FMPM: Futuristic Mobility Prediction Model for Mobile Adhoc Networks Using Auto-Regressive Integrated Moving Average

Sa？etak The mobility of the node plays a crucial role in the route discovery process for Mobile Ad hoc networks (MANET). The high-speed node affects the routing process by packet relaying, packet delivery. This paper proposes a mobility prediction model FMPM using Auto-Regressive Integrated Moving Average (ARIMA) method to estimate the futuristic speed values of the nodes in MANET. It uses the autocorrelations of previous data by time series forecasting approach. The ARIMA model applied to mobile nodes for predicting its future speeds in the network. This mobility prediction supports the route discovery process to select the moderate mobility nodes for the reliable routing in the network. The nodes are trained by ARIMA model using neural network time-series tool of Matlab. Roughly, it trains each node with ten iterations. Simulation results show that the forecasted values almost match with the simulated node speed values. Performance analysis of the predicted mobility values exhibits the improved results compared to existing works on performance metrics such as mean square error, covariance, and computation overhead. The proposed FMPM produces more perfection in the speed prediction of a mobile node with a maximum of 80-100 % throughout the time series and has the lowest error values as 0.1085. But, the neural based prediction has the lowest MSE as 0.852. Therefore, the ARIMA prediction has approximately 0.75 reduced error values