@article{oai:ipsj.ixsq.nii.ac.jp:00218922,
 author = {田中, 毅 and 馬込, 卓弥 and 山口, 徹 and 合田, 徳夫 and Takeshi, Tanaka and Takuya, Magome and Touru, Yamaguchi and Norio, Goda},
 issue = {7},
 journal = {情報処理学会論文誌},
 month = {Jul},
 note = {スポーツ選手強化へIT技術の導入が進み，特にウェアラブルセンサを活用した選手の怪我予防が期待される．しかし，従来のセンサデータによる怪我のリスク評価は人の経験による判断や週単位などの異常検知に基づき，現場で必要な日々の客観的な評価が困難であった．本研究では日単位などの細かな時間分解能で傷害予兆検知を目的とし，選手の加速度・GPSデータから動作種別を自動判別することにより条件の揃ったデータを抽出し，1回のトレーニング中の動きから怪我が発生しやすい特徴を抽出する手法を提案した．提案手法でU-18サッカーチームの練習時の動きを3日間計測して分析した結果，選手の前方ランニング時の左右バランスを示す指標において，計測後の受傷者と非受傷者の間に有意差があり，日単位の怪我リスク評価に向けて見通しを得た．, To manage training for prevention of player's injury, daily quantitative assessment of player's condition with tracking system is required. However, the conventional index of injury risk calculated from tracking data is based on weekly anomaly detection. Therefore, to realize daily assessment of sports injury, we propose new approach to extract more detail motion features from tracking data of field sports. Our approach is extracting motion features in section of specific basic activity only from tracking data automatically. By focusing the daily variation of same training and activity, we can extract detail motion features correlated to daily player's condition. For evaluation, we measured tracking data of the players in U-18 football team for 3 days. In this experiment, the standard deviation of mediolateral acceleration of forward running in game training is statistically significant different between the injured players and the other players. We confirmed that our proposed method can extract detail motion features to evaluate daily injury risk from tracking data.},
 pages = {1321--1330},
 title = {スポーツ傷害のリスク評価に向けた運動データの特徴抽出手法},
 volume = {63},
 year = {2022}
}