Indoor intrusion detection method based on wireless channel state information

Abstract

The invention provides an indoor intrusion detection method based on wireless channel state information. The indoor intrusion detection method includes the steps of two parts that the first part, a detection threshold in a silent environment is obtained, that is, a data acquisition node is used for acquiring the wireless channel state information in the indoor unmanned state, the acquired channelstate information is subjected to noise filtering and separation point removal, the amplitude and phase of the acquired channel state information are extracted, and a standard deviation is calculatedto be used as the detection threshold; and the second part, real-time intrusion detection is conducted, that is, the data acquisition node is used for acquiring the wireless channel state information,a short-term mean standard deviation (SVR) and a long-term mean standard deviation (LVR) are used for real-time intrusion detection, the SVR is first used for detecting whether intruders enter a monitoring area or not, once the intruders are detected, the LVR is used for monitor whether the intruders are indoors all the time or not; meanwhile, a K-nearest neighbor algorithm and a dynamic time structuring method are used for classifying intrusion behaviors to accurately judge the intrusion behaviors. The indoor intrusion detection method based on the wireless channel state information overcomes the limitations of a computer vision method on the demand of visible light, the great influence on user privacy and the like, and meanwhile, has the characteristics of large detection range, low equipment cost, and the like.

Description

technical field [0001] The invention relates to an indoor intrusion detection method based on wireless channel state information, belonging to the field of intrusion detection. Background technique [0002] With the rapid development of society, the traditional manual monitoring system has become more and more difficult to adapt to the requirements of today's society, intelligent monitoring system will become the mainstream of the next generation monitoring system. The most basic task of a monitoring system is to detect whether someone has entered an area, which has always been a research hotspot in the security field. A commonly adopted method is the image-based detection method. This method deploys single or multiple camera devices in an area, and then uses image processing methods to extract features from the obtained images, and finally determines whether there is human activity. Although this method is widely used, it has the following disadvantages: 1. There are dete...

AI Technical Summary

Problems solved by technology

Although this method is widely used, it has the following disadvantages: 1. There are dead spots in detection, and it is expensive to deploy a large number of cameras

2. There is a problem of violating user privacy information

The disadvantage of this method is that the detection area is relatively single, and it can only be deployed on doors and windows. Once the intruder avoids these areas through some methods, the intruder cannot be detected due to the limitation of the infrared detection distance.

Images