Vehicular Ad-hoc network (VANET) is an imminent technology having both exciting prospects and substantial challenges, especially in terms of security. Due to its distributed network and frequently changing topology, it is extremely prone to security attacks. The researchers have proposed different strategies for detecting various forms of network attacks. However, VANET is still exposed to several attacks, specifically Sybil attack. Sybil Attack is one of the most challenging attacks in VANETS, which forge false identities in the network to undermine communication between network nodes. This attack highly impacts transportation safety services and may create traffic congestion. In this regard, a novel collaborative framework based on majority voting is proposed to detect the Sybil attack in the network. The framework works by ensembling individual classifiers, i.e., K-Nearest Neighbor, Naïve Bayes, Decision Tree, SVM, and Logistic Regression in a parallel manner. The Majority Voting (Hard and Soft) mechanism is adopted for a final prediction. A comparison is made between Majority Voting Hard and soft to choose the best approach. With the proposed approach, 95% accuracy is achieved. The proposed framework is also evaluated using the Receiver operating characteristics curve (ROC-curve).

Diagnosis of heart disease is a difficult job, and researchers have designed various intelligent diagnostic systems for improved heart disease diagnosis. However, low heart disease prediction accuracy is still a problem in these systems. For better heart risk prediction accuracy, we propose a feature selection method that uses a floating window with adaptive size for feature elimination (FWAFE). After the feature elimination, two kinds of classification frameworks are utilized, i.e., artificial neural network (ANN) and deep neural network (DNN). Thus, two types of hybrid diagnostic systems are proposed in this paper, i.e., FWAFE-ANN and FWAFE-DNN. Experiments are performed to assess the effectiveness of the proposed methods on a dataset collected from Cleveland online heart disease database. The strength of the proposed methods is appraised against accuracy, sensitivity, specificity, Matthews correlation coefficient (MCC), and receiver operating characteristics (ROC) curve. Experimental outcomes confirm that the proposed models outperformed eighteen other proposed methods in the past, which attained accuracies in the range of 50.00–91.83%. Moreover, the performance of the proposed models is impressive as compared with that of the other state-of-the-art machine learning techniques for heart disease diagnosis. Furthermore, the proposed systems can help the physicians to make accurate decisions while diagnosing heart disease.

Smart devices are effective in helping people with impairments, overcome their disabilities, and improve their living standards. Braille is a popular method used for communication by visually impaired people. Touch screen smart devices can be used to take Braille input and instantaneously convert it into a natural language. Most of these schemes require location-specific input that is difficult for visually impaired users. In this study, a position-free accessible touchscreen-based Braille input algorithm is designed and implemented for visually impaired people. It aims to place the least burden on the user, who is only required to tap those dots that are needed for a specific character. The user has input English Braille Grade 1 data (a–z) using a newly designed application. A total dataset comprised of 1258 images was collected. The classification was performed using deep learning techniques, out of which 70%–30% was used for training and validation purposes. The proposed method was thoroughly evaluated on a dataset collected from visually impaired people using Deep Learning (DL) techniques. The results obtained from deep learning techniques are compared with classical machine learning techniques like Naïve Bayes (NB), Decision Trees (DT), SVM, and KNN. We divided the multi-class into two categories, i.e., Category-A (a–m) and Category-B (n–z). The performance was evaluated using Sensitivity, Specificity, Positive Predicted Value (PPV), Negative Predicted Value (NPV), False Positive Rate (FPV), Total Accuracy (TA), and Area under the Curve (AUC). GoogLeNet Model, followed by the Sequential model, SVM, DT, KNN, and NB achieved the highest performance. The results prove that the proposed Braille input method for touch screen devices is more effective and that the deep learning method can predict the user's input with high accuracy.

Technology is advancing rapidly in present times. To serve as a useful and connected part of the community, everyone is required to learn and update themselves on innovations. Visually impaired people fall behind in this regard because of their inherent limitations. To involve these people as active participants within communities, technology must be modified for their facilitation. This paper provides a comprehensive survey of various user input schemes designed for the visually impaired for Braille to natural language conversion. These techniques are analyzed in detail with a focus on their accessibility and usability. Currently, considerable effort has been made to design a touch-screen input mechanism for visually impaired people, such as Braille Touch, Braille Enter, and Edge Braille. All of these schemes use location-specific input and challenge visually impaired persons to locate specified places on the touch screen. Most of the schemes require special actions to switch between upper and lowercase and between numbers and special characters, which affects system usability. The key features used for accessing the performance of these techniques are efficiency, accuracy, and usability issues found in the applications. In the end, a comparison of all these techniques is performed. Outcomes of this analysis show that there is a strong need for application that put the least burden on the visually impaired users. Based on this survey, a guideline has been designed for future research in this area.

Remote healthcare systems and applications are being enabled via the Internet of Medical Things (IoMT), which is an automated system that facilitates the critical and emergency healthcare services in urban areas, in addition to, bridges the isolated rural communities for various healthcare services. Researchers and developers are, to date, considering the majority of the technological aspects and critical issues around the IoMT, e.g., security vulnerabilities and other cybercrimes. One of such major challenges IoMT has to face is widespread ransomware attacks; a malicious malware that encrypts the patients’ critical data, restricts access to IoMT devices or entirely disable IoMT devices, or uses several combinations to compromise the overall system functionality, mainly for ransom. These ransomware attacks would have several devastating consequences, such as loss of life-threatening data and system functionality, ceasing emergency and life-saving services, wastage of several vital resources etc. This paper presents a ransomware analysis and identification architecture with the objective to detect and validate the ransomware attacks and to evaluate its accuracy using a comprehensive verification process. We first develop a comprehensive experimental environment, to simulate a real-time IoMT network, for experimenting various types of ransomware attacks. Following, we construct a comprehensive set of ransomware attacks and analyze their effects over an IoMT network devices. Furthermore, we develop an effective detection filter for detecting various ransomware attacks (e.g., static and dynamic attacks) and evaluate the degree of damages caused to the IoMT network devices. In addition, we develop a defense system to block the ransomware attacks and notify the backend control system. To evaluate the effectiveness of the proposed framework, we experimented our architecture with 194 various samples of malware and 46 variants, with a duration of sixty minutes for each sample, and thoroughly examined the network traffic data for malicious behaviors. The evaluation results show more than 95% of accuracy of detecting various ransomware attacks.