A novel random forwarding protocol for distributed wireless sensor networks (WSNs) is reported in this paper. The proposed protocol features the utilization of azimuth angle of the nodes involved and opportunistic selection of the relaying node via contention among neighbors. First, the protocol with precise angle information is discussed and its multi-hop performance is evaluated by means of both simulation and analysis in terms of average number of hops to the sink node. Simulation results show that it performs well especially in network connectivity. As a complement, node mobility is also introduced to evaluate multi-hop performance in real environments. Then, a simple MAC scheme to ensure that the node with the highest angular advantage will be selected as a relay is proposed, and subsequently energy and latency performances based on it are evaluated. It is shown that our protocol can effectively deliver data with half number of active neighbors required in geographical random forwarding (GeRaF). Finally, a practical method is presented to estimate azimuth angle, combined with which, our forwarding protocol is shown to achieve the performance very close to the case with precise angle information. tain delay constraints [2]. This imposes two important factors, energy efficiency and transmission latency, into the consideration of multi-hop packet forwarding or routing strategies in WSNs.In many routing protocols, nodes should have full knowledge of the local or even global network topology such as STEM [3], GEAR [4], or GEM [5], etc. Though topology information can be obtained at some prices of signaling and computing, it is not easy to do so in a WSN, for (1) energy and latency being two AnRaF FOR WSNs 991 ϑ = min(θ 1 , θ 2 , · · · θ k ), k > 0 2π, k = 0 (2) Throughout the above assumptions, the event that the two nodes think they are the relay can be avoided.
Energy and Latency AnalysisIn this section, we analyze the energy and latency performance of our protocol under the assumption that N MaxAtt → ∞ and N MaxColl → ∞. AnRaF FOR WSNs 997 5.1.2. ReceivingAll active neighbors in communication region of the transmitter will have the same probability of being the winner to act as a relay, and the number of nodes obeys Poisson distribution with density M. Thus, the probability that a neighbor wins the contention is ∞ k=1