Acquisition and cognition information of human body swing

Abstract

Man plays an important role as the moving sensor platform in information systems, including 6G communication and intelligent applications. In particular, personal self-position and attitude are critical for firefighters performing rescues, athletes competing in sports, and the shooter with a shoulder launcher aiming at a target. Due to the swinging of the human body and respiratory fluctuations, the real-time miniature inertial measurement unit (MIMU) data combines body motion noise and degrades navigation precision. Two scenarios are discussed in this paper, including aiming at a stationary and moving target. The angular rate and acceleration of MIMU data for a stationary target are obtained from the turntable and shooter's shoulder, respectively. Data are collected via a stimulated launcher mounted on the shooter's shoulder and the theoretical solution for a moving target. A technology known as error cumulative amplitude spectrum (CAS) is adapted to extract the specific frequency caused by the shooter's body swing, which exhibits low-frequency characteristics. Thus, a mixed filter frame composed of an adaptive notch filter (ANF) and a cubature Kalman filter (CKF) is described in detail to reduce both high- and low-frequency measurement errors. The results of series tests indicate that the acceleration and angular rate errors caused by shooter swing are reduced by more than 98% and more than 81%, respectively, in the stationary and moving cases. As a result, the presented method effectively suppresses body swing and enhances navigation performance.