metallic substrate. The platform enables the control of droplet position and speed for sample processing (mixing and dilution of samples), as well as the control of temperature based on acousto-heating, offering embedded processing capability. It can be operated remotely while recording the measurements in cloud databases toward integrated in-field diagnostic applications such as disease outbreak control, mass healthcare screening, and food safety.

Flexible multifunctional platform based on piezoelectric acoustics for human-machine interaction and environmental perception.

Flexible human-machine interfaces show broad prospects for next-generation flexible or wearable electronics compared with their currently available bulky and rigid counterparts. However, compared to their rigid counterparts, most reported flexible devices (e.g., flexible loudspeakers and microphones) show inferior performance, mainly due to the nature of their flexibility. Therefore, it is of great significance to improve their performance by developing and optimizing new materials, structures and design methodologies. In this paper, a flexible acoustic platform based on a zinc oxide (ZnO) thin film on an aluminum foil substrate is developed and optimized; this platform can be applied as a loudspeaker, a microphone, or an ambient sensor depending on the selection of its excitation frequencies. When used as a speaker, the proposed structure shows a high sound pressure level (SPL) of ~90 dB (with a standard deviation of ~3.6 dB), a low total harmonic distortion of ~1.41%, and a uniform directivity (with a standard deviation of ~4 dB). Its normalized SPL is higher than those of similar devices reported in the recent literature. When used as a microphone, the proposed device shows a precision of 98% for speech recognition, and the measured audio signals show a strong similarity to the original audio signals, demonstrating its equivalent performance compared to a rigid commercial microphone. As a flexible sensor, this device shows a high temperature coefficient of frequency of -289 ppm/K and good performance for respiratory monitoring.

Ensemble Stochastic Configuration Networks for Estimating Prediction Intervals: A Simultaneous Robust Training Algorithm and Its Application.

Obtaining accurate point prediction of industrial processes' key variables is challenging due to the outliers and noise that are common in industrial data. Hence the prediction intervals (PIs) have been widely adopted to quantify the uncertainty related to the point prediction. In order to improve the prediction accuracy and quantify the level of uncertainty associated with the point prediction, this article estimates the PIs by using ensemble stochastic configuration networks (SCNs) and bootstrap method. The estimated PIs can guarantee both the modeling stability and computational efficiency. To encourage the cooperation among the base SCNs and improve the robustness of the ensemble SCNs when the training data are contaminated with noise and outliers, a simultaneous robust training method of the ensemble SCNs is developed based on the Bayesian ridge regression and M-estimate. Moreover, the hyperparameters of the assumed distributions over noise and output weights of the ensemble SCNs are estimated by the expectation-maximization (EM) algorithm, which can result in the optimal PIs and better prediction accuracy. Finally, the performance of the proposed approach is evaluated on three benchmark data sets and a real-world data set collected from a refinery. The experimental results demonstrate that the proposed approach exhibits better performance in terms of the quality of PIs, prediction accuracy, and robustness.