Research On Breath Detection Algorithm Based On Wireless Sensor

Breathing,an important basis for life,is a key indicator of health status for human be-ing.In the literature,with contact-based devices,some breathing signal detection methods have been proposed,which can achieve high-precision,high signal-to-noise ratio perfor-mance.However,these methods require user to be contacted with the devices,which can cause a series of problems,such as hindering the movement of users or bringing some unpleasant feeling to users.Therefore,there is an urgent need to call for a contactless solution to extract respiratory signals.Recently,with the popularity of the indoor wireless devices,breath detection with wireless sensors has drawn a lot of attention.However,the multipath effects,which com-monly exist in indoor environments,have serious impact on the propagation of wireless signals,leading to signal attenuation and poor received quality.Moreover,although the channel state information(CSI)can be obtained with the commodity WiFi such as Intel5300 network card,the phase of the CSI is distorted due to various offsets introduced during the receiving and transmitting of the wireless signals,due to which the CSI phase cannot be directly utilized.In this thesis,we try to resolve the challenges mentioned above,and the main contri-butions are follows:In Chapter 1 and 2,we introduce the motivation of using wireless sensors for breath detection,and review the state-of-the-art related works.Then,we introduce the propaga-tion model for the WiFi signal,and some commonly used neural network models.In Chapter 3,we propose a novel breath detection algorithm using WiFi signals.Specifically,we first use the multiple signal classification(MUSIC)algorithm and RAP-MUSIC algorithm to estimate the angle of arrival(AOA)of reflected targets.Then,the signals on different paths are utilized to eliminate the phase offsets.Finally,the sanitized CSI phase information is used to perform breath detection.In Chapter 4,we propose to utilize the data output from vector network analyzer to perform breath detection and human recognition.The high-quality breath signal ex-tracted from vector network is labeled and utilized to train the neural networks for human recognition.We also propose to use the transfer learning technique to improve recognition accuracy.In Chapter 5,we show the experimental results of the algorithms proposed in this thesis with comprehensive comparison and analysis.The breath detection algorithms proposed in this thesis can provide a solution to the long-term contactless breath monitoring in indoor environments,and thus a new way to continuously monitor the health status of human being.Also,the human recognition al-gorithm can provide home security through the identification of indoor targets.