Research On Delay Coordination Of OOSM Data Fusion Node In UWSN

Underwater wireless sensor network (UWSN) consists of a large number of nodes with the ability of awareness and wireless communication. Due to energy restriction of sensor node, data fusion is adopted to remove redundancy information between data packets and reduce data transmitted, in order to decrease nodes’energy consumption and extend network lifetime.Data fusion need to consider a problem that how to make a choice between time constraints and energy conservation. Namely, data fusion aims to research maximizing energy savings of network consumption on the condition of the given time of network. The purpose of a latency allocation algorithm is to satisfy the requirement of real-time performance and fusion gain. The algorithm allocates network allowable delay to the fusion node reasonably to avoid the node waiting in idle or blind, ensure real-time performance of network and save energy to a certain extent. However, influenced by water flow, marine environment, node communication and so on, there are usually different time delays in transmitting the measurement from the same target arrive out of sequence. Thus, phenomenon of "Out-Of-Sequence" Measurement (OOSM) occurs. The OOSM’s influence on the latency allocation algorithm directly affects the effectiveness of the algorithm.Therefore, refering to the existing latency allocation algorithm based on wireless sensor network, a kind of latency allocation algorithm is put forward which is suitable for UWSN, and then analysis the impact of OOSM phenomenon on the algorithm, the main content is as follows:(1) The thesis uses OPNET to build an underwater acoustic channel model and an underwater acoustic data link layer model, analyzes the influence of change of sea depth and temperature on sound velocity. The experimental results prove that the underwater acoustic channel model applied to calculate the propagation delay of the underwater environment.(2) Classical data fusion latency allocation algorithms in wireless sensor network are analyzed and then transplanted into UWSN. After simulated analysis, the thesis founds that the necessary delay of end-to-end transmission has the same allocation strategy as redundant time in the SSF-M/G/1 algorithm. As a result, when network topology changes, complexity of the algorithm is high, but its performance may be low.(3) A latency allocation algorithm suitable for UWSN is proposed. Fusion waiting time delay is calculated through the method of weighted path and fusion contribution based on the nodes’ location and the number of child nodes they own in the network. On the premise of meeting the real-time of network, the nodes save energy consumption by controlling their fusion waiting time delay. Through OPNET simulation, the thesis proves that the algorithm can further reduce network transmission delay and average energy consumption of the nodes and prolongs the network lifetime.(4) The influence of OOSM phenomenon on the latency allocation algorithm is analyzed when distribution of delay time is constant, periodic variation or the Poisson distribution. The concept of OOSM waiting time delay is introduced to improve the data accuracy of data fusion at the cost of increasing network data transmission delay.