ANN Based Framework for Energy Efficient Routing In Multi-Hop WSNs

Abstract

This paper proposes a soft computing based framework for optimal path routing in the large scale WSNs. The proposed approach works in two phases; initialization phase and the operational phase. In order to evaluate route costs, the paper first proposes an ANN based integrated link cost (ILC) measure. ILC is a function of residual sensor node energy, average end to end delay (EED) and network throughput. In the initialization phase the framework sets up and self-organizes the WSN and creates routing tables and initial set of s-t paths through cluster heads. In the operational phase optimal routes under given timing constraints are evolved using BB-BC optimization approach. Timing constraints are imposed due to dynamic conditions imposed by energy expenditure of nodes. Once the near shortest path/optimal routes are available, data transmission for a predefined interval takes place in the WSN. The ILC based dynamic shortest path routing approach improves throughput, reduces average end to end delay and improves the life time of the WSN. We implemented the proposed framework in MATLAB and its performance on optimal path enumeration was simulated. The framework was observed to be working extremely efficiently by evaluating near least cost path, thus keeping track on throughput, end to end delay and energy efficiency of the given WSN.