|제목(국문)||기계학습을 이용한 주파수 분석을 통한 퍼팅 판별 알고리즘|
|제목(영문)||Golf Putting Detection through Frequency Analysis using Machine Learning|
권혁준 (Hyuk-Jun Kwon ,한국과학기술원 기계공학과 ) ▷공저자네트워크등록하기
이두용 (Doo Yong Lee ,한국과학기술원 기계공학과 ) ▷공저자네트워크등록하기
박수경 (B.S.K ,한국과학기술원 기계공학과 ) ▷공저자네트워크등록하기
Auto-scoring related wearable devices for golf have emerged to meet consumer demands. Wrist-worn devices capable of automatic scoring using inertial sensors are providing user convenience. Impact of putting is, however, relatively small , and it is difficult to extract the putting signal due to influence of the hand grip. In this study, we designed a putting-discrimination algorithm for auto-scoring using inertial sensors. The putting-determination consists of selecting a putting phase from arbitrary motion, and extracting a putting signal in the phase. The putting phase was determined by using Convolution Neural Network (CNN). It was determined based on manual labeling by inputting the acceleration and angular velocity data. Putting-signal extraction exploits the phenomenon that vibration caused by natural frequency occurs even for a small impact. And the putting signal is extracted by designing a bandpass filter which passes only the natural frequency in the putting phase. The natural frequency of the putting is analyzed by assuming the putter as a beam, The parameters are estimated to identify the grip boundary conditions. The impact is derived through the frequency analysis, and experimentally verified. The method presented can be employed to increase accuracy of the wrist-type automatic scoring devices for other sports.
|keyword||골프 퍼팅 (Golf putting), 모션 추출 (Motion detection), 기계 학습 (Machine Learning), 주파수 분석 (Frequency Anaylsis)|
|저널명||대한기계학회 추계학술대회 ▷관련저널보기|