An Integrated Neuro-Fuzzy and Evolutionary Model for Energy-Efficient Multipath Routing and Node Localization in Wireless Sensor Networks

Abstract

ABSTRACT
Wireless Sensor Networks (WSNs) are significant in a number of fields for their low cost and flexibility; however, they encounter many energy-efficient, reliable data transmission, and network lifetime aspects. Multi-path transmission is an attempt to improve data reliability, however, it is often inaccurate in routing, thus prone to consume lot of energy. It is also valuable to know the nodes' location accurately for efficient data routing, but it usually costs a lot of energy to infer naive the position of the node. To overcome these problems, the present paper introduces a new hybrid ANFIS based multipath routing, an IRNN model for node localization, and employs a mutation CSO to optimize these two models as an integrated routing and localization approach. To accomplish power efficient routing, the model selectively chooses best paths and refines routing precision via ANFIS while accurately localizing nodes incorporating the use of IRNN and MCSO. The performance of the model is verified through extensive simulations and the impact of packet delivery ratio, end-to-end delay, energy consumption, localization error, and total network life are evaluated. Experiments show the effectiveness of the hybrid model to enhance data transfer reliability and energy efficiency, which is better than previous method by reducing localization error, energy consumption and extending the network lifetime. This mechanism is an ideal solution for the problem of improving the routing accuracy as well as saving energy effectively in WSNs, and can be used to construct the applications that need to realize the reliable and sustainable data transmission.