A Real-time CSI-based Passive Intrusion Detection Method

Passive intrusion detection plays a vital role in many specific applications such as area monitoring and elderly care. Nowdays passive intrusion detection based on wireless signals can achieve robust intrusion detection, but previous detection methods usually fail to realize real-time detection. In our paper, a real-time passive intrusion detection method based on channel state information (CSI) is proposed, making use of common commercial wireless network equipment to realize real-time passive intrusion detection. We extract the power spectral density from CSI as a feature to describe the changes of wireless signal, using dynamic threshold algorithm to achieve real-time intrusion detection. Meanwhile, exponential moving average algorithm are employed to update threshold. For evaluate our method, we conduct detection experiments in a laboratory. The experimental results indicate that the intrusion detection accuracy of our method is over 99%, and real-time detection can be completed in about 1s. In addition, this paper also discusses different influencing factors.