Towards Reliability in Smart Water Sensing Technology: Evaluating Classical Machine Learning Models for Outlier Detection.

In recent years, smart water sensing technology has played a crucial role in water management, addressing the pressing need for efficient monitoring and control of water resources analysis. The challenge in smart water sensing technology resides in ensuring the reliability and accuracy of the data collected by sensors. Outliers are a well-known problem in smart sensing as they can negatively affect the viability of useful analysis and make it difficult to evaluate pertinent data. In this study, we evaluate the performance of four sensors: electrical conductivity (EC), dissolved oxygen (DO), temperature (Temp), and pH. We implement four classical machine learning models: support vector machine (SVM), artifical neural network (ANN), decision tree (DT), and isolated forest (iForest)-based outlier detection as a pre-processing step before visualizing the data. The dataset was collected by a real-time smart water sensing monitoring system installed in Brussels' lakes, rivers, and ponds. The obtained results clearly show that the SVM outperforms the other models, showing 98.38% F1-score rates for pH, 96.98% F1-score rates for temp, 97.88% F1-score rates for DO, and 98.11% F1-score rates for EC. Furthermore, ANN also achieves a significant results, establishing it as a viable alternative.

Trustworthy Environmental Monitoring Using Hardware-Assisted Security Mechanisms.

Environmental monitoring is essential for safeguarding the health of our planet and protecting human health and well-being. Without trust, the effectiveness of environmental monitoring and the ability to address environmental challenges are significantly compromised. In this paper, we present a sensor platform capable of performing authenticated and trustworthy measurements, together with a lightweight security protocol for sending the data from the sensor to a central server anonymously. Besides presenting a new and very efficient symmetric-key-based protocol, we also demonstrate on real hardware how existing embedded security modules can be utilized for this purpose. We provide an in-depth evaluation of the performance and a detailed security analysis.

Evaluating the Performance of Pre-Trained Convolutional Neural Network for Audio Classification on Embedded Systems for Anomaly Detection in Smart Cities.

Environmental Sound Recognition (ESR) plays a crucial role in smart cities by accurately categorizing audio using well-trained Machine Learning (ML) classifiers. This application is particularly valuable for cities that analyzed environmental sounds to gain insight and data. However, deploying deep learning (DL) models on resource-constrained embedded devices, such as Raspberry Pi (RPi) or Tensor Processing Units (TPUs), poses challenges. In this work, an evaluation of an existing pre-trained model for deployment on Raspberry Pi (RPi) and TPU platforms other than a laptop is proposed. We explored the impact of the retraining parameters and compared the sound classification performance across three datasets: ESC-10, BDLib, and Urban Sound. Our results demonstrate the effectiveness of the pre-trained model for transfer learning in embedded systems. On laptops, the accuracy rates reached 96.6% for ESC-10, 100% for BDLib, and 99% for Urban Sound. On RPi, the accuracy rates were 96.4% for ESC-10, 100% for BDLib, and 95.3% for Urban Sound, while on RPi with Coral TPU, the rates were 95.7% for ESC-10, 100% for BDLib and 95.4% for the Urban Sound. Utilizing pre-trained models reduces the computational requirements, enabling faster inference. Leveraging pre-trained models in embedded systems accelerates the development, deployment, and performance of various real-time applications.

XCycles Backprojection Acoustic Super-Resolution.

The computer vision community has paid much attention to the development of visible image super-resolution (SR) using deep neural networks (DNNs) and has achieved impressive results. The advancement of non-visible light sensors, such as acoustic imaging sensors, has attracted much attention, as they allow people to visualize the intensity of sound waves beyond the visible spectrum. However, because of the limitations imposed on acquiring acoustic data, new methods for improving the resolution of the acoustic images are necessary. At this time, there is no acoustic imaging dataset designed for the SR problem. This work proposed a novel backprojection model architecture for the acoustic image super-resolution problem, together with Acoustic Map Imaging VUB-ULB Dataset (AMIVU). The dataset provides large simulated and real captured images at different resolutions. The proposed XCycles BackProjection model (XCBP), in contrast to the feedforward model approach, fully uses the iterative correction procedure in each cycle to reconstruct the residual error correction for the encoded features in both low- and high-resolution space. The proposed approach was evaluated on the dataset and showed high outperformance compared to the classical interpolation operators and to the recent feedforward state-of-the-art models. It also contributed to a drastically reduced sub-sampling error produced during the data acquisition.

A New Method for Acoustic Priority Vehicle Detection Based on a Self-Powering Triboelectric Acoustic Sensor Suitable for Low-Power Wireless Sensor Networks.

Traffic congestion is, on a daily basis, responsible for a significant amount of economic and social costs. One of the critical examples is the obstruction of priority vehicles during fast trajectories, which potentially costs lives and property in case of delay that is too great. By means of visual sensing methods, solutions and schedules have already been proposed for adjusting traffic light sequences depending on a priority vehicle's position. However, these mechanisms are computation and power intensive. Deploying and powering a large-scale network will have a crucial economical cost. Furthermore, these devices will not always have access to sufficient power. To provide a solution, we developed an acoustic and self-powered device that can detect priority vehicles and can be cost effectively deployed to define a sensor network. The device combines the detection of priority vehicles and the harvesting of sound energy through triboelectrification. This paper will introduce the use of triboelectric energy harvesting, specifically in a self-powered wireless sensor network for priority vehicle detection. Furthermore, it shows how to increase the power performance of such a generator. Finally, the results are analyzed.

CABE: A Cloud-Based Acoustic Beamforming Emulator for FPGA-Based Sound Source Localization.

Microphone arrays are gaining in popularity thanks to the availability of low-cost microphones. Applications including sonar, binaural hearing aid devices, acoustic indoor localization techniques and speech recognition are proposed by several research groups and companies. In most of the available implementations, the microphones utilized are assumed to offer an ideal response in a given frequency domain. Several toolboxes and software can be used to obtain a theoretical response of a microphone array with a given beamforming algorithm. However, a tool facilitating the design of a microphone array taking into account the non-ideal characteristics could not be found. Moreover, generating packages facilitating the implementation on Field Programmable Gate Arrays has, to our knowledge, not been carried out yet. Visualizing the responses in 2D and 3D also poses an engineering challenge. To alleviate these shortcomings, a scalable Cloud-based Acoustic Beamforming Emulator (CABE) is proposed. The non-ideal characteristics of microphones are considered during the computations and results are validated with acoustic data captured from microphones. It is also possible to generate hardware description language packages containing delay tables facilitating the implementation of Delay-and-Sum beamformers in embedded hardware. Truncation error analysis can also be carried out for fixed-point signal processing. The effects of disabling a given group of microphones within the microphone array can also be calculated. Results and packages can be visualized with a dedicated client application. Users can create and configure several parameters of an emulation, including sound source placement, the shape of the microphone array and the required signal processing flow. Depending on the user configuration, 2D and 3D graphs showing the beamforming results, waterfall diagrams and performance metrics can be generated by the client application. The emulations are also validated with captured data from existing microphone arrays.

An Ultrasonic Multiple-Access Ranging Core Based on Frequency Shift Keying Towards Indoor Localization.

This paper describes a new approach and implementation methodology for indoor ranging based on the time difference of arrival using code division multiple access with ultrasound signals. A novel implementation based on a field programmable gate array using finite impulse response filters and an optimized correlation demodulator implementation for ultrasound orthogonal signals is developed. Orthogonal codes are modulated onto ultrasound signals using frequency shift keying with carrier frequencies of 24.5 kHz and 26 kHz. This implementation enhances the possibilities for real-time, embedded and low-power tracking of several simultaneous transmitters. Due to the high degree of parallelism offered by field programmable gate arrays, up to four transmitters can be tracked simultaneously. The implementation requires at most 30% of the available logic gates of a Spartan-6 XC6SLX45 device and is evaluated on accuracy and precision through several ranging topologies. In the first topology, the distance between one transmitter and one receiver is evaluated. Afterwards, ranging analyses are applied between two simultaneous transmitters and one receiver. Ultimately, the position of the receiver against four transmitters using trilateration is also demonstrated. Results show enhanced distance measurements with distances ranging from a few centimeters up to 17 m, while keeping a centimeter-level accuracy.

Ultrasonic multiple-access ranging system using spread spectrum and MEMS technology for indoor localization.

Indoor localization of persons and objects poses a great engineering challenge. Previously developed localization systems demonstrate the use of wideband techniques in ultrasound ranging systems. Direct sequence and frequency hopping spread spectrum ultrasound signals have been proven to achieve a high level of accuracy. A novel ranging method using the frequency hopping spread spectrum with finite impulse response filtering will be investigated and compared against the direct sequence spread spectrum. In the first setup, distances are estimated in a single-access environment, while in the second setup, two senders and one receiver are used. During the experiments, the micro-electromechanical systems are used as ultrasonic sensors, while the senders were implemented using field programmable gate arrays. Results show that in a single-access environment, the direct sequence spread spectrum method offers slightly better accuracy and precision performance compared to the frequency hopping spread spectrum. When two senders are used, measurements point out that the frequency hopping spread spectrum is more robust to near-far effects than the direct sequence spread spectrum.

SoundCompass: a distributed MEMS microphone array-based sensor for sound source localization.

Sound source localization is a well-researched subject with applications ranging from localizing sniper fire in urban battlefields to cataloging wildlife in rural areas. One critical application is the localization of noise pollution sources in urban environments, due to an increasing body of evidence linking noise pollution to adverse effects on human health. Current noise mapping techniques often fail to accurately identify noise pollution sources, because they rely on the interpolation of a limited number of scattered sound sensors. Aiming to produce accurate noise pollution maps, we developed the SoundCompass, a low-cost sound sensor capable of measuring local noise levels and sound field directionality. Our first prototype is composed of a sensor array of 52 Microelectromechanical systems (MEMS) microphones, an inertial measuring unit and a low-power field-programmable gate array (FPGA). This article presents the SoundCompass's hardware and firmware design together with a data fusion technique that exploits the sensing capabilities of the SoundCompass in a wireless sensor network to localize noise pollution sources. Live tests produced a sound source localization accuracy of a few centimeters in a 25-m2 anechoic chamber, while simulation results accurately located up to five broadband sound sources in a 10,000-m2 open field.

Extending the IEEE 802.15.4 security suite with a compact implementation of the NIST P-192/B-163 elliptic curves.

Typically, commercial sensor nodes are equipped with MCUsclocked at a low-frequency (i.e., within the 4-12 MHz range). Consequently, executing cryptographic algorithms in those MCUs generally requires a huge amount of time. In this respect, the required energy consumption can be higher than using a separate accelerator based on a Field-programmable Gate Array (FPGA) that is switched on when needed. In this manuscript, we present the design of a cryptographic accelerator suitable for an FPGA-based sensor node and compliant with the IEEE802.15.4 standard. All the embedded resources of the target platform (Xilinx Artix-7) have been maximized in order to provide a cost-effective solution. Moreover, we have added key negotiation capabilities to the IEEE 802.15.4 security suite based on Elliptic Curve Cryptography (ECC). Our results suggest that tailored accelerators based on FPGA can behave better in terms of energy than contemporary software solutions for motes, such as the TinyECC and NanoECC libraries. In this regard, a point multiplication (PM) can be performed between 8.58- and 15.4-times faster, 3.40- to 23.59-times faster (Elliptic Curve Diffie-Hellman, ECDH) and between 5.45- and 34.26-times faster (Elliptic Curve Integrated Encryption Scheme, ECIES). Moreover, the energy consumption was also improved with a factor of 8.96 (PM).

Active self-testing noise measurement sensors for large-scale environmental sensor networks.

Large-scale noise pollution sensor networks consist of hundreds of spatially distributed microphones that measure environmental noise. These networks provide historical and real-time environmental data to citizens and decision makers and are therefore a key technology to steer environmental policy. However, the high cost of certified environmental microphone sensors render large-scale environmental networks prohibitively expensive. Several environmental network projects have started using off-the-shelf low-cost microphone sensors to reduce their costs, but these sensors have higher failure rates and produce lower quality data. To offset this disadvantage, we developed a low-cost noise sensor that actively checks its condition and indirectly the integrity of the data it produces. The main design concept is to embed a 13 mm speaker in the noise sensor casing and, by regularly scheduling a frequency sweep, estimate the evolution of the microphone's frequency response over time. This paper presents our noise sensor's hardware and software design together with the results of a test deployment in a large-scale environmental network in Belgium. Our middle-range-value sensor (around €50) effectively detected all experienced malfunctions, in laboratory tests and outdoor deployments, with a few false positives. Future improvements could further lower the cost of our sensor below €10.

On the ability of consumer electronics microphones for environmental noise monitoring.

The massive production of microphones for consumer electronics, and the shift from dedicated processing hardware to PC-based systems, opens the way to build affordable, extensive noise measurement networks. Applications include e.g. noise limit and urban soundscape monitoring, and validation of calculated noise maps. Microphones are the critical components of such a network. Therefore, in a first step, some basic characteristics of 8 microphones, distributed over a wide range of price classes, were measured in a standardized way in an anechoic chamber. In a next step, a thorough evaluation was made of the ability of these microphones to be used for environmental noise monitoring. This was done during a continuous, half-year lasting outdoor experiment, characterized by a wide variety of meteorological conditions. While some microphones failed during the course of this test, it was shown that it is possible to identify cheap microphones that highly correlate to the reference microphone during the full test period. When the deviations are expressed in total A-weighted (road traffic) noise levels, values of less than 1 dBA are obtained, in excess to the deviation amongst reference microphones themselves.

Using sound-taste correspondences to enhance the subjective value of tasting experiences.

The soundscapes of those places where we eat and drink can influence our perception of taste. Here, we investigated whether contextual sound would enhance the subjective value of a tasting experience. The customers in a chocolate shop were invited to take part in an experiment in which they had to evaluate a chocolate's taste while listening to an auditory stimulus. Four different conditions were presented in a between-participants design. Envisioning a more ecological approach, a pre-recorded piece of popular music and the shop's own soundscape were used as the sonic stimuli. The results revealed that not only did the customers report having a significantly better tasting experience when the sounds were presented as part of the food's identity, but they were also willing to pay significantly more for the experience. The method outlined here paves a new approach to dealing with the design of multisensory tasting experiences, and gastronomic situations.