Information leaks out: Attacks and countermeasures on compressive data gathering in wireless sensor networks

Compressive sensing (CS) has been viewed as a promising technology to greatly improve the communication efficiency of data gathering in wireless sensor networks. However, this new data collection paradigm may bring in new threats but few study has paid attention to prevent information leakage during compressive data gathering. In this paper, we identify two statistical inference attacks and demonstrate that traditional compressive data gathering may suffer from serious information leakage under these attacks. In our theoretical analysis, we quantitatively analyze the estimation error of compressive data gathering through extensive statistical analysis, based on which we propose a new secure compressive data aggregation scheme by adaptively changing the measurement coefficients at each sensor and correspondingly at the sink without the need of time synchronization. In our analysis, we show that the proposed scheme could significantly improve data confidentiality at light computational and communication overhead.