Extending the LBS-framework TraX: Efficient proximity detection with dead reckoning

Abstract

Proximity detection is defined as the capability of a location-based service (LBS) to automatically detect when a pair of mobile targets approaches each other closer than a pre-defined proximity distance. For realizing this function the targets have to be permanently tracked. To this end they are equipped with a cellular mobile device with an integrated GPS receiver, which passes position fixes obtained by GPS to a central location server. In order to save valuable bandwidth, reduce monetary costs for bearer services and to limit the power consumption at the mobile device, the number of messages exchanged between server and device needed for keeping track of the target should be reduced as far as possible. In the paper a novel strategy for efficient proximity detection is presented, which is based on an adoption of dead reckoning. In contrast to existing work the strategy considers the movement patterns of the observed targets. The paper presents results that have been achieved in various simulations comparing the proposed strategy to known approaches with regard to the amount of messages that pass the air interface. Also, it is described how proximity detection interplays with related functions like clique and k-nearest-neighbor detection. Therefore, the layered architecture of the LBS middleware framework TraX is presented, which has a special focus on community-LBSs.