In order to analyze the biomechanical characteristics of snowboard cross motion in snowboarding, a rigid body dynamics-based human body analysis model was constructed, and we developed a dual-channel method for biomechanical feature analysis. In the human body dynamics model, bones are represented by straight lines and joints are represented by circles, forming an approximate rigid body representation of the human body in snowboarding snowboard cross motion. In the construction of deep learning networks, two channels were designed to establish a direct connection between historical data and the data to be evaluated. The experimental results show that the proposed methods can effectively analyze different technical movements in snowboarding and snowboard cross, and form curve tracking of the angle changes of each joint of athletes.
Li et al. (Fri,) studied this question.