An analysis of probabilistic forwarding of coded packets on random geometric graphs

We consider the problem of energy-efficient broadcasting on large ad-hoc networks. Ad-hoc networks are generally modelled using random geometric graphs (RGGs). Here, nodes deployed uniformly in a square area around origin, and any two which within Euclidean distance 1 assumed to be able receive each other’s broadcast. A source node at origin encodes k data packets information into n(>k) coded transmits them all its one-hop neighbours. The encoding is such that, that receives least out n can retrieve original packets. Every other network follows probabilistic forwarding protocol; upon reception previously unreceived packet, forwards it with probability p does nothing 1−p. interested minimum ensures fraction decode from source. deem this near-broadcast. performance metric interest expected total number transmissions probability, where expectation over both protocol as well realization RGG. In comparison no coding, our treatment indicates judicious choice n, possible reduce while ensuring