Search

The next generation of wireless networks is poised to deliver dependable, extensive, and extremely low-latency mobile broadband communications across various industries. In this regard, reconfigurable intelligent surface (RIS)-aided non-terrestrial base station (NTBS) is crucial for developing three-dimensional (3D) networks that aim to integrate the terrestrial and non-terrestrial infrastructures. The integration of RISs with NTBSs has resulted in additional capabilities to attain faster and more flexible mobile coverage. In this paper, we employ a novel RIS-aided NTBSs with non-orthogonal multiple access in downlink heterogeneous networks (HetNets), in which the RIS is utilized to improve the transmission of signals from multiple NTBSs to ground users (GUs). We aim to maximize the system sum rate (SSR) and energy efficiency (EE) by mitigating inter-cluster and intra-cluster interference. The formulated optimization problem is non-convex due to the joint optimization of NTBS 3D positions, RIS reflection angles, RIS reflection coefficients, and successive interference cancellation among GUs. To address this problem, a modified gray wolf optimization -based meta-heuristic algorithm is proposed. In particular, the original optimization problem is divided into three sub-problems (i.e., RIS reflection angles, RIS reflection coefficients, and NTBS 3D position), which are then addressed alternately using the proposed optimization technique. The novelty of our proposed work is presented by reformulating the distance equations to minimize the distance between the wolves and their prey and enhance the search process’s precision. The simulation results stated that the suggested algorithm outperforms the traditional schemes in terms of both SSR and EE. Furthermore, the results show that incorporating RIS into multi-NTBS HetNets effectively boosts overall performance by improving the channel quality between NTBSs and their respective GUs while reducing inter-NTBS interference.

Diabetic Retinopathy (DR) is a prevalent outcome of diabetic mellitus. This causes lesions to form on the retina, impairing eyesight. Most likely, blindness can be avoided if the DR condition is discovered at an initial stage. Since DR is a non-reversible condition, early detection and treatment can significantly reduce the chance of visual loss. Fundus images manually detect DR, which is a laborious and error-prone procedure. In assessing and categorizing medical images, machine learning and deep learning have emerged as the most efficient methods, surpassing human performance, common image processing methods, and other computer-aided detection systems. For this article, the most recent approaches for utilizing fundus images to classify and detect DR using machine learning and deep learning methods have been researched and evaluated. The freely accessible DR Datasets consisting of fundus images have also been discussed. We reviewed several DR pipeline components, including the datasets that researchers frequently used and the preprocessing and data augmentation steps, feature extraction methods, commonly used detection and classification algorithms, and the generally used performance metrics. This paper ends with a discussion of current challenges that have to be tackled by researchers working in this field to translate the research methodology into actual clinical practice. Finally, we conclude with a discussion of the future perspectives of DR.

Goal: FetSAM represents a cutting-edge deep learning model aimed at revolutionizing fetal head ultrasound segmentation, thereby elevating prenatal diagnostic precision. Methods: Utilizing a comprehensive dataset–the largest to date for fetal head metrics–FetSAM incorporates prompt-based learning. It distinguishes itself with a dual loss mechanism, combining Weighted DiceLoss and Weighted Lovasz Loss, optimized through AdamW and underscored by class weight adjustments for better segmentation balance. Performance benchmarks against prominent models such as U-Net, DeepLabV3, and Segformer highlight its efficacy. Results: FetSAM delivers unparalleled segmentation accuracy, demonstrated by a DSC of 0.90117, HD of 1.86484, and ASD of 0.46645. Conclusion: FetSAM sets a new benchmark in AI-enhanced prenatal ultrasound analysis, providing a robust, precise tool for clinical applications and pushing the envelope of prenatal care with its groundbreaking dataset and segmentation capabilities.

Autonomous Vehicles (AVs) have advanced rapidly in recent years as they promise to be safe and minimize the burden coming from the driving task. AVs share the road with various categories of vehicles as Emergency Vehicles (EMVs) (e.g police and ambulance vehicles). When being approached by an active EMV, it is natural to expect all vehicles to cooperate with EMV, such that the EMV travel time is minimized. The decision-making block of an AV includes the responsibility of instructing the AV to change lanes, which is typically handled by the Lane Change Decision (LCD) model. A typical LCD model tends to overlook the presence of EMVs around, as they neglect the impact of the lane change on the EMV utility. To address this challenge, this paper proposes an Emergency Vehicle Aware LCD via utilizing Deep Reinforcement Learning. To our best knowledge, this is one of the pioneering works that propose a DRL solution for the problem, addressing important limitations that have been identified. The proposed solution was evaluated against a rule-based LCD known as MOBIL in terms of safety and level of cooperativeness with the EMV. Some key results found from the comparison between the proposed solution and MOBIL are (1) identical safety levels,(2) proposed solution is takes far less time to give up the lane when being approached by an EMV, and (3) proposed solution never blocks the path of the EMV, whereas MOBIL occasionally block the path.

About one-quarter of all car collisions in the United States are caused by distracted driving, and this ratio is expected to rise. As vehicles are equipped with more elaborate and complex technology, human-vehicle interaction via dashboard displays and controls will become more complex and distracting. Human-vehicle interaction via voice-based technology offers a less distracting alternative. In this study we aim to develop a voice-based car assistant, with a focus on Arabic language speech recognition. We prepare a new 4000-word domain-specific lexicon to comprehensively support driver-vehicle interactions, and we create corresponding text and speech corpora. Then we extract acoustic feature vectors and use various acoustic models to support speech recognition. The language model is created using an n-gram model. Then acoustic and language models, and the lexicon are combined to generate a decoding graph. The text corpus consists of 6110 elements, including words, phrases, and sentences. The speech corpus has more than 60000 recordings (almost 50 hours). For the decoding of noise-free audio, a Deep Neural Network + Hidden Markov Model provided 94.832% accuracy, a Subspace Gaussian Mixture Model + Hidden Markov Model provided 94.2% accuracy, and the best Gaussian Mixture Model + Hidden Markov Model provided 94.13% accuracy. For the decoding of noisy audio, a Deep Neural Network + Hidden Markov Model provided 93.316% accuracy, a Subspace Gaussian Mixture Model + Hidden Markov Model provided 92.62% accuracy, and the best Gaussian Mixture Model + Hidden Markov Model provided 91.82% accuracy. A usability study was conducted on the system with 10 participants. Almost all of the results of that study showed usability ratings of greater than 4.0 out of 5.0. These usability ratings indicate that the proposed system was seen by the participants as important, and useful for reducing driver distraction.

Continuous and low-cost broadcast authentication is a fundamental security service for distributed sensor networks. This paper presents a novel development of a continuous and low-overhead broadcast authentication protocol named enhanced Infinite timed-efficient stream-loss tolerant authentication (enhanced Inf-TESLA) protocol, based on the Inf-TESLA protocol, whose continuous authentication is limited to the duration of its keychains. The enhanced Inf-TESLA protocol satisfies important security properties, including lower communication and computational overhead; a continuous generation of keychains without the need to establish synchronization packets; scalability to a large network; and resistance to masquerading, modification, man-in-the-middle, and replay attacks. We also highlighted an unaddressed authentication issue in the last packets of the original TESLA protocol and proposed a corresponding solution. We performed a simulation analysis using JAVA and proved that, compared to the Inf-TESLA protocol, the enhanced Inf-TESLA protocol can continuously authenticate packets for the entire lifetime of the receiver. We also compared the enhanced Inf-TESLA protocol with the original TESLA protocol in terms of time complexity and critical authentication processes. The results revealed the superiority of the enhanced Inf-TESLA protocol over the original TESLA protocol in terms of the message authentication code (MAC) value generation time and packet authentication time, which we believe can significantly improve the lifetime and lower the energy expenditure of Internet of Things devices with limited power sources.

In recent years, the global pervasiveness of smartphones has prompted the development of millions of free and commercially available applications. These applications allow users to perform various activities, such as communicating, gaming, and completing financial and educational tasks. These commonly used devices often store sensitive private information and, consequently, have been increasingly targeted by harmful malicious software. This paper focuses on the concepts and risks associated with malware, and reviews current approaches and mechanisms used to detect malware with respect to their methodology, associated datasets, and evaluation metrics.

Lung cancer is reported to be the second most common cancer disease. This paper proposes a comprehensive and comparative global and local feature extraction framework for lung cancer detection using CT scan images. This framework consists of three main phases: data collection, global training and testing, and local training and testing. A set of 1000 CT scan images is used in this study. During the global training and testing phase, the collected images are preprocessed through image warping and cropping. Global features are then extracted from images to represent each image with feature vectors, using ten different image feature types. The feature vectors are then used to build detection models with six different machine learning algorithms. In the local training and testing phase, each image is divided into a set of local blocks. Those feature types that performed well in the global phase are then extracted from each of these blocks, to represent each block with feature vectors. These feature vectors are then used to build detection models for all of the image blocks, using the learning algorithms that performed well in the global phase. The results show that the Gabor Filter, the Histogram of Oriented Gradients (HOG), and the Haar Wavelet feature types outperformed the other seven feature types. The results also show that Support Vector Machine (SVM) outperforms the other five learning algorithms. Of most importance, the proposed local feature extraction approach outperforms the traditional global one. In the local phase, using SVM with Haar Wavelet features achieved 90% accuracy, 88% sensitivity, and 91% specificity. Using SVM with HOG features achieved 88% accuracy, 85% sensitivity, and 89% specificity. Finally, using SVM with Gabor Filter features achieved the best accuracy, sensitivity, and specificity rates of 97%, 96%, and 97%, respectively.

In multi-hop wireless communication, a sensor node should consume its energy efficiently for relaying of data packets. However, most IoT-devices are equipped with limited battery power and computing resources for wireless communications, and thus energy optimization becomes one of the major concerns in wireless sensors routing design. The wireless technologies usually use unlicensed frequency bands of 2.4 GHz to transmit the data. Due to the broadcasting medium, the wireless transmission interferes with the reception of surrounding radios. As a result, data transmission failure increases resulting in low-communication quality. Therefore, one of the best solutions to this problem is to select the hop distance node that has a few neighbor nodes to disseminate packets until it reaches the ultimate receiver. The proposed routing selects the node that has few neighboring nodes and thus less interference. In another word, the scheme finds a better load balancing, and thus minimizes the probability of overload on a sensor node. It also introduces a new clustering algorithm around a single base station to shorten the transmission distances. This approach periodically selects the cluster heads (CHs) according to its location based distance from the final destination. The extensive simulation studies reveal that the proposed algorithm finds the best routing node and clustering formation to forward the traffic and thereby minimizes the interference ratio. In addition, the proposed protocol achieves low-energy consumption and longer network lifetime than other popular protocols.

The focal and non-focal epilepsy is seen to be a chronic neurological brain disorder, which has affected $\approx ~60$ million people in the world. Hence, an early detection of the focal epileptic seizures can be carried out using the EEG signals, which act as a helpful tool for early diagnosis of epilepsy. Several EEG-based approaches have been proposed and developed to understand the underlying characteristics of the epileptic seizures. Despite the fact that the early results were positive, the proposed techniques cannot generate reproducible results and lack a statistical validation, which has led to doubts regarding the presence of the pre-ictal state. Various methodical and algorithmic studies have indicated that the transition to an ictal state is not a random process, and the build-up can lead to epileptic seizures. This study reviews many recently-proposed algorithms for detecting the focal epileptic seizures. Generally, the techniques developed for detecting the epileptic seizures were based on tensors, entropy, empirical mode decomposition, wavelet transform and dynamic analysis. The existing algorithms were compared and the need for implementing a practical and reliable new algorithm is highlighted. The research regarding the epileptic seizure detection research is more focused on the development of precise and non-invasive techniques for rapid and reliable diagnosis. Finally, the researchers noted that all the methods that were developed for epileptic seizure detection lacks standardization, which hinders the homogeneous comparison of the detector performance.