Less Channel Hitting Measuring System for Table Tennis Monitoring

Abstract

Sports play a crucial role in daily life, and precise monitoring of muscle states during sports is valuable for guiding training programs, preventing injuries, and detecting fouls. Despite significant advancements in monitoring technologies, there are still unexplored areas in detecting various sports actions and processing the associated data. In this study, we present a method to reduce communication channels in a table tennis racket sensor array and use the fabricated monitoring system to detect the player's status through signal analysis and machine learning algorithms. The system includes a flexible piezoelectric PVDF sensor array, Bluetooth chip, and battery. Unlike conventional designs that assign an individual communication channel to each sensor, this approach uses a single striped sensor on each row and column. By analyzing the axis of the sensor with signal, the hitting position and force are determined. After that, a kNN classification algorithm is built with the help of electromyography (EMG) test, aiming to assess the player's status based on their hitting input. The system effectively differentiates various player statuses with an accuracy exceeding 90%. The hardware innovation and signal processing techniques also have potential applications in other wearable sensing fields.