Improving energy expenditure efficiency in tennis through robotic ball recovery and delivery system

Engineering of Sport 15 - Proceedings from the 15th International Conference on the Engineering of Sport (ISEA 2024)

Thus far, tennis is played with the assistance of ball attendants or personal effortsin returning balls to hand. Yet this method is either expensive or inefficient in terms of time and energy, a leading factor in Tennis having one of the highest playing downtime among other sports. Tennis is split into playing and returning the ball back after a point. Previous efforts have been made in modernizing tennis by implementing Tennis robots that picks the balls up, making use of spatial recognition and AI to recognize tennis balls. However, they face limited functionality in matches due to ability to only collect and hindering of playing area. In this work, we present a robotic solution to automate the ball collection and delivery to the players on court based on an AI-enabled ball launcher that is capable of tracking and locking to the player, as well as, tossing the ball on demand via body gesture recognition. The system includes a set of nets, ball channels and a ball lift to enable a fully automated ball reloading to the robotic ball launcher. For the base pose detection, we utilize Mediapipe framework and added another layer of feature extraction algorithm which are used to train an XGBoost model to protect the system against false tracking and locking of the designated player, especially in the case of doubles game, and false recognition of gestures.