Preliminary study on the badminton backhand skill analysis based on the tactile glove

Many badminton amateurs lack professional and scientific guidance. With the advancement of technology, intelligent and quantitative approaches have been applied to badminton training. This paper presents a novel method for intelligently analyzing the badminton backhand skills based on a self-developed tactile glove. First, a dataset of stroke tactile data including four high-level badminton players performing four different backhand skills is built. Second, a neural network model is proposed to classify these backhand strokes, achieving the classification accuracies of 99.35%, 99.75%, 99.87%, and 98.87% for each skill, with an overall accuracy of 99.46%. Third, the contribution of each finger to the backhand skills is analyzed. For example, in active straight-shot, the thumb makes the greatest contribution, followed by the index finger, while the left three fingers just play a supporting role. These analysis results are of great importance for amateurs not only to improve their badminton performance but also to protect them from some injuries caused by the wrong finger exerting manners. In conclusion, this paper demonstrates the potential of AI-based methods in enhancing badminton training, which paves a promising road to intelligent sports.