The traditional intrusion detection technology has some shortcomings, such as high hardware requirements and harsh detection conditions etc. This paper proposes an environment intrusion detection technology based on WiFi, which uses the existing WiFi network to realize security monitoring function, covers a wide range and does not expose privacy. Firstly, the technology uses median filtering to denoise the subcarriers in the channel, and then using the self-organizing competitive neural network algorithm for fingerprint feature extraction and establish the intrusion signal. Finally, the statistical model of the nonlinear dependence between the intrusion and the fingerprint database is obtained by using the classification of normalized exponential function, to achieve the purpose of intrusion detection. The experimental results show that the recognition rate of this technology is improved by nearly 8% compared with the existing methods, reaching 98%, which has a good development prospect.

传统的入侵检测技术存在硬件要求高，检测条件苛刻等缺点。本文提出了一种基于WiFi的环境入侵检测技术，该技术利用现有的WiFi网络来实现安全监控功能，覆盖范围广，可以做到。不公开隐私。该技术首先使用中值滤波对信道中的子载波进行去噪，然后使用自组织竞争神经网络算法提取指纹特征并建立入侵信号。最后，通过归一化指数函数的分类，获得了入侵者与指纹数据库之间非线性相关性的统计模型，从而达到了入侵检测的目的。