Pattern prediction in infrastructured wireless networks: Directional vs temporal statistical approach

The main desire of end-users of a telecommunication system is to take advantage of satisfactory services, in terms of Quality of Service (QoS), especially when they pay for a required needing. Unfortunately, when hosts are moving into an infrastructured network, it is very important to mitigate hand-over effects, which may cause heavy flow degradations or disruptions. In this paper, we are interested in investigating how the continuity of services can be guaranteed in QoS networks, by the analysis of users mobility from two different points of view: the first based on a directional modeling of users mobility, the second based on a time-based modeling of Cell Permanence Time (CPT). After a hand-over event, bandwidth availability should always be granted for those users who need to take benefits from QoS networks: the only way to face this issue is represented by the employment of a mobility prediction scheme, in order to know 'how and when' users move among infrastructured coverage areas. Based on a large number of experiments, the proposed schemes are then compared, in order to observe what are the benefits of the considered policies and when it is preferred to use each of them. © 2013 IEEE.