Performance of received power and traffic-driven handover algorithms in urban cellular networks

This paper presents an investigation on the Handover Algorithms suitable for implementation in urban mobile cellular networks based on TDMA. More precisely, the class of Power and Traffic Driven Handover Algorithms, which are based on absolute and relative measured values of power and on actual traffic load, is analyzed. The algorithms that are usually investigated in the literature (and often implemented in actual mobile networks) belong to this class. We show the performance of traditional algorithms, and also propose some new ones, in an effort to reduce the probability of drop-out due to sudden changes of received power, and to distribute the traffic load over several cells; they show performance improvements with respect to those previously known from the literature. The performance of the different Handover Algorithms is evaluated by means of a simulation tool which allows the consideration of complex scenarios, suitable mobility and propagation models, power control, cell sectorization, non-uniform users distribution, etc. The metrics used for performance comparison are the outage, blocking, drop-out and satisfaction probabilities, whose different definitions are discussed in the paper, and the average number of handovers per call. Several urban scenarios, characterized by uniform and non-uniform traffic distributions, are taken into account. Our results show that one of the proposed algorithms, based on traffic estimates, yields the best performance in all the considered scenarios.