S-ARMA MODEL FOR NETWORK TRAFFIC PREDICTION IN WIRELESS SENSOR NETWORKS
S. Periyanayagi, V. Sumathy · 2014
Future network traffic in WSN can be predicted by time series models. The knowledge of traffic can be used for routing, load balancing and QoS provisioni ng. S-ARMA model has been proposed to predict the future traffic in WSN.The abnormality in traffic is predicted and it indicates the possibility for Dos attack and it initiates frequency hopping to avoid this. I ncrease in the frequency hopping time is identified by SARMA model, alerts the network to avoid the anomaly channel. Effectiveness of this model is been prove d to be efficient in detecting the anomaly channel fr om the simulation results since the information abo ut the attackers in the channel can be known using swarm i ntelligence (ants).