Research On UAV Intrusion Detection In Building Area Based On WiFi Perception

Nowadays,unmanned aerial vehicle(UAV)has been applied in various fields to help the development of society.However,due to the imperfect control system of UAV and people’s weak sense of safety,problems such as personal injury and privacy leakage of UAV frequently occur,which disturb social order.Therefore,UAV is forbidden to fly near many buildings.UAV intrusion detection must first detect the existence of UAV,and then locate it and find its location.WiFi sensing technology is often used in intrusion detection and indoor positioning,which has the advantages of low cost and easy deployment,and is a good technical choice for UAV intrusion detection.The main work of this paper is the intrusion detection of UAV based on WiFi awareness technology,mainly from the following two aspects:(1)for UAV detection,based on the principle that the amplitude of Channel State Information fluctuates in different degrees due to sports with different intensities in the environment,this paper proposes an UAV detection method based on amplitude variance.firstly,the amplitude information is extracted from CSI data,and the abnormal values are removed by Hampel filter to obtain reliable data,and then the amplitude variance is calculated,which is used as the characteristic fingerprint to construct an offline fingerprint database.During UAV detection,the real-time CSI data are processed the same as before,and the amplitude variance of online data is obtained.The classification results are obtained by matching the data with the fingerprint database through KNN algorithm.At the same time,combined with multi-antenna voting mechanism,the accuracy of UAV detection is improved.(2)For UAV positioning,the traditional two-dimensional positioning method is not applicable because it belongs to three-dimensional space positioning,with large amount of data and high similarity.In this paper,an unmanned aerial vehicle(UAV)positioning method based on distance and position double fingerprints is proposed.The positioning process is divided into two steps: first,coarse positioning is carried out by ranging,and then fine positioning is carried out after determining the approximate area of UAV.At the same time,the grid sampling method is replaced by the spherical sampling method,and the sampling points are set on the spherical surface with the position of the signal receiving end as the spherical center and the distance as the radius,and the space angle is taken as the standard of sampling point spacing,thus establishing a more reasonable spatial model.In this paper,Adaboost algorithm based on single-layer decision tree is used as the positioning algorithm to predict the position of UAV.Experimental results show that this method can effectively improve the positioning accuracy of UAV.