An Efficient Privacy Preserving Scheme for Distributed Data Aggregation in Smart Grid

Privacy-preserving data aggregation in smart grids has been intensively studied to balance utility and privacy requirements. Existing contributions usually rely on a trusted aggregator to process electric power data and then apply privacy preservation strategies to protect the aggregate data. However, the privacy of individuals will be revealed once the aggregator is hacked or cannot be trusted. By injecting independent Gaussian noise into the response to the query before performing aggregation, a distributed Gaussian mechanism is designed to eliminate the negative effect of the untrusted aggregator. Equipping the distributed Gaussian mechanism with a post-processing step based on the moving average filtering technique, an efficient privacy-preserving scheme is proposed to improve individual electric power data utility. The experiments conducted on the real-world datasets show that the proposed scheme can improve the utility of electric power data while protecting the privacy of individual data.