ANN based link stability estimation model and enhanced routing protocol for MANETs

Abstract

Link stability, defined in terms of a link life, is an important information in order to select stable links for route establishment in dynamic network topology. By using link stability as a metric and biasing routing protocol to choose stable link we can reduce computational overhead on the nodes involved and route control overhead on the network. The network throughput and QoS performance of the network can also be improved. In this paper, we propose Artificial Neural Network based link stability estimation model for MANETs. The model estimates link stability considering signal strength, signal strength variation with respect to time and link lifetime distribution. The estimation of the model is shown to be superior to that of other link selection methods. The performance of link stability estimation model is further evaluated by incorporating it with AODV routing protocol. The modified AODV routing protocol is termed as ALS-AODV. The performance of ALS-AODV is evaluated in terms of network throughput, routing overhead, average delay and average hop count. The simulation results show that ALS-AODV is capable in reducing routing overhead and increasing network throughput.