| A wireless sensor network is comprised of a large number of distributed sensor nodes withlimited energy, and part of the nodes energy depletion will lead to the entire network paralysis. Inorder to prolong the lifetime of the whole network, how to use the battery capacity of sensor nodesefficiently becomes a very important research subject. Reasonable setting of the buffer space toreduce the probability of packet loss is a very critical issue in wireless sensor network.Firstly, the concepts of system components, protocol structure, features and severalapplications of wireless sensor networks are described in this thesis. Then, according to the differentwork modes of sensor nodes and the duty-cycle of MAC protocol, a strategy focuses on the size ofthe nodes buffer space is discussed in order to save the energy of the whole network, and two finitebuffer Markov models are established for sensor nodes. Secondly, based on analyzing the transitionprobabilities of two sensing patterns of nodes and the packet loss rate of sensor nodes, anappropriate buffer algorithm is proposed. Meanwhile, combined with the actual node arrival rateand the service rate or adjusting the duty ratio, the appropriate size of the buffer space is set tobalance the energy of the whole network. Finally, according to the traditional routing policy inwireless sensor network, a cross-layer optimization to the buffer algorithm is proposed. Rising fromthe perspective of a single node to all the interconnected nodes in the sensor network, and accordingto the different types and the number of nodes in the hierarchical routing protocol, the moreappropriate size of the buffer space can be set, at the same time, the goal of improving networkperformance can be achieved.MATLAB simulation results implied that the buffer algorithms can be used to set the size ofbuffer space according to different applications in order to reduce the packet loss rate, improveenergy efficiency and prolong the lifetime of the entire network. |
PDF Full Text Request