Research On Flooding Probability In Wireless Sensor Networks

As a traditional routing protocol, flooding is an essential routing technique in wireless sensor network. In order to decrease the redundant data, a lot of improved solutions were put forward to overcome the objection of blind flooding and probabilistic scheme is concerned mainly in this paper. In probabilistic flooding, flooding probability is always obtained from substantial examinations instead of theory method. Flooding probability is studied in this paper as follows:First of all, the characteristic of probabilistic flooding is analyzed and the existence of probability threshold is proved by percolation theory and random graph theory. When rebroadcasting probability is greater than the threshold, the network can maintain a high message coverage rate. A method for calculating critical probability of probabilistic flooding is explored by treat the message transmission as a Markov chain. But unfortunately this method is merely fitted to grid model.To overcome the limitation of topology of networks, using branching process, a valid flooding probability is obtained in fixed probabilistic flooding. Researches show that dynamic probabilistic flooding has better performance than fixed probabilistic flooding. By applying the result to dynamic probabilistic flooding, a new dynamic probabilistic flooding algorithm based on the number of neighbor nodes is obtained. The result of simulation confirms that the algorithm is valid.