Channel Aware Encryption and Decision Fusion for Wireless Sensor Networks

Abstract

In wireless sensor networks (WSNs), security and energy consumption have been considered as long-lasting technical challenges as sensors usually suffer from complexity and energy constraints. In this paper, we study a simple and efficient physical-layer security to provide data confidentiality in a distributed detection scenario. In particular, to prevent passive eavesdropping on transmitting data from sensors to an ally fusion center (AFC), we propose a novel encryption scheme and decision fusion rules for a parallel access channel model. The proposed scheme takes advantage of a free natural resource, i.e., randomness of wireless channels, to encrypt the binary local decision of each sensor in such a way that the binary local decision is flipped according to instantaneous channel gain between the sensor and AFC. The location-specific and reciprocal properties of wireless channels enable the sensor and AFC to share the inherent randomness of wireless channels which are not available to an eavesdropper. Furthermore, it is shown that the scheme is well-suited to a low complexity and energy efficient modulation technique, noncoherent binary frequency shift keying. To evaluate performances of the proposed scheme, log-likelihood-ratio-based decision fusion strategies at the AFC are analyzed, and comparisons of decision performances are carried out. In addition, we prove that the proposed scheme achieves perfect secrecy with a simple structure that is suited for sensors of limited complexity.