Energy optimization of wireless sensor network using neuro-fuzzy algorithms

Wireless sensor network (WSN) is one of the recent technologies in communication and engineering world to assist various civilian and military applications. It is deployed remotely in severe environment that doesn’t have an infrastructure. Energy is a limited resource that needs efficient management to work without any failure. Energy efficient clustering of WSN is the ultimate mechanism to conserve energy for long time. The major objective of this research was to efficiently consume energy based on the Neuro-Fuzzy approach particularly adaptive Neuro fuzzy inference system (ANFIS). The significance of this study was to examine the challenges of energy efficient algorithms and the network lifetime on WSN so that it could assist several applications. Clustering is one of the hierarchical based routing protocols, which manage the communication between sensor nodes and sink via Cluster Head (CH); CH is responsible for sending and receiving information from multiple sensor nodes and multiple sink nodes. There are various algorithms that can efficiently select appropriate CH and localize the membership of cluster with fuzzy logic classification parameters to minimize periodic clustering which consumes more energy and we have applied neural network learning algorithm to learn various patterns based on the fuzzy rules and measured how much energy was saved from random clustering. Finally, we compared it to our Neuro-Fuzzy logic and consequently demonstrated that our Neuro-Fuzzy model outperformed by saving more than 32% of energy than the random model with 50 and 100-sensor node deployment. We confirmed that by increasing the number of sensor nodes, it was possible to increase the energy utilization but not the energy saved from the network