The advances in artificial intelligence (AI) technology can transform education. However, the growing infusion of AI technologies into academic environments raises important ethical issues that are essential for safeguarding academic integrity and quality. This paper proposes a detection framework that utilizes linguistic features and a Bidirectional Gated Recurrent Unit (Bi-GRU) model to identify AI-generated texts. The framework extracts perplexity values, readability measures, syntactic complexity metrics, and lexical diversity indicators, which are fed into a Bi-GRU classifier. Trained on an extended version of the DAIGT dataset and evaluated on the Deepfake dataset, the model achieved an accuracy of 98 % with F1 score of 97 % when tested on the DAIGT dataset. It also achieved an accuracy of$\mathbf{7 2 \%}$and an F1 score of$\mathbf{7 9 \%}$on the Deepfake dataset, outperforming state-of-the-art methodologies in these datasets. These findings highlight the potential of combining linguistic feature analysis with deep learning to develop efficient, interpretable, and domain-adaptive systems for AI text detection, addressing the critical need for automating authenticity and maintaining integrity in content generation.

Unmanned Aerial Vehicles (UAVs) are increasingly utilized across diverse industries, yet their communication protocols remain vulnerable to cyber threats and performance bottlenecks. MAVLink, the widely adopted standard for UAV communication, offers lightweight messaging but lacks robust security measures. This paper introduces eMAVLink, an enhanced protocol that integrates advanced cryptographic mechanisms to improve security and efficiency. We conduct a comparative evaluation of hashing and encryption algorithms utilized in eMAVLink, assessing the performance on both cryptographic co-processors and general-purpose CPUs. Our analysis examines end-to-end encryption, mutual authentication, replay attack prevention, and message integrity verification, quantifying their computational overhead across hardware-accelerated and software-based implementations. Additionally, we explore the trade-offs between security strength and real-time performance in dynamic UAV environments. The results highlight eMAVLink’s ability to achieve enhanced security while maintaining low computational overhead through optimized cryptographic offloading. This study provides key insights into the feasibility of deploying secure and efficient UAV communication protocols across diverse hardware architectures.

Federated Learning (FL) allows healthcare organizations to train models using diverse datasets while maintaining patient confidentiality collaboratively. While promising, FL faces challenges in optimizing model accuracy and communication efficiency. To address these, we propose an algorithm that combines feature selection with synthetic data generation, specifically targeting medical datasets. Our method eliminates irrelevant local features, identifies globally relevant ones, and uses synthetic data to initialize model parameters, improving convergence. It also employs a zero-trust model, ensuring that data remain on local devices and only learned weights are shared with the central server, enhancing security. The algorithm improves accuracy and computational efficiency, achieving communication efficiency gains of 4 to 14 through backward elimination and threshold variation techniques. Tested on a federated diabetic dataset, the approach demonstrates significant improvements in the performance and trustworthiness of FL systems for medical applications.

The emergence of neuromyths, or false beliefs about brain function and learning, has been a significant challenge in the field of education. These myths often hinders the learning process. Our study delves into this challenge by combining the cutting-edge large language models (LLMs) and humanoid robotics, aiming to dispel neuromyths in an educational setting. By integrating OpenAI's advanced Generative Pre-trained Transformers (GPT) model with a humanoid robot. This study focuses on neuromyth dispelling to investigate the integration of advanced large language models with humanoid robotics in education. We developed a system combining a Pepper robot with OpenAI's GPT model to educate users about common neuromyths. The system was evaluated through interactive sessions, where our queries were assessed on relevance, context, accuracy, response time, and speech recognition fluency. The findings reveal that the system effectively delivers relevant, contextually appropriate, and accurate information. Although response time was moderately rated, it did not significantly impact the overall user experience. The speech recognition system showed high efficiency, ensuring smooth interactions. These results demonstrate the potential of AI-humanoid robot integration as a valuable educational tool, especially for neuromyths. Our research contributes to educational technology by highlighting the effectiveness of such integrations in enriching learning experiences.

Early treatment is required to control the symptoms and serious complications caused by schizophrenia (SZ). People suffering from SZ require lifelong treatment. The use of machine learning (ML) models to detect various health problems such as SZ has received considerable attention from researchers in recent years. This study investigated the effectiveness of various ML models to detect and predict SZ using electroencephalogram data. A dataset of 14 healthy schizophrenic patients was used, and 12 features were extracted after applying independent component analysis. Three traditional ML models (logistic regression, support vector machine, and K-nearest neighbors) and a convolutional neural network (CNN) were trained, and their performance was compared. Results demonstrated that the CNN model outperformed the other three models with the highest accuracy score of 95% on validation data. Our results highlight the potential of using ML in the early detection and prediction of SZ, which can help in timely and effective treatment.

The brain-computer interface (BCI) and eye-tracking technologies can potentially improve the learning environment in education. Cognitive BCIs can give a deep knowledge of brain functioning, enabling the creation of more effective learning approaches and improving brain-based abilities. This study proposes a neurofeedback strategy based on BCI and eye-tracking to collect factual data (monitoring students' brainwaves and eye movement) and analyze their cognitive capacities during online learning. This study aims to create patterns regarding students' learning behavior based on brain and eye movement responses to learning activities as part of the learning environment. As a result, teachers may adapt to new pedagogical ideas and a more flexible delivery style.

Education is essential for achieving many Sustainable Development Goals (SDGs). Therefore, the education system focuses on empowering more educated people and improving the quality of the education system. One of the latest technologies to enhance the quality of education is Artificial Intelligence (AI)-based Machine Learning (ML). As a result, ML has a significant influence on the education system. ML is currently widely applied in the education system for various tasks, such as creating models by monitoring student performance and activities that accurately predict student outcomes, their engagement in learning activities, decision-making, problem-solving capabilities, etc. In this research, we provide a survey of machine learning frameworks for both distributed (clusters of schools and universities) and centralized (university or school) educational institutions to predict the quality of students' learning outcomes and find solutions to improve the quality of their education system. Additionally, this work explores the application of ML in teaching and learning for further improvements in the learning environment for centralized and distributed education systems.

The outbreak of the coronavirus disease (COVID-19) has had a profound impact on education worldwide. The rise of remote learning is one of the most significant changes in this regard, as many schools and universities were forced to close down by regional health authorities. This has also caused people to become more conservative in trade-offs between healthcare and education. Google Trends is the most common tool for analyzing online search behaviors. It is a free resource that provides information on the trends and changes in users' online interests over time based on certain terms and subjects. The online search queries on Google can be used to assess users' behaviors concerning online learning to forecast their choices regarding online education. This paper examines the frequency of users' web searches for online communication tools, courses, and learning terms. We statistically compared users in the Middle East and North Africa regions by using the volumes of searches recorded on Google Trends from January 2016 to August 2022. Moreover, we used machine learning techniques to identify differences among the keywords used. The findings statistically show that COVID-19 has led to an increase in the extent of students' attention to and interest in online learning.

This paper discusses neurodevelopmental disorders and their effects on the brain and nervous system. Autism, attention deficit hyperactivity disorder (ADHD), dyslexia, and cerebral palsy are just a few examples of the many conditions classified as neurodevelopmental disorders. These conditions can potentially influence wide-ranging facets of a person's life, including communication skills, behavior, ability to learn (and thus overall educational achievement), and motor skills. Hereditary and environmental variables are thought to contribute to the development of various developmental disorders, although the exact causes are unknown. The effects of neurodevelopmental disorders on a student's educational opportunities are highly variable according to the particular condition and specific requirements of the individual. Students with dyslexia, for instance, may have trouble reading and writing, whereas those with ADHD may have difficulty paying attention and controlling their impulses. Autism, meanwhile, is characterized by social interaction and communication difficulties. Students with neurodevelopmental problems need to be identified and treated as soon as possible to obtain the appropriate assistance and accommodations to facilitate success in their educational endeavors. In the classroom, teachers have a significant opportunity to recognize and respond to the needs of particular students on an individual and group level. Brain-computer interfaces (BCIs) are an exciting new field of study with potentially substantial ramifications for how persons with neurodevelopmental problems fare in educational settings. In particular, BCIs could help students with neu-rodevelopmental problems to enhance their cognitive function, ability to communicate, and motor skills. Accordingly, this paper investigates the potential applications of BCIs for people with neurodevelopmental disorders. Specifically, we conducted a systematic review and a meta-analysis of the literature indexed in four digital databases. The findings of this investigation may lead to major insights and conclusions that, if implemented, may improve the educational outcomes for individuals affected by neurodevelopmental disorders

Unmanned aerial vehicles (UAVs) have been more prevalent over the past several years with the intent to be widely deployed in many industries, including agriculture, cinematography, healthcare, delivery, and disaster management missions due to their ability to provide real-time situational awareness. However, various limitations such as the battery capacity, the charging method, and the flying range make it difficult for most applications to carry out routine tasks in vast areas. In this paper, a deep reinforcement learning (DRL) method for multi-UAV path planning that considers a cooperative action amongst UAVs in which they share the next destination to avoid visiting the same location at the same time. The Deep Q-Network algorithm (DQN) enables UAVs to autonomously plan their fastest path and ensure the continuity of the mission by deciding when to schedule a visit to a charging station or a data collection point. An objective function with a tailored reward is designed to maintain the stability of the model and ensure the quick convergence of the model. Lastly, the proposed strategy has been demonstrated by the experiments on different scenarios and showed its effectiveness in ensuring the continuity of the mission with a fastest path possible.

As security becomes a strong factor in daily activities, finding secure ways to unlock machines and smartphones is a challenge due to hardware limitations and the high risk of hacking. Considering the level of security and privacy in the digital world, attackers tend to be one step ahead. Therefore, this technical paper introduces a brain-computer interface (BCI) for increasing subject-based security using unique biometric features as a solution to build complex passwords. The BCI measures brain changes and extracts relevant bio-features from each subject using non-invasive electroencephalogram (EEG) tests. The proposed system allows users to gain access to their devices using brain waves (bypass) instead of inserting their password manually (normal path), which saves the user time and upgrades the level of privacy as no physical actions are required during this process. This system is also well suited for individuals with mobility impairments. We used the P300-based BCI controlling paradigm which depends on reading the electrical brain activity of the user when observing a particular object. The other feature of the system is that it can extract unique features of each individual brain to produce a network that uniquely identifies them, which is used as a security layer. Users need to enter their unique network to access their device with failed attempts requiring an EEG test to identify the user. The system plays an active role in facilitating user processes for authentication while accessing devices. The system establishes an urgent call whenever the user’s brain currents command it to. The project outcomes were assessed by simulating the BCI before real-time implementation to determine errors and resolve the validity of the project scope.

Emotions play an important role in the health and well-being of humans. It is associated with feedback on human interaction with the surrounding environment, decision-making, and intelligence. Electroencephalography (EEG)-based brain-computer interfaces (BCI) technology can be used to sense the emotional state of humans. Therefore, this research introduces a non-invasive BCI system that provides solutions for psychiatrists to treat patients suffering from chronic sadness, depression, and anxiety without medications. Here, we propose an EEG-based neurofeedback system for decoding and modulating human emotions. This system decodes three emotions: happiness, sadness, and neutral emotions. From the decoded emotion, the system generates visual and auditory feedback to train the patient to regulate his/her brain activity to improve his/her mental health. We collected EEG data corresponding to each emotion from twelve female participants while watching multiple stimuli to develop a support vector machine (SVM) model with a radial basis function (RBF) kernel. The SVM model decoded the desired emotions with 92.3% accuracy. Then, EEG- Neurofeedback sessions decode the patient’s emotions in real-time and generate visual and auditory feedback using the decoded emotions.

Brain computer interface (BCI)-based virtual environment control has found broad applications in solving and pursuing factual healthcare issues concerning efficiency, safety, and costs. In this technical paper, an unmanned ground vehicle (UGV) robot with a simulator-equipped BCI system was utilized. The Gazebo simulator was employed to develop a simulated setting. The software CitySim World allowed rendering the simulated milieu more down-to-earth. A non-invasive electroencephalogram (EEG)-based BCI was used to follow the brain signals and extract the P300 component, a kind of simultaneous BCI controlling procedure for safe, fast, and inexpensive implementation. This UGV control system using human brain activity can be beneficial for the real UGV platform control. It enables the discovery of the probable errors in the physical implementation. All the steps implementing our BCI system were appropriately provided (data acquisition system, user interface design, BCI data architecture, ROS/ robot Jackal, and implementation and tests). Furthermore, the project implementation and some solutions to possible issues were posed. The project outcomes were assessed by employing BCI in a simulation; before implementing real-time, to determine errors and resolve the validity of the project scope.

Over the last few years, the use of unmanned aerial vehicles (UAVs) has grown, with the goal of being widely deployed in sectors such as deliveries, rescue operations, mining fields, patrolling, and monitoring. However, the limitations of the onboard battery capacity and the flying range pose a problem to most applications while performing daily tasks such as parcel delivery or aerial communications in large areas. This paper proposes a reinforcement learning method to compute optimal trajectories for a UAV, considering both visiting delivery locations and recharging stations. The use of wireless power transfer (WPT) technology allows UAV s to wirelessly recharge their batteries on the fly and therefore to extend their flying range further. In this scenario, we consider several WPT-enabled charging stations placed around the serviced area. The proposed approach leverages a reinforcement learning strategy, and the performance results obtained show its effectiveness in finding an optimal trajectory by minimizing the UAV's travel and service time.

Many advanced researches on natural user interfaces methods based on user-centered design have been using speech, gestures and vision to interact with environment and/or control internet of things (IoT) devices. Brain computer interfaces (BCIs) technology could make this interaction/control more natural, faster, and reliable, and effective. In this paper, we propose a decoding algorithm for controlling a drone in a three-dimensional (3D) space using steady state visually evoked potential (SSVEP)-based BCI modality. SSVEP-based BCI has the great potential for use in virtual reality environment, which enables the user to control the drone using his/her brain activity in an first-person-view mode. Therefore, the user will be in a full control over the flight using BCI system by commanding the drone to take off, land, go forward, stop, and turn right/left. This system yields a super convenient way for normal people with no prior experience to interact with the drone and control a flight mission in a little to no time, over traditional manual control which takes longer time to learn and perfect. in the decoding phase, a various convolutional neural networks (CNN) models were built to accommodate different control criteria such as the generality of the model. This proposed EEG-decode-pipeline has been implemented on an open-source data-set which consists of 8-channel EEG data from 10 subjects performing 12 target SSVEP-based BCI task. A high multi-class BCI classification results were achieved with an accuracy ranging around 80-90% for performing a successful online simulation of the drone control.

The assessment of course learning outcomes is an essential component in the continuous efforts of course improvement. The assessment is a tedious process and often incurs for many educators an overhead to the teaching and learning operation. Thus the need to investigate efficient methods to improve the process of course assessment by minimizing unnecessary efforts for the planning, preparation and execution of the assessment process. Automating the assessment process is instrumental in taking away its tediousness allowing teachers to focus their efforts on the improvement of the teaching and learning quality. For the case of information technology (IT) curriculum, one main concern is the difficulties encountered by students in learning programming skills; thus the need for an assessment-driven course improvement for programming courses. In this paper, we propose an automated proactive assessment method for assessing the learning outcomes of a course by embedding the assessment instruments in the tests and student homeworks. We selected a network programming course for its suitability to embed assessment instruments as part of the programming library used by students during their test and homeworks. The embedded instruments consist of a set of use-case routines to test the validity of each design and implementation component of the developed protocol. This approach streamlines the process of learning outcomes assessment as well as the continuous improvement of the course.

Remote photoplethysmography (rPPG) is a non-contact method to measure physiological signals, such as heart rate (HR) and respiratory rate (RR), from facial videos. In this paper, we constructed a central difference convolutional attention network with Huber loss to perform more robust remote physiological signal measurements. The proposed method consists of two key parts:1) Using central difference convolution to enhance the spatiotemporal representation, which can capture rich physiological related temporal context by gathering time difference information 2) Using Huber loss as the loss function, the gradient can be smoothly reduced as the loss value between the rPPG and ground truth PPG signal is closer to the minimum. Through experiments on multiple public datasets and cross-dataset evaluation, the good performance and robustness of the rPPG measurement network based on central difference convolution are verified.

Brain–computer interfaces (BCIs) are an emerging technology with applications for persons with disabilities as well as the able-bodied. In this paper, we present a new framework of cooperative BCI control system for robot–drone interaction using P300-based BCI. This system is aimed at supporting and assisting complex and cooperative multitask military applications. In our online experiments, a robot “BB-8” and a parrot drone are separately mind-controlled to execute cooperative tasks using noninvasive brain measurements. We use real-time electroen-cephalography (EEG) signals to drive cooperative mission-based tasks by exchanging control information between two BCI users. The proposed cooperative BCI system is based on controlling the mobile robot and drone using the P300 speller modality and exchanging mapped messages between these two wearable EEG-headset-based systems. Using the EEG Unicorn Hybrid Black equipment, we quickly construct an interface that contains predefined visual cues for robot movements to be selected by the first BCI user; these cues are sent online as commands to the robot through the JavaScript server code. Another interface has been constructed that contains visual cues that are to be converted to drone-movement commands through the Python server code. The robot will perform a ground survey while the drone performs an aerial survey, and the final task will be performed by mutual communication between them. This cooperative BCI application can be operated by two soldiers. For instance, the actions of a robot may indicate different signs for the drone to undertake specific actions. This novel BCI application is evaluated based on the ability of two users to send commands using their brain activity, as well as the capability of the control algorithm to receive, send, and map the commands between the drone and the BB8 robot to allow them to achieve their mission.

Paralysis is the most inhibiting among all the severe motor disabilities. Indeed, people are inflicted with paralysis as the result of an accident or a medical condition that affects - completely or partially, the way muscles and nerves function. However, these patients are cognitively aware, and their mental abilities are unimpaired, and can still be autonomous and more useful in many other ways than many able-bodied people. Brain-computer interface (BCI) technology is now being incorporated into the treatment of physically impaired patients offering them an improved mobility and thus autonomy. In this paper, we propose to develop a smart brain-controlled wheelchair with autonomous navigation system for people with severely impaired motor functions. Our proposed solution allows its users to move around in indoor premises with great flexibility and minimum instructions. That is, high-level commands such as “Go to location X” is enough for the wheelchair to move to the desired location while finding its way around obstacles and obstructions. This system relies on two main components: a BCI interface to issue high level commands to the wheelchair, and a component for autonomous indoor navigation system which integrates all the elements of path planning obstacle detection and avoidance. In addition, the solution relies on the use of trained models that are deployed in cloud and provided as facility specific services.

In this paper, we present a new way of controlling drone by using a P300-based brain-computer interface that supports the military field as assistive technology. The main idea is that the drone can be controlled by the soldier's brain activity using electroencephalogram (EEG) to chase other drones and discover hidden enemy areas. We assumed that we have two able-bodied users, the first role played by a user as a soldier aims to control the drone by using brain activity and the other role played by another user as an enemy aims to control manually the drone using Python program. This scenario allowed us to test the ability of chasing the enemy's drone. The results for this application was evaluated by the ability of the user to calibrate very well with the software and the ability of the program to receive and send commands using EEG signal to the drone for execution.

Swarm technology for autonomous unmanned aerial vehicles (UAVs) has gained popularity in the last few years due to their potential for civilian and military applications. Intelligent swarm systems are very efficient at solving group-level problems and their capability to accomplish complex missions with no or little human intervention. One of the most challenging problems is operating in environments with surrounding obstacles such as buildings, thus, often obstructing inter-UAV communication. A UAV swarm is required to maintain continuous communication between its members and preserve the stability of its formation while flying towards an ultimate goal. In this paper we propose MSCS, a new cooperative and adaptive scheme for multi-clustered autonomous UAV swarms. This schemes allows several UAVs operating in a swarm formation to coordinate their navigation and path planning operations by using an adaptive multi-clustered leader-follower approach. The swarm members follow a dynamically elected leader based on the UAV with best fit to lead the swarm towards the destination. The coordination of the UAVs is achieved through SBP (Swarm Broadcast Protocol), a single-hop broadcast UAV-to-UAV (U2U) communication protocol, which allows swarm members to exchange information about their current location and their local cluster leader. We present a set of performance results, which show that this scheme contributes efficiently at maintaining the swarm formation's stability at acceptable density and disconnection ratio.

This paper describes a cybersecurity framework for protecting brain computer interface (BCI) technology. This framework consists of cybersecurity risk scenarios related to user safety/privacy and best practices to manage them. This framework provides solutions for privacy and safety issues of the existing noninvasive BCIs (e.g., electroencephalography (EEG)-based BCI). We chose to design a P300-based BCI application because it is the most popular modality, simulate some common cybersecurity attacks, and find a relevant solution to protect the user and/or integrated EEG hardware-software system. In this paper, we describe how cybersecurity risks could affect BCI form streaming/recording EEG signal in real-time until sending commands. We used EEG Equipment for measuring brain activity and Python programing language to build our experimental paradigm, record EEG signal, classify P300 components, send a message to another user, simulate some attacks, and find perfect solutions for assuring high BCI protection. This paper gives an overview of the framework, some description of BCI hacking challenges and their impact on BCI users as well as a preliminary demonstration of a P300-based BCI system with two common simple attacks.

The present study describes neuromagnetic Geminoid control system by using single-trial decoding of bilateral hand movements as a new approach to enhance a user's ability to interact with a complex environment through a multidimensional brain-computer interface (BCI). Two healthy participants performed or imagined four types of bilateral hand movements during non-invasive magnetic field measurements to control a human-like robot (Geminoid HI-2) in real-time. By applying a nonlinear support vector machine (SVM) method to classify the four movements regarding magnetoencephalography (MEG) sensors obtained from the sensorimotor area, we found the mean accuracy of a 2-class classification using the amplitudes of neuromagnetic fields to be particularly suitable for real time control applications, with accuracies comparable to those obtained in previous studies involving unilateral hand movement. Moreover, our results demonstrated that decoding bilateral movements in real-time is a promising option to design multidimensional-control based BCI applications.