Collection and Classification of Tennis Swings Using a Virtual Racket

Computerized learning systems are popular these days due to the advances in artificial intelligence and decision support. Learning sports using a computer is a new field of research but it requires additional effort in the areas of motion sensing and modeling, and data mining. We are designing a tennis e-learning system using the Nintendo Wii remote as a virtual racket for practicing swings. This work introduces the swing motion data collection process. Classification of the swing data is explored using various techniques such as principal component analysis and K-means clustering. It is evident from the graphical data that different types of tennis swings have dissimilar characteristics in the 3-D space. The distinct envelope shape of the swings can be characterized and differentiated using descriptive statistics. Classification results are presented with emphasis on the swing consistency of tennis learners as well as the similarity of the swing motions which are important in the eventual learning process.