Multisensor data fusion of motion monitoring system based on BP neural network

Abstract

Recently, many researchers have proposed a number of respiratory monitoring methods to monitor the respiratory amplitude and respiratory rate of athletes under different conditions. However, the problem with such methods is that their accuracy is not high. Therefore, this paper proposes a respiratory monitoring method based on a BP neural network combined with multisensor fusion technology. First, we fix multiple sensors on the chest and back of the human body so that we can accurately measure the acceleration and angular velocity changes caused by the contraction and expansion of the chest contour when athletes breathe under different motion conditions. After a coordinate measurement is performed on the data measured by the back measurement unit, the BP network algorithm is used to obtain the acceleration curve under the simple respiratory motion state. Finally, the respiratory frequency and respiratory depth parameters, which are important indicators for evaluating athletes’ physical fitness, can be obtained for the athletes under different motion states. Verification was carried out in an experimental platform, and the experimental results for different postures of the human body were compared with the standard respiratory mask measurement results. The accuracy rate was over 90%, thus realizing the networking, portability, and wearability of the breathing state sensors, which provide real-time accurate measurements.