Efficient Non-Planar Routing around Dead Ends in Sparse Topologies using Random Forwarding

Geographic forwarding in wireless sensor networks (WSN) has long suffered from the problem of bypassing "dead ends," i.e., those areas in the network where no node can be found in the direction of the data collection point (the sink). Solutions have been proposed to this problem, that rely on geometric techniques leading to the planarization of the network topology graph. In this paper, a novel method alternative to planarization is proposed, termed ALBA-R, that successfully routes packets to the sink transparently to dead ends. ALBA-R combines nodal duty cycles (awake/asleep schedules), channel access and geographic routing in a cross-layer fashion. Dead ends are dealt with by enhancing geographic routing with a mechanism that is distributed, localized and capable of routing packets around connectivity holes. An extensive set of simulations is provided, that demonstrates that ALBA-R is scalable, generates negligible overhead, and outperforms similar solutions with respect to all the metrics of interest investigated, especially in sparse topologies, notoriously the toughest benchmark for geographic routing protocols.