On Using Handoff Statistics and Velocity for Location Management in Cellular Wireless Networks

This paper studies the location management problem in cellular wireless networks. We propose a HandoffVelocity Prediction (HVP) scheme to minimize the paging cost in searching a mobile terminal. HVP is based on the assumptions that the movement behavior of mobile terminals has temporal and spatial properties. Based on handoff statistics, the system maintains a handoff graph to describe the movement probabilities of a mobile terminal in a cell to the neighboring cells within a location area. Combining with the velocity information of a mobile terminal, we calculate the probabilities of finding the mobile terminal in the cells within the paging area. Then, the paging of the mobile terminal follows the cell probabilities to minimize the paging cost of a mobile terminal. Analysis on HVP has been performed to calculate the optimal threshold for update generation to minimize the total cost in location management. A group-paging scheme based on non-linear programming technique is suggested to limit the paging delay within the quality of services (QoS) requirement in call connection delay and at the same time to minimize the paging cost. In 3G networks and the next generation wireless networks, different connection requests may have different QoS requirements in connection delay. Extensive experiments have been performed to investigate the performance characteristics of HVP as comparing with the adaptive distance-based (ADB) method, the direction-based location update (DBLU) method and the basic velocity paging (BVP) method under different system settings. The results have shown that HVP gives a better performance comparing with ADB, DBLU and BVP for different CMR values and update cost to paging cost ratios.