Wearable motion sensor based phasic analysis of tennis serve for performance feedback

Recent trends show that wearable devices embedded with high-range sensors play a significant role in tracking health and fitness, including sports activities. For crucial analysis of tennis as a sport, this paper describes a serve analytics engine that provides feedback to players for enhancing their serve performance while preventing potential injuries. By utilizing the information of inertial sensors from the wrist of a player and using serve kinetics, the engine segregates sensor signals into various serve keypoints: start, trophy pose, cocking position, impact and finish. Motion between these keypoints constitutes serve phases like backswing, pronation, follow-through, which are compared against corresponding phases of professionals or related statistics in biomechanical studies. Such comparisons using machine learning techniques enable us to provide players with insights into their playing styles and corrective feedback.