Probabilistic Flow-based Spreading Geographical Routing for Wireless Sensor Networks

摘要

Routing protocols for wireless sensor networks must ad dress goals simultaneously in time, to prolong the network operational time, and in space, to keep effective spatial cov erage by reducing the chances of congestion. This dual na ture of application goals requires the routing protocol not only to be energy aware, but also traf?c load aware. In this paper we propose a routing protocol (Probabilistic Flow based Spread -^sPFS) that probabilistically forwards packets based on localized traf?c ?ow information. Our protocol spreads traf?c to multiple paths by locally distributing load uniformly to its next hop neighbors. In doing so we avoid over-utilizing any single node and so quickly draining its energy. Our protocol reacts to dynamic load conditions so that traf?c will not converge to any single node or path. In order to measure the performance of the protocols we propose the following metric: Network Lifetime × Pack ets Received, which captures the performance in both the time and the spatial domains. We show through simulations that our approach outperforms greedy forwarding by 45% and maximum energy by 35% when evaluating the Network Lifetime × Packets Received metric. In addition, it sustains double the data rate of maximum energy forwarding with out suffering congestion.