A novel temporal perturbation based privacy-preserving scheme for real-time monitoring systems

In real-time monitoring systems, participant’s privacy could be easily exposed when the time-series of sensing measurements are obtained accurately by adversaries. To address privacy issues, a number of privacy-preserving schemes have been designed for various monitoring applications. However, these schemes either lack considerations for temporal privacy or have less resistance to filtering attacks, or cause time delay with low utility. In this paper, we introduce a lightweight temporal perturbation based scheme, where sensor readings are buffered and disordered to obfuscate the temporal information of the original sensor measurement stream with differential privacy. Besides, we design the operations on the system server side to exploit the data utility in measurements from large number of sensors. We evaluate the performance of the proposed scheme through both rigorous theoretical analysis and extensive simulation experiments in comparison with related existing schemes. Evaluation results show that the proposed scheme manages to preserve both the temporal privacy and measurement privacy with filter-resistance, and achieves better performance in terms of computational overhead, data utility of real-time aggregation, and individual accumulation.