Joint Adaptive Mobility Prediction and Signal Strength Prediction Based Cell Selection Algorithm in Ultra-Dense Networks

Ultra-dense network(UDN) improves the regional throughput, however, it has caused the problems of frequent handover and handover failure. Mobility prediction and signal strength prediction is a focus in recent research to solve these problems. In this paper, we propose a joint adaptive mobility prediction and signal strength prediction based cell selection algorithm for handover in UDN. Firstly, we predict the user's path through adaptive order Lagrange interpolation and the order is adjusted according to user's moving state. Then, we use the predicted path to estimate the time user stay in the neighbor cells, i.e., time of stay(ToS). Taking both the ToS and the prediction result of signal strength into account, the cell selection process is modeled as a multi-objective decision-making problem, and we use the TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) framework to solve it. The research reveals the influence of ToS on frequent handover and the advantage of adaptive parameters adjustment on mobility prediction. Numerical results show that the proposed algorithm reduces the number of handovers by 16.8% and handover failures by 56.9% compared with the contrast algorithm.