運用慣性測量單元開發多剛體物件於空間中之姿態感測系統

隨著科學的進步以及各種日新月異的技術，人們的生活有了高速的發展，然而隨著知識的拓展，在各領域的所需的專業知識越來越多，而且在使用功能越加強大但是其操作難度也越加困難的工具時，或許是受限於數學上的嚴謹抑或是其物理上的限制，人們往往無法相當直觀地去使用這些工具，甚至還需要花上相當長的一對時間去學習。因此，人機互動和虛實整合的概念在此便相當地重要，要如何讓人的各種感受以及控制能力透過科技產品得以延伸，又要如何讓這一些科技產物能夠快速、精確地辨別人的各種動作與表現，在最近幾年來變得相當地被重視，為了實現此一概念，其中要如何感測人的姿體動作並明確地辨識其動作意圖，將會在當中扮演著相當重要的環節。如何使人與機械之間的互動更為簡單與直觀，便是此研究的核心動機。本研究運用微機電慣性測量單元，建立一動態擷取技術，並開發一套能夠擷取物體於空間中之準確姿態，量測表現與光學式編碼器進行比較後，方均根誤差約1度；標準差為0.75度，根據數據顯示此項技術足以應用許多應用上。期望未來此項技術能結合各領域並運用在不同用途上，能夠完成更多、更加有趣的事情。

關鍵字

慣性測量單元；姿態感測；動作擷取

並列摘要

With the progress of science and a variety of ever-changing technology, people's life develop dramatically. However, people need more and more expertise in various fields because of the expansion of knowledge. When using powerful but complex tools, people often can not quite intuitive to use these tools , and even takes on quite a long time to learn. It is probably because people require more expertise to use it. Thus, human-computer interaction and the concept of the Cyber-Physical System will be quite important. How to extend people’s feeling and control ability and how to develop technologies to identify people’s motion become more and more important in recent years. In order to answer these questions, sensing people’s postures will play an important role. How to make the interaction between people and machine more simple and intuitive is the motivation of this research. This research uses MEMS type inertial measurement units and comes up the algorithm to estimate the orientation multi-rigid-body. The performance shows that the root-mean-square error is 1 degree and the standard deviation is 0.75 degrees compared with the optical encoder. This data shows that this research is enough to be applied to various applications.

並列關鍵字

Inertial measurement unit ； posture estimation ； motion capture

參考文獻

[1] G. Schirner, D. Erdogmus, K. Chowdhury, and T. Padir, "The future of human-in-the-loop cyber-physical systems," IEEE Computer, pp. vol. 46, no. 1,pp. 36–45, 1 2013.

[2] "Human-CPS," University of California, Berkeley, [Online]. Available: http://human-cps.eecs.berkeley.edu/.

[3] V. Koltun, Motion Capture, Stanford University, 2009.

[7] Miller, N,Jenkins, O.C. , Kallmann, M. , and Mataric, M.J., "Motion capture from inertial sensing for untethered humanoid teleoperation," in Humanoid Robots, 2004.

[11] R. Mosher, "Industrial Manupulators," Scientific American, pp. vol. 211, issue 4, pp. 88-95, 10 1964.

國際替代計量

運用慣性測量單元開發多剛體物件於空間中之姿態感測系統

全文下載

主題瀏覽