Cognitive-state assessment in real-time is a key part of adaptive learning, neuroergonomics, and safety-critical systems. Electroencephalography (EEG) is helpful in monitoring neural oscillations, as it allows fast monitoring of changes in attention and working memory status of neural fluctuations that can respond to changes in attention and working memory. In this paper, an interpretable EEG-based framework is proposed to use sensitive fronto-temporal and temporo-parietal measures of dynamic working-memory volumes, profile attentional functions of target regions, and a transparent fuzzy-logic mechanism with detailed case-study analysis to represent the perception of cognitive status. EEG signals from bipolar pairs F7-T3 and F8-T4 capture verbal and visuospatial working-memory activities, and T4-T6 monitors attentional engagement across multimedia and text-type activities. The dynamic weighting model generates a composite working-memory index which is integrated with the attention measure to generate a continuous cognitive state index (CSI) indicative of passive, moderate and focused engagement. Case studies depict temporal EEG behavior, working memory shifts, changes in attention and the respective CSI trajectories. This indicates clear neurophysiological patterns, with pronounced right fronto-temporal fluctuation for high cognitive load, stable temporo-parietal activity for memory maintenance and consistently lower level steady activity on the left frontal regions.

Fast and accurate algorithms are essential for processing EEG signals effectively. However, the high sampling rate of EEG data generates a vast number of data points, posing challenges for predicting working memory capacity in cognitive studies. Therefore, to resolve this issue and increase the accuracy and reliability of the model, cost-benefit analysis has been used. It focuses on minimizing classification error and maximizing accuracy. The work also focuses on analyzing neural activity through EEG recordings to predict cognitive performance, particularly using tasks like the N-back, a widely used paradigm for Working memory (WM) studies. Write full form(WMACBA) understands how attention is allocated based on the relative costs and benefits associated with focusing on specific information in working memory. Participants' EEG signals were recorded while performing tasks that gradually increased in difficulty to capture the varying cognitive load. Functional data analysis has been used to handle the high-dimensional EEG data effectively. A key obj ective is to develop predictive models that link EEG signal patterns with WM abilities, offering a non-invasive way to estimate cognitive performance. This approach aims to strike a balance between the complexity and practical utility of the predictive model, potentially making it suitable for applications like early cognitive impairment detection and personalized.

Intrusion Detection Systems (IDS) are crucial to secure cloud-based infrastructures from ever-evolving cyber-attacks. However, traditional IDS models struggle with issues such as redundant features, ineffective attack classification, and outdated datasets, limiting their adaptability to emerging cyber threats. This paper presents FSEGM (Feature Selection and Ensemble Generative Model), an advanced IDS framework that integrates both feature selection and ensemble generative models to improve the efficiency and accuracy of detection. This proposed framework follows a multi-stage methodology: first, a stepwise forward and backward elimination algorithm is employed for feature selection, reducing dimensionality while preserving essential information. Next, an ensemble classification approach is applied, combining several methods, including Decision Tree (J48), Random Forest, Random Tree, and Naïve Bayes with Attribute Penalization (NBPA). A generative learning technique is then used to merge probability distributions from base classifiers, improving attack recognition. The final response phase categorizes threats into misuse-based and anomaly-based activities for proactive mitigation. Experimental investigations for the proposed model were carried out with the CIC-IDS2017 dataset, a benchmark for network intrusion detection. The results demonstrate that the proposed FSEGM method significantly outperforms state-of-the-art models across multiple evaluation metrics. The Random Forest classifier achieved 98.99

With the quick advancement of wireless communication technologies and the rising demand for spectrum resources, integrating learning and reasoning capabilities into cognitive radio networks (CRN) has become essential. This study explores the spectrum sensing capabilities of Secondary Users (SUs) within a CRN. Various supervised machine learning (ML) techniques, including Bayes Network, Naive Bayes, and Decision Tree, are employed to assess the performance of the Secondary Users in spectrum sensing. These algorithms are used in order to predict the trustworthiness of Secondary Users by considering sensing reputation values altogether. It is a topic of discussion, to identify the most effective ML algorithm that can guarantee the maximum level of system accuracy & efficiency. In this study, the open-source data mining tool WEKA is used to create a wide range of classification models for the evaluation of performance and correct prediction of trustworthy users. The best possible model accuracy is attained with the aid of Receiver Operating Characteristics (ROC) curves and cost-benefit analysis of three different classifiers. The evaluation of SUs’ trustworthiness and spectrum sensing repute in CRN concludes that the Decision Tree classifier provides best performance which enables the system for correct classification among malevolent users, suspicious users, and honest users.

The centralized nature of software-defined networks (SDN) makes them a suitable choice for vehicular networks. This enables numerous vehicles to communicate within an SD-vehicular network (SDVN) through vehicle-to-vehicle (V2V) and with road-side units (RSUs) via vehicle-to-infrastructure (V2I) connections. The increased traffic volume necessitates robust security solutions, particularly for Sybil attacks. Here, the attacker aims to undermine network trust by gaining unauthorized access or manipulating network communication. While traditional cryptography-based security methods are effective, their encryption and decryption processes may cause excess delays in vehicular scenarios. Previous studies have suggested machine learning (ML) like AI-driven approaches for Sybil attack detection in vehicular networks. However, the primary drawbacks are high detection time and feature engineering of network data. To overcome these issues, we propose a two-phase detection framework, in which the first phase utilizes cosine similarity and weighting factors to identify attack misbehavior in vehicles. These metrics contribute to the calculation of effective node trust (ENT), which helps in further attack detection. In the second phase, deep learning (DL) models such as CNN and LSTM are employed for further granular classification of misbehaving vehicles into normal, fault, or Sybil attack vehicles. Due to the time series nature of vehicle data, CNN and LSTM are used. The methodology deployed at the controller provides a comprehensive analysis, offering a single- to multi-stage classification scheme. The classifier identifies six distinct vehicle types associated with such attacks. The proposed schemes demonstrate superior accuracy with an average of 94.49% to 99.94%, surpassing the performance of existing methods.

The security and privacy of data are major concerns in the mobile crowdsensing (MCS) environment due to the huge amount of heterogeneous data received from various users and devices automatically or manually regarding their surrounding environment. User participation in the MCS approach is highly essential to have a vast dataset for analysis that will provide the required information or beneficial solution for society. However, it is difficult to achieve due to huge energy consumption, the need for internet connectivity for data transmission, and the security and privacy of data. Therefore, it is essential to have a network coverage model in which data transmission can be done with minimal energy consumption and the need for internet connectivity can be removed from the user’s side. The user’s sensitive data needs to be protected from internal and external attackers to improve the efficiency of the solution provided by the MCS environment with genuine data. This work is based on data collection from users based on their experience for a certain location using the hybrid network coverage model based on clustering, in which each location may have just one or multiple heterogeneous cluster heads. Discrete event-based CrowdSenSim Simulator has been used to design a simulation environment in urban spaces in which 2000 users will move to any location randomly among considered 40 locations and provide feedback data for the location. In this paper, a novel security mechanism based on multiple heterogeneous cluster heads per location has been presented, and it provides better security against attackers than the security model with one cluster head per location. The proposed multiple-cluster heads per location (MCHL)-based mechanism has been compared with the vulnerable one-cluster head per location (OCHL)-based mechanism on the basis of the average number of rounds attackers attacked, average number of locations attackers attacked, average coverage and average efficiency of attackers, and average efficiency of system security.

The content-based order of urban sound classes is an important aspect of several emerging techniques and applications as a result, the research issue has gotten a lot of attention recently. The goal of this work is to develop an effective Machine Learning(ML) based approach for urban sound categorization in cities for intelligent object detection and recognition through sound for environmental study. Audio samples from urban space has been classified through various ML algorithms based on Mel spectrogram technique. Where, mel spectrogram techniques is a frequency based approach for extraction of features from audio samples for classification. The Urban Sound dataset, which contains a total of 8732 number od sound samples belongs to ten different classes, has been used in this work. This may be used for smart object detection through sound coming from objects in different applications, such as Traffic Management, Agriculture, Smart city, etc.

A smart wheelchair provides mobility assistance to persons with motor disabilities by processing sensory inputs from the person. This involves accurately collecting inputs from the user during various movement activities and using them to determine their intended motion. These smart wheelchairs work by collecting brain signals in the form of electroencephalography (EEG) signals and by processing them into a quantized format to provide movement assistance to people. Such systems can be referred to as brain–computer interface (BCI) systems that work with EEG signals. Acquiring data from human beings in the form of brain signals through EEG, along with processing of those signals and ensuring the correctness of actions instigated by those brain signals involve a huge amount of data. In this work, we carried out an experiment by taking 100 human subjects and recording their brain signals using a NeuroMax device. Typical wheelchairs are constrained by design as the motion of those is limited either by manual operation or controlled by haptic sensors and actuators. The main objective in this work was to design a wheelchair with better usability and control using machine learning-based knowledge, which is typically a data-driven approach. However, the proposed approach was designed to take inputs from human gestures and brain sensory activities to provide better usability to the wheelchair. The attention meditation cost–benefit analysis (AMCBA) proposed in this paper aims to reduce the risk of inappropriate results and improve performance by considering various cost–benefit parameters. The said classifier aims to improve the quality of emotion recognition by filtering features from EEG signals using methods of feature selection. The operation of the proposed method is described in two steps: in the first step, we assign weights to different channels for the extraction of spatial and temporal information from human behavior. The second step presents the cost–benefit model to improve the accuracy to help in decision-making. Moreover, we tried to assess the performance of the wheelchair for various assumptions and technical specifications. Finally, this study achieves improved performance in the most difficult circumstances to provide a better experience to persons with immobility.

A smart wheelchair can be perceived as a wheelchair that can provide autonomous mobility and features to persons who are incapable of using their own body for self-mobility. There are broadly two categories of smart wheelchairs. In the first category, there are provisions for the usage of wheelchairs using in-built functions built into the wheelchair. In the second category, the smart wheelchair can use the signal obtained from human brain using Brain–Computer Interface (BCI), for the movement of human being who are incapable of using their own body for the various functions of a wheelchair. It has been substantial research on providing accuracy for the movement of smart wheelchairs. In this regard, there can be significant differences between the accuracy received from the user-supplied dataset and from the operational aspect of the wheelchair when put to the real operational environment. We have hence tried to improve the accuracy of our model by incorporating a cross-validation approach to our model. Cross-validation is a statistical method for evaluation of machine learning methods by segregating two different sections where the first section is used to develop and learn a model using training dataset. Further, it has been validated using testing dataset. Finally, test the stability of our machine learning model with real-time dataset.

Emotions play a significant part in a person's social connections, decision-making, and perception of the world. Elicited emotions cause a change in a person's physiological and psychological states. As Electroencephalography (EEG) facilitates a close study of brain activity, it is becoming a standard method among the research community for reliable recognition of human emotions. This work demonstrates various advancements in emotion recognition utilizing EEG signals and points out major changing trends by making a comparison of previously available research in this field. In addition to the survey a detailed explanation of the procedure for refining EEG for emotion recognition has been explained in this work. This aims to help researchers, especially beginners, have a thorough understanding of the developmental research in this field.

Cloud computing has been of greater convenience in recent years owing to its flexibility and on-demand service availability. Cloud computing services allow multiple users to access the available resources concurrently. So, trust is a key parameter from every client’s point of view. While accessing the resources in the cloud security becomes the primary concern for every participant and cryptography provides a secure room for every client. The cryptographic solution can be achieved in two different ways such as the conventional mechanism and the homomorphic encryption (HE) mechanism. The main objective of these two solutions is to maintain the confidentiality, integrity, and availability (CIA) of the information and resources stored in the cloud storage. The conventional mechanism provides a secure means to encrypt and decrypt the information at the sender and receiver sides respectively. But for any kind of operation, the information needs to be decrypted which increases the computational overhead which shows the path to the HE. In HE, the user can directly perform any kind of operation on the encrypted data which can help to reduce the computational time and cost to a much more significant value. In this current research work, the main focus is on the HE. Several HE mechanisms such as the RSA, Elgamal, Paillier, and DGHV have been implemented and the performance has been analyzed in terms of computational time. The empirical analysis demonstrates that the RSA homomorphic algorithm shows an average encryption and decryption time of 138.5 and 206.25 s which is lower as compared to others. Considering average CPU utilization, RSA shows around 27% with an average of 30 MB memory utilization which is found lower as compared to other algorithms.

Mobile Crowdsensing (MCS) is a major source of a vast dataset containing heterogeneous types of data collected from various sources and stored in the local or remote server.Proper analysis of MCS data helps in better decision-making.However, MCS data suffers from data integrity issues, such as validity, accuracy, and reliability, that affect decision-making.Therefore, ensuring data integrity in the MCS environment is essential as it is a major source of a huge dataset.The proposed work considers user review data collection and analysis using a mobile application developed for the purpose.To ensure the data integrity, identification of fake and invalid reviews in the dataset need to be determined.This work proposes two approaches to solve data integrity issues.The first approach is to detect and eliminate fake/ invalid reviews from the dataset.The second is to identify the sources of fake/ invalid reviews and block them to protect the dataset from future fake reviews.Machine learning (ML) models are proposed to solve these issues and to ensure data integrity by filtering out fake reviews from real-time data sets.The proposed model uses data fuzzification over a purely mathematical model that categorizes users or customers as honest, suspicious, or malicious and their reviews/ feedback as genuine or fake using ratings provided by the user in the MCS Environment.Using the developed mobile application, user can give feedback about the desired location through various devices, which is stored in a cloud platform.The dataset can be analyzed through a fuzzy logic-based mathematical model followed by an ML algorithm and cost-benefit analysis to detect genuine reviews for maintaining data integrity.Further accuracy of the proposed models is compared with popular ML algorithms such as Naive Bayes (NB), Bayes Net(BN), Support Vector Machine(SVM), Decision Tree(J48), and Random Forest(RF).Initially, it achieves 99.79% of accuracy using the Random Forest algorithm that has been enhanced to 100% using cost-benefit analysis in cross-validation mode.

In this paper, a trust framework is proposed for misbehavior detection in software defined vehicular networks (TFMD-SDVN) to detect the correct events in the network reported by the trusted or untrusted nodes. The trust value of a node is calculated based on rating, recommendation, and similarity. If the trust value is greater than a threshold, then the event reported by the event reporting node (ERN) is assumed to be correct. The performance of the proposed work is evaluated using OMNeT++ network simulator and SUMO traffic simulator in Veins hybrid framework. The performance parameters taken are True Positive Rate (TPR), False Positive Rate (FPR), Detection Time (DT), and Packet Delivery Ratio (PDR). Simulation results show that the proposed approach performs better than ART scheme, RPRep scheme, and BYOR scheme.

Software-defined networking (SDN) is an upcoming network model that emphasizes the separation of the control plane and the data plane, resulting in more flexibility, programmability, and network management. The problem of placement of controllers deals with the count of controllers that are needed and their location in the network to maintain the network structure. In a large network where multiple controllers are needed, the quantity and position of the controllers have a major impact on SDN’s performance and reliability. The network performance degrades significantly in case of controller failure. To avoid this situation, planning is required. A moth flame optimization algorithm is proposed for controller placement with planning, CPWP_MFO and compared with a genetic algorithm CPWP_GA. The aim is to place m controllers in the network to reduce the worst-case and average propagation latency considering controller failure and capacity constraint. The simulation results show that CPWP_MFO yields better results than CPWP_GA.

Background: In India, the population of elderly is predicted that it will be increased from 8% in 2015 to 19% in 2050. Geriatric population contributes around 9% of the total Odisha population and 86.3% of them reside in rural areas. The study aimed to estimate the prevalence of diabetes mellitus and find out its risk factors among rural geriatric population in Tigiria block of Odisha, India. Methods: This was a community-based study, cross-sectional in design among 725 rural geriatric populations of Odisha. Socio-demographic information was collected following the standard census of India operational definitions. Self-reported diabetes mellitus status was collected and classified as "present" or "absent". Statistical analysis was performed using "R version 4.0.4". Results: Among the total elderly, 88 (12.13%) participants were diagnosed with diabetes. Common factors found to be significant with diabetes were illiterates (AOR=0.32, CI=0.125-0.817), not working elderly (AOR=2.51, CI=1.103-5.723), high socioeconomic status (AOR=3.79, CI=1.351-10.632) and overweight elderly (AOR=2.19, CI=1.286-3.753) respectively. Conclusion: The frequency of diabetes mellitus among rural geriatric population is less but the risk is high among those not working, literate, with higher SES and overweight elderly group. The researcher should emphasize real-time diagnosis of blood sugar levels using standardized measures among the rural elderly population. Keywords: diabetes mellitus; geriatric population; rural; Odisha