LOOP: A location based routing scheme for opportunistic networks

As a key enabling technology for pervasive computing, opportunistic networks have attracted intensive research efforts recently. In this paper, we present a new routing scheme for opportunistic networks that aims at forwarding messages to a destination location/area, instead of forwarding to specific nodes. Our routing scheme, termed LOOP for LOcation based routing for OPportunistic networks, exploits the regularity embedded in human moving pattern. As human movements often exhibit a high degree of repetition including regular visits to certain places and regular contacts during daily activities, we can predict a mobile node's future locations based on its mobility trace with high confidence. We formulate the movement pattern mining as a multi-label classification problem and construct a Bayes' predictive model to explore the mobility history and learn the movement pattern. This movement pattern will then be used to predict the node's future movement. Based on the prediction, the ability of the node to deliver a message to the destination is quantified through defined metrics. These metrics will be the determining factor for choosing proper relaying nodes in several proposed strategies. Our scheme can preserve privacy as no information, including location information, needs to be exchanged among nodes. At the same time, our scheme achieves total distributed control as each node can choose its individual forwarding strategy without involving network wide changes. Our analytical and simulation results show that LOOP can achieve significant performance gains over well known existing strategies for routing in opportunistic networks.