International Journal of Wireless and Ad Hoc Communication

With the use of deep learning algorithms, we provide in this work a novel approach, called "DeepDiffNet," to investigate the most recent advancements in the comprehension of coded diffraction patterns. Comprehensive tool DeepDiffNet decodes complicated coded diffraction patterns using deep neural networks. Encoding, decoding, and preprocessing are the three main algorithms used in the method.Preprocessing is an essential initial step in preparing coded diffraction patterns for analysis. It includes bringing intensity data into a standard range and employing a windowing tool to minimize noise and emphasize features.  The Encoding Algorithm leverages a convolutional neural network (CNN) to extract valuable data from the diffraction patterns that have been analyzed. Critically significant patterns and structures are recognized by the CNN via encoding them as feature vectors, which is how it learns to evaluate input. To reconstruct the original objects or specimens from the encoded information, the Decoding Algorithm uses a recurrent neural network (RNN). The RNN models the relationships between these features and the spatial arrangements of things to reconstruct them properly. We use many measures, such as Mean Absolute Error (MAE), the Structural Similarity Index (SSI), and the Peak Signal-to-Noise Ratio (PSNR), to evaluate DeepDiffNet's performance. These measures guarantee the reliability and efficacy of our approach to pattern reconstruction. When compared to conventional approaches, DeepDiffNet is light years ahead in terms of accuracy, precision, recall, and processing efficiency when analyzing coded diffraction patterns. The method's outstanding efficacy, flexibility, and resilience make it a priceless resource for a wide range of scientific, medical, and industrial endeavors.

In recent years, cloud computing was and still is one of the most pragmatic and popular topics of research because of its advantages. Cloud storage allows organizations to store information of service providers at remote sites. However, cloud computing has encountered challenges, notably security issues and scheduling problems, primarily stemming from concerns related to data confidentiality and efficient resource allocation among users. These challenges are inherent to cloud computing, where data and computational resources are shared among multiple users and often hosted on remote servers operated by third-party providers. Hence, our objective is to identify and analyze the challenges associated with cloud computing, with a particular focus on data security. in addition to conduct scientific review and compare multiple recent research studies. The focus will be on identify challenges and advantages of cloud computing and data security when going through various data security measures that are currently employed in cloud computing. eventually we will come up with valid recommendations based on the findings.

  The convergence of Industry 4.0 and sustainability has brought forth a new era of manufacturing, where data-driven approaches play a pivotal role in achieving operational efficiency while minimizing environmental impact. This paper presents an innovative framework for sustainable smart manufacturing through data-driven predictive maintenance planning. By integrating advanced analytics and machine learning, we propose a preemptive equipment management approach that not only optimizes production processes but also fosters environmental responsibility. Our methodology combines the power of Long Short-Term Memory (LSTM) networks for pattern modeling and the Sea Lion Optimization Algorithm for feature selection. We demonstrate the effectiveness of our approach through a comprehensive empirical analysis conducted on a real case study, where the results indicate significant improvements over baseline studies, as evidenced by reduced Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE), along with higher R-squared (R2) values. Our findings emphasize the synergy between technological innovation and sustainability imperatives, positioning our approach as a catalyst for reshaping modern manufacturing practices.  

The technology that was developed during the fourth industrial revolution has contributed to the recent surge of interest that has been seen in the field of medicine. In particular, the importance of personal medical information obtained via knowledgeable self-diagnosis is becoming more apparent. However, the disclosure of such private medical information raises several concerns regarding trustworthiness and security. Accidents involving personally identifiable medical information could happen on the computer, but more frequently than not, they take place during the process of information exchange and data transfer. So, the goal of this research is to improve the trustworthiness of managing such sensitive data by making blockchain technology better. The objective of the project was to create smart healthcare systems by utilising blockchain technology and the Internet of Things (IoT). Moreover, they utilised various measuring instruments to collect data and carry out an individual electrocardiogram assessment. Through an examination of the fused threshold, the observed biosignals were analysed to provide a tailored diagnostic. In this article, we describe the implementation of a monitoring system that analyses individual biometric information by making use of measuring devices. Machine learning has been included in the deployed system, which has resulted in better dependability and security of the system's information.

The Internet of Things (IoT) is a cutting-edge piece of cybernetic infrastructure that will eventually link all manner of previously disconnected physical objects to the web. The IoT is rapidly expanding into many facets of human life. IoT's attack surface has grown as a result of the technology's hyper-connectivity and inherent heterogeneity. In addition, IoT devices are used in both managed and unmanaged settings, leaving them open to innovative attacks. Fog computing is used in the proposed intrusion detection system for IoT applications to implement intrusion detection in a decentralised manner. Attack detection at fog nodes and summarization on a cloud server make up the proposed system's two parts. The local fog nodes in the IoT environment examine the traffic, and then they send a report to the cloud server that summarises the current global security state of the IoT application. According to the results of the experiments, the fog nodes are able to identify the attack 27% more quickly while also reducing the number of false alarms. The work that has been recommended provides a beginning point for the creation of a fog-based intrusion detection system that can be used for applications related to the IoT. The proposed system has a false alarm rate of only 0.32% and an accuracy of 98.15 percent. The proposed method can only identify attacks that conform to specific patterns.

Large number of small sensor nodes exists in WSN’s for sensing and collecting information from the environment. In today’s time, these sensor nodes were applied in under water, military area, health care, earthquake sensing and in dedicated areas with recent technologies. Sensor nodes have limited life time and have supplementary network life. Network lifecycle depends on many factors such as connectivity, residual energy, topology types, single hop, multi hop, distance from base station, distance to cluster heads and much more. Among the various solutions given, clustering is considered to be good solution and optimal cluster head selection leads to efficient energy consumption. This paper proposes fuzzy based multi-attributes clustering that balances load among sensor nodes and also gives energy efficient clustering. Here we have used some attributes such as delay, residual energy, distance to CH, standard deviation to average network lifetime and standard deviation to residual energy. Results and experimental analysis validates that the proposed methods outperforms other compared algorithms.