Data Persistency in Wireless Sensor Networks Using Distributed Luby Transform Codes

In large-scale distributed sensor networks without a fusion center, increasing persistence and lifetime of the sensed data is important. In this paper, we therefore propose a distributed algorithm that generates redundant data based on Luby transform coding. In the proposed algorithm, sensed data is propagated in random according to a probabilistic forwarding table that is an extension of the Metropolis-Hasting weighting method for a Markov chain with non-uniform stationary distribution. Compared with the previous methods, the proposed algorithm here benefits from lower decoding overhead at sink and communication cost of data dissemination, which is verified by simulations.