Indoor localization is important for many applications such as navigation, movement tracking, geotagging, and augmented reality. Most studies have used either Wi-Fi or image signals to determine the user’s location. However, each localization method has advantages and disadvantages. In this study, we propose a hybrid localization system combining the advantages of Wi-Fi and image-based methods. The localization is calculated based on the best four outputs of either image or Wi-Fi localization system. The system was evaluated by comparing the accuracy and unit errors of image-based, Wi-Fi-based, hybrid (image + Wi-Fi), hybrid (Wi-Fi + image) methods. The results showed accuracies of 77.2%, 49.5%, 73.1%, and 81.6% in the image-based, Wi-Fi-based, hybrid (image + Wi-Fi), and hybrid (Wi-Fi + image) methods, respectively. The hybrid (Wi-Fi and image) method has the lowest error and highest accuracy of the four methods compared. In addition, the image-based localization system shows the highest error, while the Wi-Fi-based localization system shows the lowest accuracy. The robot tests prove that the proposed hybrid system can achieve excellent performance in indoor localization. The proposed hybrid system uses both image processing and Wi-Fi fingerprinting methods to determine the mobile device's location by creating the two-phase framework, which can help improve the accuracy of indoor localization.

Enhancing the performance of a chemo-resistive gas sensor is often challenging due to environmental humidity influencing their sensitivity and baseline resistance. One of the most promising ways of overcoming this challenge is through Ultraviolet (UV) illumination onto the sensing material. Most research has focussed on using UV with in-house developed sensors, which has limited their widespread use. In this work, we have evaluated if UV can enhance the performance of commercially available MOX-based gas sensors. The performance of five different MOX sensors has been evaluated, specifically SGX Microtech MiCS6814 (thin-film triple sensor), FIGARO TGS2602 (n-type thick film), and Alphasense VOC sensor (p-type thick film). These sensors were tested towards isobutylene gas under UV light at different wavelengths (UV-278nm & UV-365nm) to investigate its effect on humidity, sensitivity, baseline drift and response, and recovery time of each sensor. Also, to understand the effect of UV on both thin and thick film sensors. The sensors were also tested at three different temperatures and with continuous and different blink speeds (at 2 sec, 500 sec). We found the response time of thin-film sensors for reducing gases was improved from 80 sec under dark conditions to less than 30 sec under UV- 365nm at normal operating temperatures. In addition, all the sensors were left in a dirty environment and the humid-gas testing was repeated. However, due to their robust design, the sensitivity and baseline drift of all the sensors remained the same. This indicates UV has only limited use with commercial gas sensors.

Milk is considered to be a complete food that provides all necessary nutrients to the human body. It is thus very important to ensure good quality milk is consumed. However, the quality of commercially available milk is inconsistent across various sources. A study conducted by the Food Safety and Standards Authority of India (FSSAI) in 2018 found that 68.7 % of milk and milk by-products produced in the country were found to be laced with polluting ingredients. Tasting the milk to ascertain its quality is not always feasible. Developing a device or a non-invasive method to quantitatively ascertain the quality of milk is the need of the hour. We have attempted to develop a system and an inexpensive device capable of detecting the quality of milk without opening the packaging. The experimental setup consists of a circuit and an electrode receptacle which is customized according to the targeted milk packaging. There has been a lot of research in this domain but correlating the capacitance directly to the quality of milk has not been attempted before. In this method, the dielectric constant of the milk is determined from the capacitance read from the experimental setup which is then correlated to the pH of the milk. pH is universally considered to be an indicator of milk quality with the pH of fresh milk being around 6.5-6.7. A regression based model is developed using the data capturing the change in capacitance and pH values over time to estimate the quality of milk. The deployment of this system in the Indian market will help in dynamic pricing based on a quantitative freshness measure and assure the customer of milk quality. The results of the research and a prototype of the device are presented in the paper.

After an Earthquake or a building collapse, victim recovery is a challenging task. In such cases, recovery methods must prioritize fast detection and procurement of the location of victims. Human speech is one such parameter that can be used in rescue operations. This research work discusses the application of Voice Activity Detection (VAD) techniques for detecting and discriminating human speech from noise. In this paper, VAD is performed on three important spectral parameters of signals namely: flux, roll-off, and centroid. Using all the three parameters and their combinations, the VAD algorithm is tested for their success rate on a set of audio samples, containing studio-recorded speech, outdoor speech recording with background noise, and pure noise signals from different sources. The change of the signal parameters over time was plotted in separate graphs. For further processing, the information from the change of speech properties over time had to be reduced to a small set of parameters. Our new approach compresses the audio signal to the average values of positive and negative peaks. The research progresses from a method of manual threshold selection technique to machine learning-based linear discriminant method and a comparative study was made to find the best performing method for detection of speech. Using the cross-validation tests based on the linear discriminant analysis model, flux and centroid individually displayed the highest success rate for all categories of test samples with a recognition rate of 78 % to 83 %. However, stability was further improved by combing these two parameters increasing the rate to 88%.

Emotion identification and categorization have been emerging in the Brain Machine Interface in current era. Audio, visual, and electroencephalography (EEG) data have all been shown to be useful for automated emotion identification in a number of studies. EEG-based emotion detection is a critical component of psychiatric health assessment for individuals. If EEG sensor data are collected from multiple experimental sessions or participants, the underlying signals are invariably non-stationary. Because EEG signals are noisy, non-linear, and non-stationary, developing an intelligent framework that can give high accuracy for emotion identification is a difficult challenge. Many research has shown evidence that EEG brain waves may be used to determine feelings. This study introduces a novel automated emotion identification system that employs deep learning principles to recognize emotions through EEG signals from computer games. EEG data was obtained from 28 distinct participants using a 14-channel Emotive Epoc+ portable and wearable EEG equipment. Participants played four distinct emotional computer games for five minutes each, with a total of 20 minutes of EEG data available for each participant. The suggested framework is simple enough to categorize four classes of emotions during game play. The results demonstrate that the suggested model-based emotion detection framework is a viable method for recognizing emotions from EEG data. The network achieves 99.99 along with less computational time.

Chemiresistive gas sensors are a crucial tool for monitoring gases on a large scale. In order to properly estimate certain gas concentrations and to differentiate between different gases, pattern recognition algorithms, such as neural networks, are used to analyze the complex signals describing the resistivity or conductivity on the sensor material, e.g. MOX or graphene. These algorithms are usually trained on experimental data based on sample sensors measured under either controlled laboratory conditions or in open scenarios involving a reference device. However, in the production process of such low cost sensor technologies, small variations in the physical properties of the sensors can occur. This means that the reaction of a single sensor to a certain concentration can slightly vary from the original sensor used for algorithm development. An example for such a variation would be the operating and heating temperature of the device. In order to study the influence of such variations on the overall performance of pattern recognition algorithms, we used a stochastic simulation model of a graphene-based gas sensor to generate data of different concentration profiles exposed to sensors with different heating settings. Subsequently, we trained machine learning models on different subsets of the synthetic data to estimate the influence of temperature variations on the prediction outcome and to study which training data configurations might increase the performance of the sensors under varying input conditions. Our results show that different temperatures can steadily lower the performance of the algorithms. Moreover, a well-balanced training set featuring several measuring temperatures can increase the robustness of the prediction algorithms.

The 3D printing process deals with the manufacture of parts by adding layers of material onto a heated printing bed. Electret microphones are widely used, low-cost and precise measuring devices. However, its response is negatively affected by higher temperatures due to the Field Effect Transistor utilized in its construction. The Pencil Lead Break (PLB) method is a standardized artificial acoustic emission source utilized for the evaluation of sensors response. The present work aims to study the electret microphone response for 3D printing monitoring, and to evaluate the efficiency of a proposed housing to reduce the printing bed temperature’s influence on the electret microphone’s response. The microphone housing was 3D-printed utilizing ABS filament, and its geometry was designed with the purpose of separating the sensor from the heated bed and creating an acoustic shell. Then, PLB tests were performed, and the raw signal was collected from housed and non-housed microphones at 5MHz sampling frequency. The sensors were tested under three temperatures of the printer bed: at 25ºC (ambient), at 65ºC (operating temperature), and finally after the temperature of the table was naturally stabilized from 65ºC to 25ºC. The signals were investigated in the time and frequency domain. The results show that the housing impacts the microphone’s response positively when operating at 25ºC, where the signals presented higher amplitudes in both domains. However, the response obtained by the housed sensor was considerably attenuated at 65ºC. Furthermore, the signals collected at 25ºC after exposing the housed microphone to heat demonstrate a “greenhouse effect”, keeping the sensor at higher temperatures for an extended period. It can be concluded that the proposed housing did not succeed in reducing the temperature effects in the sensor’s response. However, these effects were shown to be significant and the need for an alternative method to attenuate them is reinforced.

The motivation of this research was to detect and quantify stress level in the phonation signal. This signal is recorded in parallel during scanning in the magnetic resonance imager (MRI) for calculation of the 3D model of the human vocal tract. An examined person is exposed by vibration and noise originated from the gradient system of the running MRI device. The mental stress can effectively be identified by the heart rate (HR), the blood pressure, or other parameters using the photo-plethysmography (PPG) signal. The amplitude of the picked-up PPG signal is usually not constant and it can often be partially disturbed or degraded. Therefore, the sensed raw PPG signal must be smoothed before the HR determination. The filtering as well as the HR determination procedures work in real-time, so the implemented algorithms must be simple, but robust and stable. For proper and safe function in the low magnetic field environment of the MRI device, the PPG sensor must be composed of a non-ferromagnetic material including the power supply part.

The paper describes design, realization, testing, and first practical measurements with two developed wearable PPG sensors working in reflectance mode with real-time Bluetooth data transfer to an external recording device enabling post-processing and storage of PPG signals. Two described realizations differ in the type of a control unit based on the Arduino platform and the used Bluetooth (BT) communication module (working in BT 2.0/BT4.0 BLE standards). In the frame of the performed practical experiments, the basic functions of the developed PPG sensors as well as the built control application were stepwise tested. The auxiliary measurement consists of real-time sensing, transmission, and storage operations (including signal filtering and HR value determination) – in the normal laboratory conditions. Comparative measurements with the oximeter device for calibration of determined HR values were also performed. The main experiment with the PPG signal sensing inside the running MRI device has shown that the PPG signal must be filtered, all PPG sensor parts must be shielded to avoid strong radiofrequency disturbance, and the communication baud rate must be decreased in comparison with the measurement in the standard condition.

The quality of the power systems is related to their capability to predict failures, avoid stoppages, and increase the lifetime of their components. Therefore, science has been developing monitoring systems to identify failures in induction motors, transformers, and transmission lines. In this context, one of the most crucial components of the electrical systems is the insulation devices like bushings, which are constantly subjected to dust, thermal stresses, moisture, etc. These conditions promote insulation deterioration, leading to the occurrence of partial discharges (PD). PDs are localized dielectric breakdown that emits ultra-violet radiation, heat, electromagnet, and acoustics waves. The most traditional techniques to identify PDs on bushings are based on the current, UHF, and acoustic emission analysis. However, thermal analysis stands out as a noise-resistant technique to monitor several components in the power systems.

Although the thermal method is applied to detect different types of faults, such as bad contacts, overloads, etc, this technique has not been previously applied to perform PD detection and evaluate its evolution on bushings. Based on this issue, this article proposes two new indexes to characterize the PD evolution based on the infra-red thermal analysis: the ARC (Area Ratio Coefficient) and the RGBRC (RGB Ratio Coefficient). Seven PD levels were induced in a contaminated bushing, and an infra-red thermal camera captured 20 images per condition, totalizing 140 images. ARC and RGBRC were used to perform the identification of PD evolution. Results indicated that values of the new indexes increase with the PD activity. Thus, the new imaging processing approach can be a promising contribution to literature, improving the reliability and maintenance planning for power transmission systems.

For effective structural health monitoring (SHM), the electro-mechanical impedance (EMI) technique is usually implemented in which smart material, namely, Lead Zirconate Titanate (PZT), is used to observe the possible changes made in the structural performance. The dual nature of PZT has enabled it to work as an actuator as well as a sensor. In this article, the proposal of the multi-sensing technique on the surface-mounted PZT sensors is offered. The investigation is performed on the concrete structures for detecting and localizing the structural damage. Multiple smart sensing units (SSU) are adhesively bonded on the top surface of the simply supported concrete beam. As each PZT sensor has a small zone of influence, therefore, the use of multiple smart sensors is recommended for effective damage detection. As the EMI measurement, the conductance signatures are obtained at different stages in the frequency range of 0-450 kHz. To quantify the changes in the conductance signatures, various damage indices, i.e., root mean square deviation (RMSD) and correlation coefficient (CC), are used. This article also presents the effective methodology for damage localization, which assumes the parallel connection of SSUs under MISO mode. The approximate position of the damage is estimated using the dynamic metrics based upon the multiplexed signatures. The methodology adopted for structural damage detection is effective, as it is verified with the experimental results performed on the concrete structures with multiple surface-mounted PZT sensors. A good agreement is observed in the trend of the numerical results and experimental results, which promotes the implementation of the present methodology for damage detection.