Development of jump motion evaluation system using wearable sensors and a deep learning technique

2020 Fiscal Year Final Research Report

• Project/Area Number: 19K19948
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 59020:Sports sciences-related
• Research Institution: Osaka City University
• Principal Investigator: Suzuki Yuta 大阪市立大学, 都市健康・スポーツ研究センター, 講師 (90747825)
• Project Period (FY): 2019-04-01 – 2021-03-31
• Keywords: 慣性センサ / 動作分析 / 関節トルク / 地面反力 / ニューラルネットワーク / 長・短期記憶 / カルマンフィルタ

Outline of Final Research Achievements

The purpose of this study was to develop a system to estimate ground reaction forces and lower limb joint moments during vertical and horizontal jumping using inertial measurement units (IMUs) and artificial neural networks. Twelve university students participated in this study. Jump motions and ground reaction forces were measured during vertical and horizontal jumps. In addition, triaxial accelerations and angular velocities of the pelvis, thigh, shank, and foot of right leg were measured using four IMUs. A neural network was developed to estimate the ground reaction forces and joint moments from the data of IMUs. The results of the present study showed the potential of estimating the ground reaction forces and joint moments during both vertical and horizontal jumping using IMUs and artificial neural networks.

Free Research Field

スポーツ科学

Academic Significance and Societal Importance of the Research Achievements

本研究の結果から，慣性センサとディープラーニングにより跳躍動作中の地面反力や関節トルクを精度良く推定できることがわかった．したがって，従来は専門的な分析が必要だった跳躍動作の詳細な評価を，慣性センサを用いることで簡便に行うことが可能となった．今後は継続的な動作評価をもとにしたコンディションやスポーツ障害のモニタリング，本システムを応用した他のスポーツ動作の評価システムの開発などが期待される．