Geographic routing based on on-demand neighbor position information in large-scale mobile sensor networks

Abstract

Geographic routing has been considered as a scalable approach in ad-hoc and sensor networks since it exploits pure position information of neighbors and a destination node instead of global topology information to route data packets. To obtain the position information of the neighbors, many geographic routing protocols assume that each node exchange periodically its own position information with its neighbors. However, these periodically position exchanges in regions without packet forwarding make nodes consume unnecessary energy. In addition, given that the nodes have mobility, when a node forwards a packet, the position information of its neighbor nodes may be invalid. Hence, in these mobile networks, it is more efficient that a mobile node collects the position information of its neighbors when it needs to forward a packet. Because every sender node in geographic routing also selects the nearest neighbor node to the destination node as the receiver node, obtaining the position information of only closer neighbor nodes to the destination node can save more wireless resources than obtaining that of all neighbor nodes. In large-scale networks, because many wireless collisions can happen due to position exchanges of many neighbor nodes, they should share limited wireless resources. Therefore, we propose a mechanism to solve these issues of geographic routing in large-scale mobile networks. Simulation results show that the proposed mechanism has better performance than the existing geographic routing protocols.