A Hybrid Mobility Prediction (HMP) strategy for PCS networks

作者：Saleh Ahmed Ibrahim Mohammed^*

来源：Pattern Analysis and Applications, 2016, 19(1): 173-206.

DOI：10.1007/s10044-015-0483-1

• PCS
• Resource reservation
• Handoff
• Location prediction

Recently, several high-speed applications have been emerged for personal communication service (PCS) networks. Therefore, locating mobile terminals (MTs) as they move from one place to another is the key to provide continuous services with unrestricted mobility. To maintain the continuity of service during handoffs, PCS involves the need to predict MTs' movements to cope with resource limitations and heterogeneity. Predicting MT's next location in PCS networks has become an inherently interesting and challenging problem. If we can correctly predict the next cell to which MTs are going, the performance of wireless applications, such as call admission control, QoS, and mobility management can be improved as well. Mobility prediction based on MTs' movement history has become an effective mean not only to decrease call-dropping probability, but also to shorten the handover latency. However, applying various basic prediction schemes to realistic environments has reported several limitations. Moreover, the randomness of MT's motion prevents the accurate prediction. In this paper, a Hybrid Mobility Prediction (HMP) strategy has been proposed. HMP provides flexible usage of the limited resources of PCS networks. It combines evidence from three different predictors, which are; probabilistic predictor (PP), group-based predictor (GP), and spatial predictor (SP). PP relies on Bayes theorem, GP uses ant colony optimization (ACO), while SP tries to detect the topological architecture of the current registration area to enhance the prediction process. HMP provides high accurate prediction of the next cell to which an MT is going, which avoids over reservation resulting in a waste of resources. Experimental results have shown that HMP outperforms recent prediction strategies as the amount of reserved resources is dynamically adjusted based on the predicted cell.

• 出版日期2016-2