Showing papers by "Linga Reddy Cenkeramaddi published in 2023"

Recent Advances in Thermal Imaging and its Applications Using Machine Learning: A Review

Abstract: Recent advancements in thermal imaging sensor technology have resulted in the use of thermal cameras in a variety of applications, including automotive, industrial, medical, defense and space, agriculture, and other related fields. Thermal imaging, unlike RGB imaging, does not rely on background light, and the technique is nonintrusive while also protecting privacy. This review article focuses on the most recent advancements in thermal imaging technology, key performance parameters, an overview of its applications, and machine-learning techniques applied to thermal images for various tasks. This article begins with the most recent advancements in thermal imaging, followed by a classification of thermal cameras and their key specifications, and finally a review of machine-learning techniques used on thermal images for various applications. This detailed review article is highly useful for designing thermal imaging-based applications using various machine-learning techniques.

Dielectric Modulated Bilayer Electrode Top Contact OTFT for Label Free Biosensing

Abstract: In this paper, dielectric modulated bilayer electrodes top contact organic field effect transistor (DMBETC-OTFT) is investigated as a biosensing device for label-free detection of biomolecules. The nanocavity used for biomolecule detection is created by etching the oxide in a conventional OTFT device. Neutral and charged biomolecules can be detected by the proposed device using their respective dielectric constants and charge densities. Subthreshold swing (SS), on-current ( I_ON ), and on-off current ratio ( I_ON/I_OFF ) are the main biosensing performance characteristics computed and compared for different gate work function (ϕ _m ) and cavity thickness ( T_gap ) for the proposed biosensor device. The change in drain current ( I_D ), as well as the I_ON/I_OFF ratio, have both been calculated to investigate the sensitivity of the proposed biosensor. The influence of the gate work function is also investigated to improve the sensitivity of the proposed device. According to the finding of this study, using a gate work function with a lower value results in a significant increase in sensitivity. For charged biomolecules ( Q_f = +1 × 10 ¹² cm ^–2 ) with dielectric constant of biomoecules (K = 12), the highest drain current sensitivity is 4.5 × 10 ³ . The drain current sensitivity achieved is four times greater, when comparing the proposed device to the latest published work of metal controlled dielectric modulated OTFT-based sensor. The proposed device also has a high I_ON/I_OFF sensitivity of 4.60 × 10 ² when V_GS = -3.0 V and V_DS = -1.5 V. In light of its high sensitivity, low cost, and bio-compatibility, the DMBETC-OTFT biosensor holds great promise for the advancement of new demanding flexible biosensing applications.

Generalization of Relative Change in a Centrality Measure to Identify Vital Nodes in Complex Networks

Abstract: Identifying vital nodes is important in disease research, spreading rumors, viral marketing, and drug development. The vital nodes in any network are used to spread information as widely as possible. Centrality measures such as Degree centrality (D), Betweenness centrality (B), Closeness centrality (C), Katz (K), Cluster coefficient (CC), PR (PageRank), LGC (Local and Global Centrality), ISC (Isolating Centrality) centrality measures can be used to effectively quantify vital nodes. The majority of these centrality measures are defined in the literature and are based on a network’s local and/or global structure. However, these measures are time-consuming and inefficient for large-scale networks. Also, these measures cannot study the effect of removal of vital nodes in resource-constrained networks. To address these concerns, we propose the six new centrality measures namely GRACC, LRACC, GRAD, LRAD, GRAK, and LRAK. We develop these measures based on the relative change of the clustering coefficient, degree, and Katz centralities after the removal of a vertex. Next, we compare the proposed centrality measures with D, B, C, CC, K, PR, LGC, and ISC to demonstrate their efficiency and time complexity. We utilize the SIR (Susceptible-Infected-Recovered) and IC (Independent Cascade) models to study the maximum information spread of proposed measures over conventional ones. We perform extensive simulations on large-scale real-world data sets and prove that local centrality measures perform better in some networks than global measures in terms of time complexity and information spread. Further, we also observe the number of cliques drastically improves the efficiency of global centrality measures.

Unified Quality-Aware Compression and Pulse-Respiration Rates Estimation Framework for Reducing Energy Consumption and False Alarms of Wearable PPG Monitoring Devices

Abstract: Due to the high demands of tiny, compact, lightweight, and low-cost photoplethysmogram (PPG) monitoring devices, these devices are resource-constrained including limited battery power. Consequently, it highly demands frequent charge or battery replacement in the case of continuous PPG sensing and transmission. Further, PPG signals are often severely corrupted under ambulatory and exercise recording conditions, leading to frequent false alarms. In this paper, we propose a unified quality-aware compression and pulse-respiration rates estimation framework for reducing energy consumption and false alarms of wearable and edge PPG monitoring devices by exploring predictive coding techniques for jointly performing signal quality assessment (SQA), data compression and pulse rate (PR) and respiration rate (RR) estimation without the use of different domains of signal processing techniques that can be achieved by using the features extracted from the smoothed prediction error signal. By using the five standard PPG databases, the performance of the proposed unified framework is evaluated in terms of compression ratio (CR), mean absolute error (MAE), false alarm reduction rate (FARR), processing time (PT) and energy saving (ES). The compression, PR, RR estimation, and SQA results are compared with the existing methods and results of uncompressed PPG signals with sampling rates of 125 Hz and 25 Hz. The proposed unified quality-aware framework achieves an average CR of 4%, SQA (Se of 92.00%, FARR of 84.87%), PR (MAE: 0.46 ±1.20) and RR (MAE: 1.75 (0.65-4.45), PT (sec) of 15.34 ±0.01) and ES of 70.28% which outperforms the results of uncompressed PPG signal with a sampling rate of 125 Hz. Arduino Due computing platform-based implementation demonstrates the real-time feasibility of the proposed unified quality-aware PR-RR estimation and data compression and transmission framework on the limited computational resources. Thus, it has great potential in improving energy-efficiency and trustworthiness of wearable and edge PPG monitoring devices.

Adversarial Unsupervised Domain Adaptation for Hand Gesture Recognition Using Thermal Images

Abstract: Hand gesture recognition has a wide range of applications, including in the automotive and industrial sectors, health assistive systems, authentication, and so on. Thermal images are more resistant to environmental changes than red–green–blue (RGB) images for hand gesture recognition. However, one disadvantage of using thermal images for the aforementioned task is the scarcity of labeled thermal datasets. To tackle this problem, we propose a method that combines unsupervised domain adaptation (UDA) techniques with deep-learning (DL) technology to remove the need for labeled data in the learning process. There are several types and methods for implementing UDA, with adversarial UDA being one of the most common. In this article, the first time in this field, we propose a novel adversarial UDA model that uses channel attention and bottleneck layers to learn domain-invariant features across RGB and thermal domains. Thus, the proposed model leverages the information from the labeled RGB data to solve the hand gesture recognition task using thermal images. We evaluate the proposed model on two hand gesture datasets, namely, Sign Digit Classification and Alphabet Gesture Classification, and compare it to other benchmark models in terms of accuracy, model size, and model parameters. Our model outperforms the other state-of-the-art methods on the Sign Digit Classification and Alphabet Gesture Classification datasets and achieves 91.32% and 80.91% target test accuracy, respectively.

Multitarget Detection and Tracking by Mitigating Spot Jammer Attack in 77-GHz mm-Wave Radars: An Experimental Evaluation

Abstract: Small form factor radar sensors at millimeter wavelengths find numerous applications in the industrial and automotive sectors. These radar sensors provide improved range resolution, good angular resolution, and enhanced Doppler resolution for short range and ultrashort ranges. However, it is challenging to detect and track the targets accurately when a radar is interfered by another radar. This article proposes an experimental evaluation of a 77-GHz IWR1642 radar sensor in the presence of a second 77-GHz AWR1642 radar sensor acting as a spot jammer. A real-time experiment is carried out by considering five different targets of various cross sections, such as a car, a larger size motorcycle, a smaller size motorcycle, a cyclist, and a pedestrian. The collected real-time data are processed by four different constant false alarm rate detectors, cell averaging (CA)-CFAR, ordered statistics (OS)-CFAR, greatest of CA (GOCA)-CFAR, and smallest of CA (SOCA)-CFAR. Following that, data from these detectors are fed into two different clustering algorithms (density-based spatial clustering of applications with noise (DBSCAN) and K-means), followed by the extended Kalman filter (EKF)-based tracker with global nearest neighbor (GNN) data association, which provide tracks of various targets with and without the presence of a jammer. Furthermore, four different metrics [tracks reported (TR), track segments (TSs), false tracks (FTs), and track loss (TL)] are used to evaluate the performance of various tracks generated for two clustering algorithms with four detection schemes. The experimental results show that the DBSCAN clustering algorithm outperforms the K-means clustering algorithm for many cases.

Performance Analysis of Spectrum Sharing Radar in Multipath Environment

Abstract: Radar based sensing and communication systems sharing a common spectrum have become a potential research problem in recent years due to spectrum scarcity. The spectrum sharing radar (SSR) is a new technology that uses the total available bandwidth (BW) for both radar based sensing and communication. Unlike traditional radar, the SSR divides the total available BW into radar-only and mixed-use bands. In a radar-only band, only radar sensor signals can be transmitted and received. In contrast, radar and communication signals can both be transmitted and received in the mixed-use band. Taking such BW sharing into account, this paper investigates the performance of SSR in an information-theoretic sense. To evaluate performance, mutual information (MI), spectral efficiency (SE) and capacity (C) metrics are used. Initially, this paper considered a clean environment (no multipath) in order to evaluate performance metrics in the mixed-use band with and without successive interference cancellation. Following that, this paper addresses the performance of BW allocation by allocating low to high BW in mixed-band. Furthermore, the performance metrics are extended to account for the multipath environment, and the same analogy as in a clean environment is used. In addition, the MI and SE of traditional radar system is taken into account when comparing the performance of SSR with and without the use of the SIC. Finally, MI and capacity results show that using the SIC scheme in a mixed-use band yields performance comparable to traditional radar and communication system. In terms of SE, the SSR with SIC scheme outperforms traditional radar and communication system.

Energy and Throughput Management in Delay-Constrained Small-World UAV-IoT Network

Abstract: Multihop data routing over a large-scale Internet of Things (IoT) network results in energy imbalance and poor data throughput performance. In addition, data transmission using a large number of hops causes more delay. In light of this, in this work, a novel method of energy and throughput management in a delay-constrained small-world unmanned aerial vehicle (UAV)-IoT network is proposed. The proposed small-world framework optimizes the number of hops required for the data transmission leading to improved energy efficiency and quality of service. The method introduces optimal long-range links between device pairs resulting in low average path length and high clustering coefficient which are called as small-world characteristics. Therefore, in this work, UAVs are deployed to collect the data from IoT devices and forward it to the ground station (GS) utilizing the small world framework. It is shown through results that the network delays corresponding to the proposed method, conventional routing method, low-energy adaptive clustering hierarchy (LEACH) protocol, modified LEACH protocol, and canonical particle multiswarm (CPMS) method are 789.39, 1602.53, 1000.92, 873.63, and 999.79 s, respectively. It is also observed that the number of dead UAVs in case of the proposed method is reduced when compared to other existing methods. It is also noticed that the proposed method results in 100% packet delivery ratio (PDR) dominating LEACH and modified LEACH protocols. Thus, it is shown that the proposed method outperforms the other shortest path methods in terms of network latency, lifetime, and PDR. Further, the effect of location of GS, velocities of UAVs, and hovering heights of UAVs is considered for the performance evaluation of the proposed method. The obtained results validate the significance of utilization of the proposed method over various network scenarios.

Detection of Depression Using Weighted Spectral Graph Clustering With EEG Biomarkers

Abstract: The alarming annual growth in the number of people affected by Major Depressive Disorder (MDD) is a problem on a global scale. In the primary scrutiny of depression, Electroencephalography (EEG) is one of the analytical tools available. Machine Learning (ML) and Deep Neural Networks (DNN) methods are the most common techniques for MDD diagnosis using EEG. However, these ML methods heavily rely on manually annotated EEG signals, which can only be generated by experts, for training. This also necessitates a large amount of memory and time constraints. The requirement of huge amounts of data to foresee emerging tendencies or undiscovered alignments is enforced. This article develops an unsupervised learning method for identifying MDD in light of these difficulties. The preprocessed EEG is used to extract three quantitative biomarkers (Band Power: Beta, Delta, and Theta), and three signal features (Detrended Fluctuation Analysis (DFA), Higuchi’s Fractal Dimension (HFD), and Lempel-Ziv Complexity (LZC)). Through the extracted features, an undirected graph is created using the features as a weight along the edges, with nodes as channels in EEG recording. The bifurcation of the subjects in either of the classes (MDD or N) is done by implementing spectral clustering. A 98% accuracy with a 2.5% of miss-classification error is achieved for the left hemisphere. In contrast, a 97% accuracy with a 3.3% $CEP$ (or miss-classification error or Classification Error Percentage) is achieved for the right hemisphere. FP1 and F8 channels have achieved the highest possible level of classification accuracy.

A Socially-Aware Radio Map Framework for Improving QoS of UAV-Assisted MEC Networks

Abstract: The expeditious growth of the Internet of Things (IoT) has accelerated the evolution of multi-access edge computing (MEC). MEC alleviates the challenges of conventional cloud computing, such as high data latency, poor data gathering reliability, increased network cost, and lack of network robustness. The primary objective of MEC is to facilitate a hierarchy of edge servers to address these quality-of-service (QoS) challenges, especially the information propagation issue due to the mobility of IoT devices (IoD). Further, social-relationship among mobile IoD is a critical parameter used to reduce the data transmission delay and queue size at the MEC. Specifically, in this work, a novel socially-aware radio map generation method is proposed to compute the fine-grained and accurate locations of QoS-deprived areas. Firstly, a novel method to compute the social relationship index (SRI) factor is proposed on the basis of current and future encounters among moving IoDs. Then the obtained SRI factor is used to form clusters of mobile IoD. The clusters’ signal to interference plus noise ratio (SINR) is then used to generate the socially-aware radio map. Following that, unmanned aerial vehicles (UAV) use this radio map, which contains rich and serviceable channel information, for 3D beamforming towards the mobile clusters. Using the obtained radio map, Kalman filter-based offline path planning of UAVs is proposed to minimize the UAVs flying distance from the initial to final locations. Furthermore, an optimization problem is formulated to assess the performance of the proposed method. Finally, the performance of the proposed method is compared with the existing methods, taking into account various network parameters such as optimum number of UAVs needed to cover the deployed area, data transmission delay, and received SINR.

Autonomous UAV Path Planning using Modified PSO for UAV-Assisted Wireless Networks

Abstract: Recently, unmanned aerial vehicles (UAVs) have attained considerable attention for providing reliable and cost-effective communication due to the flexibility of deployment and line of sight (LoS) propagation. Efficient UAV path planning is one of the key aspects that need to be addressed to minimize energy consumption and satisfy the rate requirements of the user. Thus, in this work, we propose a novel framework that utilizes the modified Particle Swarm Optimization (PSO) algorithm for UAV path planning to support the rate requirements of the user. In the proposed framework, the problem of joint path planning and energy consumption is formulated to improve the instantaneous sum rate of the user. In order to solve the formulation, the proposed framework involves two steps. Initially, the line of sight probability is used to obtain an optimal destination location at which the UAV is in LoS with the user and offers the required downlink rate. Following that, the modified PSO is used to find the most energy-efficient path from the source to the destination. Through experiments, we show that the proposed framework provides a three-dimensional (3D) path in a complex environment, and has the ability to avoid obstacles in the path. In addition, it minimizes energy consumption and travel time and improves the user rate as compared to the state-of-the-art methods. Finally, the performance of the framework is tested in three different scenarios.

Generalized multi-cavity laser self-mixing interferometry based on scattering theory.

Abstract: We present a generalized mathematical model and algorithm for the multi-cavity self-mixing phenomenon based on scattering theory. Scattering theory, which is extensively used for travelling wave is exploited to demonstrate that the self-mixing interference from multiple external cavities can be modelled in terms of individual cavity parameters recursively. The detailed investigation shows that the equivalent reflection coefficient of coupled multiple cavities is a function of both attenuation coefficient and the phase constant, hence propagation constant. The added benefit with recursively model is that it is computationally very efficient to model large number of parameters. Finally, with the aid of simulation and mathematical modelling, we demonstrate how the individual cavity parameters such as cavity length, attenuation coefficient, and refractive index of individual cavities can be tuned to get a self-mixing signal with optimal visibility. The proposed model intends to leverage system description for biomedical applications when probing multiple diffusive media with distinct characteristics, but could be equally extended to any setup in general.

Multitarget Angle of Arrival Estimation Using Rotating mmWave FMCW Radar and Yolov3

Abstract: It is still challenging to accurately localize unmanned aerial vehicles (UAVs) from a ground control station (GCS) using various sensors. The mmWave frequency-modulated continuous wave (FMCW) radars offer excellent performance for target detection and localization in harsh environments and low lighting conditions. However, the estimated angle of arrival (AoA) of targets in the captured scene is quite poor. This article focuses on improving AoA estimation by combining the cutting-edge machine learning (ML) algorithms with a mechanical radar rotor setup. An mmWave FMCW radar system is mounted on a programmable rotor to capture range–angle maps of targets at various locations. The range–angle images are then labeled and trained further with the Yolov3 algorithm. Subsequent testing reveals that for detected target objects, the centroid of the bounding boxes from the detected objects provides accurate AoA estimation with very low root mean square error (RMSE). The results show that the proposed approach outperforms traditional methods in terms of performance and estimation accuracy.

Real-Time Missing Data Estimation in Water Networks

Abstract: Estimating missing data (e.g., pressure) in geographically distributed nodes is one of the fundamental challenges in a water distribution network (WDN). The Model-Based Kalman Filter is a simple solution that can estimate missing data in systems with a linear Gaussian state model. However, in large WDN, the state model is non-linear, and accurate system dynamics are unknown. Thus, we investigate the use of Data-Driven Kalman Filter (DDKF), in which the Recurrent Neural Network (RNN) updates the Kalman Gain. This data-driven approach has the potential to reduce approximation errors caused by the state model’s nonlinearity. As a result, this work demonstrates the application of a DDKF in a WDN. In addition, we demonstrate numerically that the DDKF can overcome approximation errors caused by flow equations and can estimate missing data in a WDN.

DepCap: A Smart Healthcare Framework for EEG Based Depression Detection Using Time-Frequency Response and Deep Neural Network

Abstract: A novel wearable consumer electronics device for detecting Major Depressive Disorder (MDD) has been developed using deep learning techniques for smart healthcare. Accurate identification of MDD through individual interviews or perceiving Electroencephalogram (EEG) signals is challenging. This study presents the concept of a novel wearable smart cap named DepCap for real-time detection of depression using EEG signals. First, spectrogram images are generated from the EEG signals of depressed and healthy patients using Short-Time Fourier Transform (STFT) to extract valuable features. Then, these spectrogram images obtained from STFT are used as input to the classification model. A deep analysis is done using various neural networks consisting of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). RNNs are used to extract temporal data from the EEG, while CNNs are used to retrieve spatial information. Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) are the two different kinds of RNNs evaluated in this work. The implemented combination models are (STFT+CNN), (STFT+CNN-LSTM) and (STFT+CNN-GRU). Four pre-trained models, Inception, AlexNet, VGG16, and ResNet50 are also implemented along with the combination models. The dataset for this work is a publicly accessible dataset with 33 major depressive disorders and 30 healthy subjects. The evaluation results show that the STFT+CNN-LSTM has much better performance in terms of accuracy, sensitivity, specificity, and precision of 99.9%, 100%, 99.8%, and 99.4%, respectively, than other implemented models. The proposed wearable device DepCap is based on the STFT+CNN+LSTM model and is also integrated with the Internet of Medical Things (IoMT) framework for real-time depression detection.