| 研究課題/領域番号 |
22K20519
|
|---|
| 研究機関 | 国立研究開発法人産業技術総合研究所 |
|---|
研究代表者 |
Shen Keli 国立研究開発法人産業技術総合研究所, 情報・人間工学領域, 産総研特別研究員 (80965179)
|
|---|
| 研究期間 (年度) |
2022-08-31 – 2024-03-31
|
| キーワード | Dance adpative skills / Motor coordination / Sports biomechanics / Motion complexity / Deep RL / Motor synergy / Motor injury prevention / Latent behavior |
|---|
| 研究実績の概要 |
In the first year of our project, we focused on analyzing dance motion data using our newly developed low-dimensional representation algorithm. Our goal was to gain a better understanding of the adaptive motor skills, coordination patterns, motion consistency, and complexity of various dance genres, by examining them from both kinematic and kinetic perspectives. We were able to create computational models that can accurately reproduce dance motor skills through deep reinforcement learning, which was a significant achievement. Our approach allowed us to explore the complex interplay between the cognitive and physical aspects of dance, and to gain valuable insights into how dancers develop and refine their skills over time.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
Our research has focused on gaining a deeper understanding of dance motion data, with a particular emphasis on music-induced dance motions. To this end, we have developed new dimension reduction methods that have allowed us to analyze the complexity of dance movements. Our analysis has been based on the identification of synergy patterns and their activations, which has provided valuable insights into the underlying principles of dance motion.
Besides, we constructed a simulation environment for modeling dance training, and developed learning algorithms to facilitate the training process. While the coding and training are still in progress, our work to date represents a significant step forward in the development of tools and techniques for improving dance training and performance.
|
| 今後の研究の推進方策 |
Our research aims to explore the adaptivity of dance motion imitated by Deep Reinforcement Learning (DRL). To this end, we have designed experiments to assess the performance of a humanoid agent as it attempts to transfer dance motions across different environments. Our goal is to observe how well the agent can adapt to new environments, which will provide valuable insights into the adaptivity of imitation dance motion. In addition, we are also investigating the hyper-adaptivity mechanism of the human body in our dance motion imitation tasks. By studying the ways in which humans adapt their movements to different environments, we hope to gain a deeper understanding of the underlying principles of dance motion and to develop new techniques for enhancing dance performance.
|
| 次年度使用額が生じた理由 |
発生した金額は書籍の発注の遅延によるものです。これは、完了した際に書籍の支払いに使用されます。
|
