Privacy-preserving Data Aggregation And Processing Techniques In Wireless Sensor Networks

Wireless sensor Network can sense data from the outside world and feed them back to users.Data aggregation is one of important techniques of wireless sensor network because it's energy-constrained.And with the rapid development of wireless sensor network,it's also vital to deal with the massive data generated by it.And meanwhile,privacy-protection must always be considered.This thesis,based on comprehensive study of the theories and development of wireless sensor network,pays close attention to prediction-based data aggregation and proposes two aggregation schemes which are energy-efficient and privacy-preserving.And a framework based on Hadoop,which aims to deal with the big data generated by wireless sensor network is proposed and tested.The main contributions of this thesis can be categorized into three parts.First,to reduce the amount of data transmitted within the network and provide privacy-preserving abilities,exponential smoothing based data aggregation scheme is proposed.Data aggregation in the network is link-based,private data is computed using predicted value.Simulation results show that this scheme presents high-level privacy-preserving abilities and has low energy consumption.Second,to further diminish energy consumption,especially reduce transmission times,ARMA based data aggregation scheme is proposed.This scheme calculates predicted values through ARMA models,updates and synchronizes models according to thresholds when needed.In the end,a Hadoop-based data processing framework is proposed,this framework can effectively deal with the big data generated by wireless sensor network and provide privacy-preserving abilities.Platform that we set up shows this framework is feasible and efficient.In conclusion,this thesis,from the perspective of data,solves the problems of data aggregation and processing in wireless sensor network,and always pays attention to privacy-protection.