Improved Markov mobility prediction mechanism for HetNets

Mobility prediction can make a network system be more sensitive and anticipate to the change of the location of its users and make its mobility management more intelligent. However, there are more challenges in the mobility prediction for Heterogeneous wireless Networks (HetNets) because of its special network topology and user mobility feature. This paper presents an Improved Markov Mobility Prediction mechanism (IMMP) for mobile users in HetNets by considering the different structure of HetNets. IMMP takes advantage of classic model of the traditional Markov mobility prediction algorithm. Meanwhile, IMMP covers the shortage of the traditional Markov mobility prediction when the prediction accuracy decreases because of not enough history location information. Simulations show that our IMMP can obtain better prediction accuracy and user adaptability performance under both short and long prediction times than the other two classic prediction algorithms.