Performance evaluation of geographic probabilistic flow-based spreading routing in wireless sensor networks

Included in the performance goal of most wireless sensor network applications are both spatial coverage and extended network lifetime. Pure energy efficient routing protocols,such as Shortest Path, cannot satisfy this need, since sensor nodes along the data path may have their energy levels drained very quickly, leading to loss of spatial coverage. Motivated by this necessity to incorporate longevity in the design of routing protocols for sensor networks, we proposed a probabilistic routing protocol called Probabilistic Flowbased Spreading (PFS), in which the intermediate nodes forward packets with a probability based on neighboring nodes' traffic load. In this paper we propose evaluating our protocol under scenarios which are representative of some typical traffic patterns that are likely to be encountered by sensor networks. We also study the performance of another protocol for sensor networks, GPSR, under the same scenarios for comparison purposes. We use the metric Lifetime -- Packets Received in order to evaluate the performance of the protocols both in the space and the time domains, and we show that our protocol consistently outperforms GPSR in all scenarios, reaching at the highest double the performance of GPSR.