Research On Sensor Node Deployment Optimization Based On Implicit Recognition Of Human Behavior

The implicit perception of human behavior is widely used because it does not need to perceive object to wear any device and involve less user privacy.Existing research In order to ensure the completion of human behavior recognition tasks,often deploy more than the actual need for the sensor nodes,so that the wireless sensor network completely cover the monitoring area.And utilize all the radio frequency signals in the data processing,and not take into account the effective location of the sensor nodes,which can lead to problems such as data redundancy,channel interference,large computation and waste of hardware cost.Based on the multi-layer link recognition model,this dissertation mainly analyzes the two aspects of implicit perception of human behavior recognition and sensor node deployment optimization in combination with relevant theoretical research.Aiming at the problem of sensor node deployment optimization,this dissertation proposes a method to find the optimal node subset based on genetic algorithm.Through a large number of experimental results,it is found that this method can effectively reduce the number of node deployment.The main contributions in the dissertation can be summarized as follow:(1)In order to ensure the precision of the data,this dissertation collect six different behavior data corresponding to six different text by using behavior switching method in the real indoor environment deployment of sensors,while using the sliding window technology to fill the lost Packet data,extract the time domain mean feature,construct the human behavior identification data set.(2)We construct the training model and recognition model by using K-nearest neighbor and support vector machine on the original data set respectively to analyze the accuracy of human behavior recognition,and then select the appropriate classification algorithm.Then we use the genetic algorithm to perform the node subset optimization operation,select the subset of the data link corresponding to the subset of candidate nodes by the algorithm to identify the human behavior,so as to find a subset of the candidate nodes whose recognition accuracy is above 92%and the node with the least number of nodes,that is the optimal node subset,find the corresponding node numbering,and deploy the sensor node according to the node numbering to implicitly perceive the human behavior.In order to verify the robustness of this method,respectively choose two different types of indoor environments for experimentation,and the redundant nodes are reduced by more than 60%.