A human body sitting posture detection method based on channel state information

Abstract

A method for detecting human sitting posture based on channel state information. There are two phases: Off-line training phase and on-line testing phase. In off-line training phase, subcarrier sequence signals are analogous to time series signals and complex network is constructed by visual method to extract network features, meanwhile corresponding statistical features are extracted from the samedata, and the fingerprint database of sitting posture is established by combining two kinds of features. In the online testing phase, we use several machine learning algorithms to detect and judge the sitting posture classification. At the same time, we combine the amplitude and phase information, and synthesize multiple pairs of antennas to further improve the accuracy and stability of posture detection. By adopting the method, the invention can greatly reduce the actual operation cost and overcome the problems and defects caused by the computer vision method, thereby effectively realizing the detection and judgment of the sitting posture of the indoor human body, and realizing the classification accuracy of more than 95% under the optimal condition.

Description

technical field [0001] The invention relates to the field of human sitting posture detection, in particular to a method for detecting human sitting posture based on channel state information. Background technique [0002] Sitting and standing posture has become a common way of people in modern life. White-collar workers operate computers at their desks to complete their work every day, students lean over their desks to study, drivers sit in the cab for a long time to control cars, etc. However, sitting and standing for a long time will have a huge negative impact on people's physical and mental state, especially some bad sitting habits, which not only easily cause myopia, but also easily cause people to suffer from waist and cervical spine diseases and muscle stiffness and strain. various chronic diseases. If it is not well prevented and corrected, it will damage the health and ultimately affect the individual's learning and work efficiency. In addition, studies have shown...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the deficiencies of the existing computer vision human body sitting posture detection methods, which have high requirements for environmental lighting, involve personal privacy violations, and high operating costs, the present invention provides a method that is more adaptable to the environment than traditional computer vision methods, and does not involve A human sitting posture detection method based on channel state information that invades privacy, has higher accuracy, and is cheaper. It can effectively realize the judgment and classification of indoor human sitting posture

Images