Research On The Application Of Location Service For Indoor Personnel Based On Wireless Perception

The application of location service based on wireless sensing refers to the use of wireless sensor network technology to count and locate people in the monitored area and WiFi authentication based on the location of people.At present,there are a lot of people location service technologies that can be used for wireless sensing,including RFID,Bluetooth,infrared,UWB,Zig Bee,WiFi and pedestrian dead reckoning(PDR)technology.In recent years,with the popularity of smart phones and other mobile devices,location service based on WiFi has also been developed rapidly.WiFi wireless network widely exists in indoor environment,and has many advantages such as easy to set up,fast transmission rate,etc.,which makes it very suitable for indoor location service technology.This paper studies the indoor personnel count and location count based on wireless perception,and proposes a WiFi authentication method based on the location information of indoor personnel.The main work completed includes:1)The number of indoor personnel and channel state information(CSI)feature matching are studied in depth.This paper analyzes the influence of the number of indoor personnel on the channel state information,and constructs a mathematical model to identify the number of indoor personnel based on machine learning.The CSI source data used in this method can alleviate the influence of signal multipath effect.The method of deconvolution phase and linear transformation is used to reconstruct CSI,so that CSI phase information can be centralized in the form of cluster,avoiding the problem of too large range of original phase distribution and too high randomness.Based on the Hampel filter,the singular data in the carrier amplitude is removed to reduce the interference caused by the environmental noise to the number characteristics of the personnel and ensure the accuracy and stability of the personnel count.Finally,support vector machine(SVM)classifier is used to identify the number of people.2)This paper proposes a CSI based indoor personnel location method.In this method,principal component analysis(PCA)is used to extract the features of the fingerprint database,which reduces the dimension of training samples and improves the time complexity of regression training.V-SVR is proposed to analyze the distance and angle of personnel position coordinates(L,?)respectively.The experimental results show that the positioning accuracy of the indoor positioning method proposed in this paper is within2.5m,and it can accurately locate people in the indoor environment.3)This paper proposes a WiFi authentication method based on the location information of people.This method builds SVM model based on CSI fingerprint database,obtains the location information of the user who initiated WiFi authentication by SVM calculation,and judges the legitimacy of the user who initiated WiFi authentication by identifying the user's location.In order to remove the noise in CSI,we use Hampel filter to remove the singular data in CSI data stream and avoid the position error caused by environmental factors.The percentile algorithm is used to improve the efficiency of parameter selection of Gaussian kernel function.