A Back propagation Neural Network Model for HWSNs Using IMIMO with a Secured Routing Mechanism

Abstract

Generally, a wireless sensor network (WSN) is made up of autonomous devices that are spatially distributed and use sensors to monitor environmental conditions. Due to some of the diverse applications, such as healthcare monitoring, management of disasters, smartphones, military, and other systems of surveillance, deployment of sensor nodes is more which are independent and distributed in harsh environments. Clustering stabilizes the network and gives the maximum efficiency of energy where sensor nodes are grouped into clusters that give preservation of energy. The proposed work is mainly on a heterogeneous wireless sensor network (HWSN) which has various levels of energy developed to overcome some of the issues. Many protocols have been developed and designed for enhancing the performance of HWSN. The proposed work presents a model with the concept of Improved-Multiple Input Multiple Output (IMIMO) which provides secured routing against attacks. The research work is focused mainly on the detection and prevention of black hole attacks using an algorithm called Modified Back Propagation Neural Network (Modified-BPNN). Considering the previous research, the objective is to provide secure routing and increase network performance. The modified-BPNN algorithm reconstructs the network by eliminating the malicious nodes which disturb the network operation. The proposed algorithm works better for data transmission with secured routes and also network performance is achieved by expanding the MAC parameters. Throughput, Packet Delivery Ratio (PDR), Packet Loss, and Routing Overhead are evaluated using NS 2.35 simulator tool and compared with the existing Protocols such as DSR, AODV, and MAODV.