Narrow Results

Clustering is an essential part of data analytics and in Wireless Sensor Networks (WSN). It becomes a problem for causes such as insufficient, unavailable, or compromised data in the face of uncertainties. A solution to tackle the instability of clusters due to missed values has been proposed. The fundamental theory determines whether to incorporate an entity into a group if it is not clear and probable. One of the main issues is identifying requirements for three forms of decision definition, including an entity in a cluster, removing an object from a group, or delaying a decision (defer) to involve or rule out a group. Current studies do not adequately discuss threshold identification and use their fixed values implicitly. This work explores using the game theory-based Possibility Clustering Algorithm for Incomplete Data (PCA-ID) framework to address this problem. In specific, a game theory will be described in which thresholds are determined based on a balance between the groups’ precision and generic characteristics. The points calculated are used to elicit judgments for the grouping of unknown objects. Experimental findings on the deep learning datasets show that the PCA-ID increases the overall quality considerably while maintaining comparable precision levels in competition with similar systems.

Typically, wireless sensor networks (WSNs) are used to monitor as well as detect different kinds of objects in realistic monitoring, where security remains as a major confront. Estimation of node trust is established to be an effectual way of enhancing the security, thus aiding in nodes collaboration and decision-making in wireless and wired networks. Nevertheless, conventional methods of trust management generally highlight on trust modeling and fail to notice the overhead issues. In this paper, a security aware ring cluster routing technique is introduced. The routing is undergone based on the multi-objectives including trust (security) parameters, energy, and distance. Here, the trust parameters include both the direct trust evaluation and indirect trust evaluation. Thereby, the lifetime of the network gets maximized even with secured manner. An innovative Self-Adaptive Deer Hunting Optimization (SA-DHO) is presented in this study because the optimization plays a significant role in selecting the neighbors as ring nodes. Finally, the superiority of the suggested approach is demonstrated in relation to various measures.

Nowadays, Wireless Sensor Networks (WSN) face more security threats due to the increased service of data transmission at high speed in almost all applications. The security of the network must be ensured by identifying abnormal traffic and current emerging threats. The most promising model for safeguarding the core network from outside attacks is Intrusion Detection Systems (IDS). This work focuses on the introduction of clustering-based intrusion detection in WSN. Initially, clustering takes place, where the nodes are grouped under certain constraints via selecting the optimal Cluster Head (CH). The considered constraints are energy, delay, distance, risk, and link quality. This optimal selection takes place by a new hybrid optimization algorithm termed as Truncate Combined Bald Eagle Optimization (TCBEO) algorithm. The subsequent process is intrusion detection, where a hybrid detection model combining a Convolutional Neural Network (CNN) & Bi-directional Gated Recurrent unit (Bi-GRU) is employed, which is trained with features like improved entropy and correlation taking into consideration of constraints like energy and distance, respectively. Eventually, the suggested work’s effectiveness is affirmed against existing techniques using various performance metrics.

The Internet of Things (IoT) is a developing technology in the world of communication and embedded systems. The IoT consists of a wireless sensor network with Internet service. The data size of the sensor node is small, but the routing of the data and energy consumption are important issues that need to be advocated. The Mobile Adhoc Network (MANET) plays a very important role in IoT services. In MANET, nodes are moving within the network. So, routes are created dynamically on demand and do not have any centralized units. The route optimization method addresses issues like selecting the best routes in terms of overhead, loop free, traffic control, balancing, throughput, route maintenance, and so on. In this paper, IoT routes are created between sensors to sink through MANET nodes with WSN routing ideology. The Krill Herd and Feed Forward Optimization (KH-FFO)-based method discovers the routes. The Krill herd algorithm clusters the network. This method increases network speed and reduces energy waste. Feed-forward optimization involves learning all the nodes in the network and identifying the shortest and most energy-efficient route from source to sink. The overall performance of the KH-FFO protocol has improved the network’s capacity, reduced packet loss, and increased the energy utilization of the nodes in the network. The ns-3 simulation for KH-FFO is tested in different node densities and observed energy utilization is increased by 28%, network life is increased by 7%, Packet delivery ratio improved by 7.5%, the End-to-End delay improved by 31% and the Throughput is 3%. These metrices are better than the existing works in the network.