Search

Eye movements are essential in human-computer interaction (HCI) because they offer insights into individuals’ cognitive states and visual attention. Techniques for adequately assessing gaze have increased in the last two decades. Notably, video-based tracking methods have gained considerable interest within the research community due to their nonintrusive nature, enabling precise and convenient gaze estimation without physical contact or invasive measures. This paper introduces a video-based gaze-tracking method that presents an affordable, user-friendly, and dependable human-computer interaction (HCI) system based on iris movement. By utilizing the MediaPipe face mesh model, facial features are extracted from real-time video sequences. A 5-point user-specific calibration and multiple regression techniques are employed to predict the gaze point on the screen accurately. The proposed system effectively handles changes in body position and user posture through real-time re-calibration using z-index tracking. Furthermore, it compensates for minor head movements that may introduce inaccuracies. The proposed system is cost-effective, with a general cost below $25{\$}$ , which may vary based on camera usage. Thirteen participants were involved in the system testing. The system demonstrates a high level of sensitivity to low light conditions, a strong response to changes in distance, and a moderate reaction to glasses, with an average frame processing time of 0.047 seconds. On average, it achieves a visual angle accuracy of 1.12 degrees with head movement and 1.3 degrees without head movement.

Autism Spectrum Disorder (ASD) is a type of neurodevelopmental disorder that affects the everyday life of affected patients. Though it is considered hard to completely eradicate this disease, disease severity can be mitigated by taking early interventions. In this paper, we propose an effective framework for the evaluation of various Machine Learning (ML) techniques for the early detection of ASD. The proposed framework employs four different Feature Scaling (FS) strategies i.e., Quantile Transformer (QT), Power Transformer (PT), Normalizer, and Max Abs Scaler (MAS). Then, the feature-scaled datasets are classified through eight simple but effective ML algorithms like Ada Boost (AB), Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (KNN), Gaussian Naïve Bayes (GNB), Logistic Regression (LR), Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA). Our experiments are performed on four standard ASD datasets (Toddlers, Adolescents, Children, and Adults). Comparing the classification outcomes using various statistical evaluation measures (Accuracy, Receiver Operating Characteristic: ROC curve, F1-score, Precision, Recall, Mathews Correlation Coefficient: MCC, Kappa score, and Log loss), the best-performing classification methods, and the best FS techniques for each ASD dataset are identified. After analyzing the experimental outcomes of different classifiers on feature-scaled ASD datasets, it is found that AB predicted ASD with the highest accuracy of 99.25%, and 97.95% for Toddlers and Children, respectively and LDA predicted ASD with the highest accuracy of 97.12% and 99.03% for Adolescents and Adults datasets, respectively. These highest accuracies are achieved while scaling Toddlers and Children with normalizer FS and Adolescents and Adults with the QT FS method. Afterward, the ASD risk factors are calculated, and the most important attributes are ranked according to their importance values using four different Feature Selection Techniques (FSTs) i.e., Info Gain Attribute Evaluator (IGAE), Gain Ratio Attribute Evaluator (GRAE), Relief F Attribute Evaluator (RFAE), and Correlation Attribute Evaluator (CAE). These detailed experimental evaluations indicate that proper finetuning of the ML methods can play an essential role in predicting ASD in people of different ages. We argue that the detailed feature importance analysis in this paper will guide the decision-making of healthcare practitioners while screening ASD cases. The proposed framework has achieved promising results compared to existing approaches for the early detection of ASD.

The fundamentals of nonlinear constraint optimization are exploited in this paper to design a decentralized PID controller for variable area coupled tank systems. Maintaining the tanks at the desired level is the main objective of the proposed decentralized control algorithm. The benchmark coupled tank systems are modeled to the first order plus dead time (FOPDT) model by designing decoupled subsystems. The control law is designed for decoupled systems to attain the servo response by minimizing loop interactions. Further, the overshoot is reduced by imposing constraints on the maximum closed-loop amplitude ratio of the system. Furthermore, the robustness of the controller is analyzed and stability is verified with the Kharitonov-Hurwitz theorem. A concise comparison is carried out between the proposed control law with existing methods in order to highlight the productive application of the proposed control algorithm. From the results, it envisaged that the proposed controller shows better responses compared to the existing methods. Besides, the efficacious nature of the proposed control scheme is validated by considering a wide range of closed-loop amplitude ratios.

Rapid and sensitive bioburden detection is of paramount importance in different applications including public health, and food and water safety. To overcome the traditional limitations of bacterial detection i.e., lengthy culture time, and complicated procedure, a low-cost, portable multichannel fluorometer coupled with machine learning (ML) has been implemented in this study. Five different strains of bacterial samples were tested along with the negative control for time-series fluorescence data collection and analysis. We applied different conventional unsupervised and supervised machine learning techniques with extracted features followed by preprocessing of the data. Initially, machine learning algorithms were applied for the qualitative detection of bacteria by binary classification followed by regression analysis to predict the level of contamination for E. coli. The multiclass classification was used to identify gram-positive, and gram-negative bacterial strains and differentiate all the bacterial strains tested. Our results show that around 97.9% accuracy can be achieved for bacterial contamination detection for as low as 1 CFU/mL while 92.1% accuracy can be achieved for differentiating the gram-positive and gram-negative strains. Additionally, with 1 minute of data, high accuracy is obtained for detecting bioburden, proving the multichannel fluorometer’s rapid detection capability. The multichannel fluorometer integrated with ML analytics is capable of automating data analysis and determining accurate and rapid bacterial detection on-site with the prediction of bioburden levels and differentiating bacterial strains and the protocol can be applied to the biosensors with a similar data type.

Seamless recovery and sustained power to critical infrastructures (CIs), after grid failure, is a crucial need arising in disaster scenarios that are increasingly becoming more frequent. Accreditation standards recommend CIs to have emergency power supply system (EPSS) in order to form a local microgrid network with backup resources (generation units/renewable resources) in case of sudden power blackouts of main grid supply. The standards also recommend enhancements such as seamless transition capability during on-grid to off-grid jump and vice versa, efficient and optimal operation while maintaining the standard regulation and power quality limits during power contingencies, and operation with sustenance of reliable power for long periods of time. This article is proposing a comprehensive design of the EPSS for uninterrupted operation of CIs by employing novel techniques, such as $\mathrm {1)}$ mode-dependent droop controlled grid-forming inverters for seamless transition capability; $\mathrm {2)}$ fast-acting optimal net-load management engine for efficient and optimal operations maintaining regulation and power quality limits; $\mathrm {3)}$ optimization-based horizon of viability engine for longer, sustained and viable operation in the aftermath of grid failure as recommended in standards. Two stage validations are conducted using defined test cases while considering various types of contingencies on a 3- ${\phi }$ , 60 Hz, 480 V, 450 kVA test system based on a realistic electrical network of a commercial scale medical center. The comprehensive controller-hardware-in-loop and power-hardware-in-the-loop based experiments corroborate the efficacy and robustness of the proposed strategy for EPSS employed in the CI. The mode-dependent droop controller enables seamless transition of the critical infrastructure between the off-grid and the on-grid modes. Inverters operate in the gird-forming mode when the CI is off-grid and also when the CI is on-grid while keeping frequency and voltage regulated between ±0.5 Hz and $\pm 10\%$ of the nominal respectively. The fast-acting net-load management engine maintains the balance in total available generation and the power demand by determining the optimal decisions within 100 ms during various contingencies while regulating the voltage, frequency, and total harmonic distortion (THD) of the system under the strict limit recommended by the grid codes. The horizon of viability optimization engine showcases advantages over conventional scheduling engines with higher utilization of critical loads and maintaining higher energy at the end of an 1-hour experiment.

In this paper, a heterogeneous communication system capable of delivering 5G/sub-terahertz signal carriers over an arbitrary long fiber and separated transmission links is presented by employing direct detection, multiplexing techniques, and advanced digital signal processing. In this experiment, the 3.5 GHz and 28.5 GHz carrier frequencies, representing 5G links, deliver 4 Gb/s 16-QAM OFDM signals to separate user ends over a 1-meter wireless link distance. Later, the sub-terahertz wireless communications of 4 to 10 Gb/s QPSK and 8-QAM signals with varying carrier frequencies of 125-,175- and 225 GHz, over wireless distances (< 80 cm) are presented and evaluated. The results indicate that by increasing optical power from 12 dBm to 13 dBm, the bit error rate decreases two orders of magnitude. Eventually, with the assistance of artificial intelligence, a nonlinear equalizer (AI-NLE) prototype is introduced. The results indicate that the AI-NLE successfully decreases the number of errors in received data by one order of magnitude. The proposed heterogeneous system is compatible with radio-over-fiber technology, cost-effective, and easy to deploy, making it a promising candidate for indoor terahertz communication.

We address the problem of access point (AP) placement in small-cell networks with partial infrastructure flexibility, i.e., a novel class of problem in Beyond 5G, resultant from the utilization of unmanned aerial vehicles (UAVs) with AP functionalities (UAV-APs), to aid fixed wireless networks in coping with momentary peak-capacity requirements. We use the signal-to-generated-interference-plus-noise ratio (SGINR) metric as an alternative to the traditional signal-to-interference-plus-noise ratio (SINR) to quantify the effects of inter-cell interference (ICI) on the per-user capacity. From average SGINR, we derive the ICI-aware distortion measure leading to the Inter-AP Lloyd algorithm to obtain throughput-optimal AP placement for a fully flexible infrastructure. We then impose a hybridity constraint to the AP placement problem which turns a fraction of the network into a fixed infrastructure composed of terrestrial APs (T-APs) while the remainder is constituted by UAV-APs with flexibility in position. This newly formulated AP placement problem is solved by the proposed Lloyd-type algorithm called Hybrid AP Placement Algorithm (HAPPA). Furthermore, we present an initialization method for the Lloyd and Lloyd-type algorithms for Gaussian mixture models (GMMs) that offers an AP allocation leading to a higher rate compared to the k-means++ initialization. Finally, computer simulations show that the Inter-AP Lloyd algorithm can improve the performance of the worst users by up to 42.75% in achievable rate, assuming a fully flexible network. By using HAPPA on hybrid networks, we achieve improvements of up to 71.92% in sum rate over the fixed network and close the performance gap with fully flexible networks down to 2.02%, when an equal number of UAV-APs and T-APs is used. Further, our proposed initialization scheme always results in a balanced AP allocation, which means a more even distribution of users per AP, whereas the k-means++ scheme results in unbalanced allocations at least 30% of the time, resulting in a worse minimum rate.

A bi-directional power exchange between the plug-in electric vehicle (PEV) and the AC electrical grid is necessary to perform the Vehicle to Grid (V2G) and Grid to Vehicle (G2V) operations. While performing these operations, different power converters and controllers play an important role as mediators between the PEV and electric grid. Various works have demonstrated the utilization of controllers for PEV’s battery power management. However, the existing conventional controllers have technical shortcomings about vulnerability to controller gain, accurate mathematical modelling, poor adaptability, sluggish response to a sudden outburst and lengthy interval execution processing. Therefore, this paper develops an adaptive neuro-fuzzy inference system (ANFIS) control strategy based bidirectional power management scheme to ensure the optimal electrical power flow exchange between the AC electrical grid and battery storage system in PEVs. This paper aims to reduce the stress on the grid power side and utilize the unused power properly. The performance of the ANFIS model is varied using two PEVs based on real-life power consumptions by different loads at home based on five operational modes. Besides, a comparative analysis between the ANFIS controller and the PI controller is carried out to demonstrate the effectiveness of the proposed control scheme. The results illustrate that the proposed ANFIS controller delivers a smoother power injection from the PEV to the AC power grid with the least harmonics as well as achieves a smoother battery profile and less distortion when power is absorbed by PEV battery.

Colorectal cancer (CRC) is the third most deadly cancer worldwide. The use of gut microbiome in early detection of the disease has attracted much attention from the research community, mainly because of its noninvasive nature. Recent achievements in next generation sequencing technology have led to increased availability of sequence data and enabled an environment for the growth of gut microbiome research. The use of conventional machine learning algorithms for automatic detection of CRC based on the microbiome is limited by factors such as low accuracy and the need for manual selection of features. Despite their success in other fields, Deep Neural Network (DNN) algorithms have limitations in microbiome-based CRC classification. These limitations include high dimensionality of microbiome data and other characteristics associated with sequence data such as feature dominance. In this paper, we propose a feature augmentation approach that aggregates data normalization methods to extend existing features of a dataset. The proposed method combines feature extension with data augmentation to improve CRC classification performance of a DNN model. The proposed model obtained area under the curve (AUC) scores of 0.96 and 0.89 on two publicly available microbiome datasets.

In the above article [1], the acknowledgment for the research grant was referencing an incorrect grant reference number.

We experimentally demonstrate pilot aided compensation of relative phase noise (RPN) in a distributed Raman amplified coherent fiber optic system with forward pumping. This technique effectively eliminates the error floor that arises when conventional phase noise compensation algorithms track RPN, especially in the context of higher order modulation formats and low baud rate signals. Extrapolating the experimental results, we also demonstrate through measurement and simulation, the effectiveness of pilot aided RPN compensation in a 25 Gbaud, 16-QAM modulated digital-subcarrier system with 2.5 Gbaud and 5 Gbaud subcarriers.

For their capability of electronic dispersion compensation, transmission systems based on direct detection of single-sideband (SSB) signals are attractive candidates as energy-efficient and cost-effective alternative solutions to intradyne digital coherent systems for interdata center and metro applications. The Kramers-Kronig (KK) receiver scheme has been shown to provide superior performance compared to other schemes in signal-to-signal beat interference (SSBI) cancellation in these direct-detection systems. In this paper, we propose a low-complexity and cost-effective scheme of generating an optical superchannel comprising multiple SSB channels, based on a single quantum-dot mode-locked laser source. The proposed system does not require additional photonic or RF components at the transmitter to generate the required SSB signal with a continuous wave (CW) carrier. It also preserves the full digital-to-analog converters' bit resolution for data modulation, in contrast to other methods based on digital generation of the CW component. Simulations of system performance with KK receiver, based on measured laser output field, show that the proposed system can achieve bit-error ratio below the hard-decision forward error correction threshold for 16-QAM Nyquist SSB signals after transmission through three amplified spans of single-mode fiber in a 240-km link. Using 8 KK channels at 23 GBaud each, the proposed scheme will be able to achieve a transmission rate of 736 Gb/s with noncoded spectral efficiency of 2.45 b/s/Hz. The impacts of carrier-to-signal power ratio, per channel launch power into the fiber, and component frequency drifting on transmission system performance are also discussed.

We propose a simplified digital phase noise compensation technique for a Nyquist pulse-shaped digital subcarrier multiplexed (DSCM) coherent optical transmission system, employing an optical frequency comb based on Quantum dot passive mode-locked laser (QD-PMLL). Our results show that the impact of dominant common mode phase noise can be efficiently compensated at the receiver by digitally mixing the data sideband with the complex conjugate of the residual carrier component. This digital mixing technique resulted in better bit error rate performance compared to the conventional mth power Viterbi-Viterbi algorithm for QPSK and blind phase noise compensation for 16-quadratic-amplitude modulation formats, especially in the presence of large phase noise. To this end, exploiting the mutual coherence between the mode-locked comb lines of QD-PMLL, we numerically demonstrate its potential applicability as a transmission source for coherent optical superchannel transmission.