| Project/Area Number |
16K06156
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Dynamics/Control
|
|---|
| Research Institution | Toyohashi University of Technology |
|---|
Principal Investigator |
Akiduki Takuma 豊橋技術科学大学, 工学(系)研究科(研究院), 助教 (40632922)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
高橋 弘毅 長岡技術科学大学, 工学研究科, 准教授 (40419693)
|
|---|
| Research Collaborator |
Zhang Zhong
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Project Status |
Completed (Fiscal Year 2018)
|
| Budget Amount *help |
¥4,680,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥1,080,000)
Fiscal Year 2018: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2017: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2016: ¥3,120,000 (Direct Cost: ¥2,400,000、Indirect Cost: ¥720,000)
|
|---|
| Keywords | アトラクタ / 動作特徴 / 個人性 / センサ寄与度 / 特異値分解 / 技能教育支援 / 身体動作データ / 加速度センサ / 運転動作 / 技能教育 / 身体動作解析 / 歩行動作 / 変数選択 / 機械力学・制御 / 記号化 / ソフトコンピューティング |
|---|
| Outline of Final Research Achievements |
In this research, we toward to develop a method to extract and analyze individual characteristics (individuality) such as habit and difference in skill levels, in human body movement from motion data, and apply it to skill education.To construct the extraction method of individual features, we discussed the following two major points: (1) a feature extraction technique using a trajectory attractor on a phase space, and (2) an automatic selection technique of variables constituting the phase space.
As a result, we have shown an analysis process for clarifying what kind of movement difference occurred in which position in and between subjects from multivariate measurement data. In addition, we have confirmed the effectiveness of the proposed method by analyzing the individuality of the walking motion and the left turn motion in the car driving.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究で開発した身体動作データの分析方法は，モデル構築のための大量の学習データや反復計算が不要なことから，少数の実験データからも効率よく，時間伸縮に対しても安定して時系列データ間の類似度を評価できる．このことから，例えば，事故誘発動作など発生が希な動作や，癖や経験の違いにより個人差の大きな動作に対しても，被験者内/間での動作の違いを比較しその差が生じた要因の推定ができる．その結果，学習者に対して動作改善のための着眼点を与えたり，個人に合わせた教習法を検討するなど，技能教育の効果と効率の向上に寄与できると考える．
|
