ACCURATE LOW-COST SENSOR-DRIVEN LOCALIZATION OF TENNIS RACKET IMPACTS

This thesis explores a low-cost method for accurately localizing ball impacts on tennis rackets using consumer-grade sensors. A combined approach leveraging the Newton/Euler method for X-axis estimation and vibration-based analysis for Y-axis estimation is proposed. The Newton/Euler method demonstrated an X-axis error of 17.51%, while the vibration method yielded a Y-axis error of 18%. When integrated, these methods achieved an XY localization error of just 11.44%, significantly outperforming current low-cost alternatives. The system is highly adaptable, requiring only basic racket parameters and a calibration phase to account for individual racket properties. While the study utilized low-velocity impacts, the results suggest that stronger signals in real-world conditions would improve accuracy. Despite challenges such as sensor noise and the influence of hand reaction forces, this work lays the groundwork for a practical, real-time impact localization system with potential applications in player performance enhancement and racket design.