Research And Realization On The Data Fusion Energy-saving Algorithm Of Wireless Sensor Based On The Regression Forecasting Algorithm

The feasibility of efficient and low-cost sensor network has improved along with the highintegration of MEMS, RF and DSPS of low energy consumption and low cost, which lays afoundation for the application and promotion of WSN. However, there is still a key problem thathas not been solved about spreading the WNS widely---the energy carried by the normal node.The replacement of battery is quite inconvenient, so it has been a research hotspot about how touse the energy of node efficiently.The research shows that the energy of working WSN node is mainly lost in datatransmission and acceptance. We can reduce energy consumption in receiving and sending databy decreasing frequency of data transmission. Thereby the operating life of WSN will beextended. The node of WSN begins to work after the automatic topology forming a network.Normal node aims at collecting the data that have been set, and then sending the data to SINKnode at certain times. Then SINK node receives the data of normal node and uploads them toServer for users.Innovations of the paper rest with:1. Analyze the internal relationships of data by using Data Fusion Algorithm which studythe data of current cycle and data of previous cycle.2. Build Data Fusion Algorithm based on Regression Forecasting by combining DataFusion Algorithm and Regression Algorithm.3. Apply the Regression Forecasting Model to WSN and use it to predict future datainformation. We can infer the next cycle data information by using historical data withthe algorithm. Compare the error by setting threshold. If the error is within theacceptable level, not upload the data of normal node,but use the calculated data of SINK node. By doing this, we can reduce theamount of transmitted data and then achieve the goal of saving the nodeenergy of WNS.