An optimal routing approach in wireless sensor networks

Abstract

The route selection and forwarding mechanism models suffer from routing overhead, data drop, higher delay and link failures. The traditional route selection techniques are based on fixed rule depending on multiple parameters in a given sequence. Also, the existing techniques does not easily adapt to the changing network environment. The elasticity of the network in the terms of varying number of nodes and area is also not considered. Thereby, not only a single parameter can be considered best for selecting an optimized route for data forwarding. Conclusively, to overcome the existing routing problems, a multiple factor tradeoff mechanism based on Machine Learning by integrating the Q- Learning reward-based Route Selection model has been proposed to improve the network performance and reduce overall delay. The multiple factor trade-off can be the decisive factor to rely upon the best parametric combination for the given routes. The simulation results have been computed for QoS parameters. The proposed technique has outperformed the existing conventional techniques in terms of Delay, Throughput, Average First Dead Node Analysis and Energy Consumption. newline