Publication: SwingPong: Analysis and suggestion based on motion data from mobile sensors for table tennis strokes using decision tree
0
0
Issued Date
2016
Resource Type
File Type
application/pdf
Other identifier(s)
2-s2.0-85015625018
Rights Holder(s)
มหาวิทยาลัยศรีนครินทรวิโรฒ
Bibliographic Citation
ACM International Conference Proceeding Series. (2016), p.-
Suggested Citation
Viyanon W., Kosasaeng V., Chatchawal S., Komonpetch A. SwingPong: Analysis and suggestion based on motion data from mobile sensors for table tennis strokes using decision tree. ACM International Conference Proceeding Series. (2016), p.-. doi:10.1145/3028842.3028860 Retrieved from: https://hdl.handle.net/20.500.14740/4844
Author(s)
Abstract
Table tennis is a sport that everyone can play, regardless of age or gender. It becomes a popular sport since it does not require much space and its equipment is affordable. For beginners, they may not have a chance to get some helpful advice to learn how to control a table tennis ball. Thus, we decided to analyze user's motion data and build a model to develop an application called SwingPong detecting basic table-tennis strokes in order to provide helpful suggestion to players. SwingPong is an android application on a smartphone analyzing user's forehand and backhand strokes and providing suggestion back to the user in order to improve his or her swing. There are two main modules in the application: Module #1: storing angular acceleration and angular velocity of strokes. When a table tennis ball is hit on the racket, the angular acceleration and angular velocity of user's swing are collected using accelerometer and gyroscope sensors in smartphones. We collected the training data set from table tennis national players, Thai youth national team table tennis athletes and coaches. J48 Decision Tree was used in order to build a model of hitting classification. Module #2: providing suggestions. We applied the model from Module#1 to evaluate user's basic strokes and provide alert sound when a ball is hit. After a user hits the ball in a given time, the application analyzes the user's strokes and shows suggestions. The experimental results show that SwingPong is able to recognize strokes whether the ball is good (or in), out, or hits the net. The accuracy of classification of forehand strokes is 77.21% and the accuracy of classification of backhand strokes is 69.63%. In addition, SwingPong can provide helpful suggestions so that users can use the feedbacks to improve their swings. © 2016 ACM.
Subject(s)
Accelerometers
Angular velocity
Decision trees
Feedback
Gyroscopes
Motion analysis
Smartphones
Sporting goods
Sports
Trees (mathematics)
Accuracy of classifications
Analysis and suggestions
Android applications
Angular acceleration
Gyroscope sensors
Mobile sensors
Table-tennis
Training data sets
Decision tables
Angular velocity
Decision trees
Feedback
Gyroscopes
Motion analysis
Smartphones
Sporting goods
Sports
Trees (mathematics)
Accuracy of classifications
Analysis and suggestions
Android applications
Angular acceleration
Gyroscope sensors
Mobile sensors
Table-tennis
Training data sets
Decision tables