Event Determination And Evolution In Underwater Sensor Networks Leveraging Practical Data Prediction

With the fast development of underwater sensor devices,the underwater sensor network is applied in the field of intelligent ocean construction.For example,ocean data collection,pollution monitoring,exploration,disaster prevention and auxiliary navigation,etc.Underwater sensor network real-time monitors the change of ocean environment,when an event occurs,policymakers need to locate the event area in time,in order to make real-time response.Besides,using the historical data to analyze the evolution trend.In this thesis,we use the real time data to predict the effective mechanism of events to determine,especially when the underwater environment changes not severe or when there is a rule,use data forecasting model,radically reduce packet to upload,so as to realize efficiently identify events.Use history to collect data information to judge the evolution of events.The main research contents are as follows:Firstly,considering the underwater sensor network nodes small computing power and limited storage space and transmission delay of sensor networks,this thesis used a simple model to capture data change trend of DBP data forecasting model,and with a kind of elastic rule to calculate the interference of handle exist at the same time.The model is adjusted by the learning window size of the control model and the tolerance of time value,and the calculation inclination is constantly iterated to adapt to the current trend of data development.Thereafter,real-time data collection of underwater sensor network requires more energy.Underwater nodes and base stations to deploy the same data forecasting model,when a node and the real and estimated values of the error,in the early set an acceptable range,the node of the current time there is no need to upload data,base station node to forecast as a real value.On the contrary,when the prediction error is too large,the node needs to upload data,while the underwater node and base station update model.The realization fundamentally reduces packet transmission to save energy consumption.Then,real-time data prediction and synchronization mechanism of underwater sensor network can effectively realize data aggregation.When the event occurs,use the collection of abnormal data,construct the network abnormal weight map,identify the event occurrence area location event source.Then,based on the historical data of continuous time slice,the evolution rule of the event is analyzed.So that policymakers can respond immediately to prevent the spread of events.Lastly,in this thesis,the effective event determination and event evolution analysis based on real-time data prediction are presented.The simulation results show that the proposed algorithm has good performance when the data change is not violent or network environment follows certain patterns.It reduces the network energy consumption,and the life cycle of sensor network is extended to a certain extent.