A non-contact sensing method for bodybuilding movement monitoring

Abstract

The application discloses a non-contact fitness exercise monitoring method. This aspect includes using a receiving device having at least two antennas to simultaneously receive signals from a transmitting device having at least one antenna in a wireless transmission manner; measuring channel state information (CSI) of the wireless transmission, and by processing the CSI, obtaining a CSI signal stream, Doppler velocity flow and Doppler displacement flow; cutting motion based on obtained Doppler displacement flow; cutting motion based on said, from CSI signal flow, Doppler velocity flow and Doppler displacement flow Extracting classification features; classifying the exercise based on the classification features, thereby determining the type of fitness exercise.

Description

technical field [0001] The invention relates to a fitness exercise monitoring technology, in particular to a non-contact fitness exercise monitoring method based on commercial wireless local area network equipment. Background technique [0002] Unarmed fitness is not borrowing any equipment and tools for fitness activities, and uses your own body as a weight-bearing exercise, such as sit-ups, squats, push-ups, etc. Non-contact bodybuilding movement monitoring refers to the precise monitoring of the bodybuilder's bodybuilding movement without any contact with the equipment, and gives feedback and evaluation of the movement situation, such as exercise type, Indexes such as exercise volume and movement standard degree. At present, wireless local area network (WiFi) devices such as mobile phones, tablet computers, laptops, and smart watches have widely existed in our daily life. If commercial WiFi devices are used to realize contactless physical fitness monitoring, exercisers w...

AI Technical Summary

Problems solved by technology

[0003] In 2015, China's Wei Wang proposed a solution at the MobiCom conference in 2015, based on the understanding and modeling of WiFi signals for human activity recognition (Understanding and Modeling of Wi-Fi Signal Based Human Activity Recognition.MobiCom15.pp.65 –76.), by using the channel state information obtained from the commercial WiFi network card (Channel State Information (CSI) reflects the amplitude and phase changes of the signal from the transmitting end to the receiving end), extracting from the CSI amplitude signal can reflect The signal characteristics of human speed are used to realize the recognition of 9 kinds of behaviors (running, walking, sitting, etc.), but because the Wi-Fi amplitude signal is greatly affected by noise, and the amplitude signal can only reflect the human speed in a coarse-grained manner, It is impossible to know the direction of the speed, so this method can only identify several behaviors of people at a coarse-grained level, and cannot realize the identification of fine-grained fitness exercises, statistics of exercise volume, and achieve the purpose of exercise monitoring

Although these two methods use the channel state information, that is, the phase information of CSI, and can estimate the magnitude and direction of human motion velocity at the same time, they regard the moving target as a whole and can only identify the direction and position of the moving target in a coarse-grained manner. It is impossible to know the fine-grained information such as the target's current exercise type, exercise state, and exercise amount

Images