Introducing a novel minimum accuracy concept for predictive mobility management schemes

In this paper, an analytical model for the minimum required accuracy for predictive methods is derived in terms of both handover (HO) delay and HO signaling cost. After that, the total HO delay and signaling costs are derived for the worst-case scenario (when the predictive process has the same performance as the conventional one), and simulations are conducted using a cellular environment to reveal the importance of the proposed minimum accuracy framework. In addition to this, three different predictors; Markov Chains, Artificial Neural Network (ANN) and an Improved ANN (IANN) are implemented and compared. The results indicate that under certain circumstances, the predictors can occasionally fall below the applicable level. Therefore, the proposed concept of minimum accuracy plays a vital role in determining this corresponding threshold.