Evidence Hub

Physics-Informed Dual-View YOLO Framework for Tennis Ball Trajectory Prediction

Huang¹ H, Zheng Q, Lin KW, Lin W

2026

Abstract

Abstract The flight behavior of spinning tennis balls constitutes a nonlinear aerodynamic system governed by the coupling among translational motion, rotational dynamics, gravity, and aerodynamic forces. Conventional trajectory prediction methods often rely on simplified projectile assumptions and therefore fail to accurately reconstruct spin-induced nonlinear trajectory evolution under realistic high-speed sports conditions. To address this limitation, this study proposes a physics-informed tennis trajectory reconstruction framework integrating synchronized dual-view motion acquisition, YOLO-based object localization, three-dimensional kinematic reconstruction, and Magnus-effect-informed aerodynamic simulation. Side-view and top-view observations are utilized to estimate key motion parameters, including initial velocity, elevation angle, and lateral motion angle, which are subsequently incorporated into a nonlinear aerodynamic trajectory model considering gravity, drag, and spin-induced lift forces. Experiments were conducted using 40 tennis strokes collected under daytime and nighttime conditions on a red clay court with physically measured landing-point references. The results demonstrate that incorporating Magnus-effect coupling substantially improves longitudinal trajectory reconstruction accuracy. Specifically, the Z-axis mean absolute error decreased from 2.49 m to 1.08 m, while the overall RMSE decreased from 2.82 m to 1.52 m compared with the non-Magnus model.…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free