Peripheral Routing Protocol – a new routing protocol proposal for a realistic WSN mobility model

Wireless sensor networks (WSNs) are changing our way of life just as the internet has revolutionized the way people communicate with each other. Future wireless networks are envisioned to be robust, have simple and efficient communication between nodes and self-organizing dynamic capabilities. When new nodes join in, a self-configuring network has to have the ability to include these nodes in its structure in real time, without human or machine interference. The need for a destination node (D) which moves at the periphery of wireless sensor networks can be argued from different points of view: the first is that different WSN scenarios require data gathering in such a way; the second point is that this type of node movement maximizes network lifetime because it offers path diversity preventing the case where the same routes are used excessively. However the peripheral movement model of the mobile destination does not resemble any mobility models presented in the WSN literature. In this thesis a new realistic WSN sink mobility model entitled the “Marginal Mobility Model” (MMM) is proposed. This was introduced for the case when the dynamic destination (D), moving at the periphery, frequently exits and enters the WSN coverage area. We proved through Qualnet simulations that current routing protocols recommended for Mobile Ad Hoc Networks (MANETs) do not support this sink mobility model. Because of this, a new routing protocol is proposed to support it called the Peripheral Routing Protocol (PRP). It will be proven through MATLAB simulations that, for a military application scenario where D’s connectivity to the WSN varies between 10%-95%, compared with the 100% case, PRP outperforms routing protocols recommended for MANETs in terms of throughput (T), average end to end delay (AETED) and energy per transmitted packet (E). Also a comparison will be made between PRP and Location-Aided Routing (LAR) performance when D follows the MMM. Analytical models for both PRP and LAR are proposed for T and E. It is proved through MATLAB simulations that, when compared with LAR, PRP obtains better results for the following scenarios: when the WSN size in length and width is increased to 8000 m and one packet is on the fly between sender and sink, PRP sends 103% more data and uses 84% less energy; when more data packets are on the fly between sender and sink, PRP sends with 99.6% more data packets and uses 81% less energy; when the WSN density is increased to 10,000 nodes PRP uses 97.5% less energy; when D’s speed in increased to 50 Km/h, PRP sends 74.7% more data packets and uses 88.4% less energy.