A reliable fault-tolerant ANFIS model based data aggregation scheme for Wireless Sensor Networks

Wireless Sensor Networks (WSNs) are widely used in many applications like forest fire monitoring, environmental monitoring, search and rescue operations and weather monitoring etc. The WSNs are organized into clusters in order to conserve energy. A high level of reliability is necessary for using WSNs particularly in mission-critical applications like battlefield surveillance, disaster management etc. The node deployment is an essential feature of WSN. It is of two types- random and deterministic. An efficient node deployment scheme not only reduces the energy cost but also enhances the network lifetime. The paper proposes four deterministic WSN cluster deployment topology models - the square, the pentagonal, the hexagonal and the 3 x 3 grid model for fixed sensor deployment. The performance of the four proposed models are first analyzed using a Reliability Block Diagram (RBD) where the reliability value for each of the proposed models is calculated using the Sum of Disjoint Products (SDP) method. Then, the proposed models are compared through simulation for different performance metrics. The paper also proposes a Reliable Neuro-Fuzzy Optimization Model (RNFOM) data aggregation technique which combines the power of an ANFIS estimator for intra-cluster and inter-cluster fault detection in WSNs for different fault types with Gorti's Enhanced Homomorphic Cryptosystem (EHC) for imparting security to the network. Both the analytical and the simulation results confirm that the 3 x 3 WSN grid cluster deployment model powered with the proposed RNFOM data aggregation technique gives comparatively better performance and a longer network lifetime than the other proposed models for any clustered WSN. (C) 2017 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University.