{"title":"基于WiFi的动作识别系统","authors":"Xianggang Zhang, Meina Dong, Ting Zhang","doi":"10.1109/AICIT55386.2022.9930165","DOIUrl":null,"url":null,"abstract":"Compared with the action recognition method based on wearable sensors and video, the action recognition technology based on WiFi has the advantages of extensive infrastructure, simplicity and no user interference. In this paper, we use the CSI data of WiFi signal to realize human actions recognition based on WiFi through the stages of data acquisition, data preprocessing and intelligent recognition. Preprocessing includes data Deduplication / interpolation, removing DC component of signal, removing abnormal data, data de-noising, segmentation of data flow. The typical U-Net is used as the action recognition network. In the experiment, through the recognition and verification of seven actions, the average accuracy rate reached 90.83%. In addition, the influence distances of the other actions are analyzed through experiments.","PeriodicalId":231070,"journal":{"name":"2022 International Conference on Artificial Intelligence and Computer Information Technology (AICIT)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"An Action Recgonition System Based on WiFi\",\"authors\":\"Xianggang Zhang, Meina Dong, Ting Zhang\",\"doi\":\"10.1109/AICIT55386.2022.9930165\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Compared with the action recognition method based on wearable sensors and video, the action recognition technology based on WiFi has the advantages of extensive infrastructure, simplicity and no user interference. In this paper, we use the CSI data of WiFi signal to realize human actions recognition based on WiFi through the stages of data acquisition, data preprocessing and intelligent recognition. Preprocessing includes data Deduplication / interpolation, removing DC component of signal, removing abnormal data, data de-noising, segmentation of data flow. The typical U-Net is used as the action recognition network. In the experiment, through the recognition and verification of seven actions, the average accuracy rate reached 90.83%. In addition, the influence distances of the other actions are analyzed through experiments.\",\"PeriodicalId\":231070,\"journal\":{\"name\":\"2022 International Conference on Artificial Intelligence and Computer Information Technology (AICIT)\",\"volume\":\"29 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-09-16\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2022 International Conference on Artificial Intelligence and Computer Information Technology (AICIT)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/AICIT55386.2022.9930165\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 International Conference on Artificial Intelligence and Computer Information Technology (AICIT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AICIT55386.2022.9930165","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Compared with the action recognition method based on wearable sensors and video, the action recognition technology based on WiFi has the advantages of extensive infrastructure, simplicity and no user interference. In this paper, we use the CSI data of WiFi signal to realize human actions recognition based on WiFi through the stages of data acquisition, data preprocessing and intelligent recognition. Preprocessing includes data Deduplication / interpolation, removing DC component of signal, removing abnormal data, data de-noising, segmentation of data flow. The typical U-Net is used as the action recognition network. In the experiment, through the recognition and verification of seven actions, the average accuracy rate reached 90.83%. In addition, the influence distances of the other actions are analyzed through experiments.