Analysis of Badminton Motion Trajectory Algorithm Based on Neural Network

Abstract

Badminton is a competitive sport with high real-time requirements. To build a badminton robot and enable human-robot sparring, real-time dynamic tracking of fast-moving badminton trajectories is essential. The tracking problem of dynamic targets is widespread in various fields, including industrial production, daily life, and military applications. Existing badminton trajectory planning algorithms face challenges in accurately tracking moving targets and evaluating the stability of badminton flight paths. To enhance the badminton trajectory planning capability, this paper presents a neural network-based algorithm for badminton trajectory prediction. Firstly, a badminton aerodynamic model is established based on the flight characteristics of a badminton shuttlecock. Then, the motion trajectory planning constraint parameters are analyzed for the parameters involved in the dynamic model. The neural network is introduced to predict the badminton trajectory and facilitate accurate tracking of the badminton path. Experimental results demonstrate that the proposed method can effectively track the dynamic path of a badminton shuttlecock in real-time, reduce the deviation of drop-off distances, and improve the accuracy of drop-off prediction.