Research On Data Aggregation Algorithm For Wireless Sensor Networks

Wireless Sensor Networks(WSNs)can use many kinds of sensors for real-time monitoring and collecting information of the objects,and send the information collected to the sink nodes.Therefore,WSNs have vast application prospects in environmental surveillance,m-health,traffic monitoring and many other fields.Typically,the energy of wireless sensor networks is limited,and the transmission of data consumes a lot.Therefore,reducing the amount of data and energy consumption,as well as prolonging the network lifetime,is of great importance in WSNs.Data aggregation is an important data processing technology of wireless sensor networks,by aggregating collected or received data,which can effectively remove redundant data.This thesis focuses on temporal and spatial correlation based data aggregation algorithm for wireless sensor networks,which aims at reducing the amount of data,saving energy consumption issues,and finally prolonging the network lifetime.First,the spatial auto-regression model based data aggregation algorithm(SMDA)is proposed according to the spatial data correlation.In SMDA,the cluster head nodes collect data from member nodes who work,select sleeping nodes using sleep scheduling algorithm.Then,cluster head nodes predict the data of sleeping nodes using delaunay triangulation and spatial auto-regression model,finally aggregate all received data and send to the sink node.Meanwhile,a spatial auto-regression model and grey model based data aggregation algorithm(SGDA)is proposed,which involves both temporal and spatial data correlation.By taking the minimum sum of error absolute as the target function to build combination forecasting model,the SGDA algorithm further reduces the prediction error.The simulation results show that the proposed approaches significantly reduce communication redundancy and balance energy consumption of nodes,evidently improve the lifetime of wireless sensor networks and ensure high data accuracy.