Research On Human Positioning And Motion Form Detection Based On Pyroelectric Infrared Wireless Sensor Network

Indoor human body positioning and motion recognition have always been hotspots in the field of contemporary artificial intelligence.There have been many research results on these two aspects,but from the aspects of cost,implementation complexity,and human privacy,the existing Some human target positioning and motion recognition algorithms have problems such as expensive equipment,high environmental requirements,and easy leakage of life privacy.In view of the above problems,this paper proposes to use the analog signal output by the pyroelectric infrared sensor when the human body moves to locate and recognize the human body.First,the pyroelectric infrared sensor is low in cost and power consumption.Second,the signal it collects is human infrared heat radiation,which neither reveals privacy nor interferes with normal life and work,and it can adapt to complex and changeable external environments.Therefore,this paper makes full use of the analog signal output by the PIR sensor,establishes a human body action recognition model through a deep learning network,and establishes a human body target positioning model through the cross-correlation characteristics of the signals,and realizes the human target positioning and motion recognition in an indoor environment.The main work of this thesis is as follows:(1)Based on the analysis of the analog signal output by the PIR sensor,a distributed pyroelectric infrared sensor system was designed and built to collect real-time signals of human motion,and provide raw data for the subsequent establishment of human motion recognition and human target positioning methods;(2)Aiming at the timing characteristics of PIR sensor node output data during different human actions,a two-layer network cascaded hybrid deep learning network model is proposed as a classification algorithm for human actions.The first layer uses a onedimensional convolutional neural network to extract features from the data,and the second layer uses a gated loop unit to save historical input information to prevent loss of effective features and uses the softmax classifier to output the recognition accuracy of five basic actions,and finally displays.The average recognition accuracy rate is as high as 98%;(3)The method for single PIR sensor to locate the human target has low positioning accuracy and poor anti-interference ability,so a human target localization model based on the correlation characteristics between the two PIR sensor signals is proposed.Use the correlation coefficient between the two PIR analog signals and the location of the target to construct a positioning function model,so that the position of the human body can be determined by the correlation coefficient between the two signals.Experiments have verified that the model has a small average positioning deviation and anti-interference ability Strong characteristics,the average positioning deviation is only 0.087 m.Experiments show that the human body motion signal collected in the human body pyroelectric infrared information acquisition system designed in this paper can be used for indoor human target positioning and motion recognition,with high accuracy,which has important theoretical significance and applications value.