Adaptive Probabilistic Epidemic Protocol for Wireless Sensor Networks in an Urban Environment

In this paper, we consider the use of multi-level wireless sensor networks (WSN) in an urban environment for scenarios such as disaster recovery or a military mission in a battlefield. The quick deployment of WSN, comprised of ground sensors and unmanned air vehicles (UAVs) flying above building height, enables a new communication network to be formed instantly. However, due to the obstruction of buildings and movement of sensor nodes, communications among the sensor nodes is intermittent resulting in frequent partitioning of the network. The epidemic routing protocol, with its simple and robust flooding nature has made communication in a partitioned network possible, exploiting the mobility of the network nodes and based on pair-wise communication. However, due to the broadcast nature of this protocol, problems such as redundancy rebroadcast, contention and collisions, collectively known as the broadcast storm problem, have been observed. We first recommend a more detailed radio propagation modelling approach for the urban environment for more realistic simulation studies. Further, we propose an adaptive probabilistic epidemic protocol for WSN in the urban environment that adapts to the network topology over time. Through simulation studies, we show that this protocol improves the performance of the WSN in terms of data packet delivery probability compared to the case without probabilistic transmission in epidemic protocol.