膝關節復健動作之即時辨識嵌入式系統

近年，姿態辨識為一項熱門的技術，許多研究與文獻提出各式的演算法，致力於達到準確的動作姿態辨識或是手勢辨識。大部分演算法所建立的數學模型過於複雜，龐大的計算量較難以利用輕巧的嵌入式系統順利完成。本文使用最長共同子序列 (Longest Common Subsequence, LCS)演算法作為動作辨識之核心演算法，並且實踐於以ATmega2560微處理器所建構之嵌入式系統，達到可攜式、即時且精準的辨識能力。在十字韌帶手術後，藉由活動式膝關節支架進行長期復健治療是目前的標準療法。本研究將單顆加速度感測器置於活動式膝關節支架，經由移動平均法與資料正規化的訊號處理方式，在即時復健動作辨識上具有高度的準確性。為發展辨識動作的多樣化，本研究進而使用雙顆加速度感測器置於膝關節支架上下肢，以增加所獲取的數據資訊，並將兩組感測資料以組合的方式，搭配辨識出多種膝關節復建動作。本文於實驗結果舉出目前膝關節復健中最普遍的七項動作，並提供不同使用者操作，其辨識準確率高達90%以上。說明此復健輔助裝置能廣泛應用於十字韌帶術後復健。

關鍵字

最長共同序列子演算法；移動平均法

並列摘要

Becoming more and more popular in recent years, the study of gesture and posture recognition has provoked suggestions of numerous algorithms in many researches and references, striving for more accurate hand gesture and body posture recognition. The mathematical models built by most algorithms are too complex with amount of calculation too large to be successfully operated by lightweight embedded systems. The Longest Common Subsequence (LCS) is employed in this study as the core algorithm for the recognition of posture movements and the design is realized using an embedded system built with ATmega2560 microcontroller. This way, portability and abilities to instantly and accurately recognize user posture and movements are achieved. Currently, it is part of the standard treatment procedure to use hinged knee braces in a long-term rehabilitation process. Here, a single inertial sensor is taken and attached on a hinged knee brace then signal processing through moving average and data normalization methods are used for real-time recognition and more accurate results of physical therapy (PT) movements. For recognizing a wide range of movements, this study took a step further and attached two separate inertial sensors on the upper and lower parts of the knee brace to increase the details of acquired data. The two sets of sensor data are taken and paired up to identify a wider variety of PT movements. As the experiment results cited the 7 most common knee joint PT movements and presented the data from several different users, it is observed that identification accuracy reach over 90%. This is a sound prove that this PT aid can be widely applied to the rehabilitation of post cruciate ligament surgeries.

並列關鍵字

Longest Common Subsequence Algorithms ； LCS ； Moving Average

參考文獻

[19] 楊晨佑，以RGB-D感測器實現全方向移動機器人自動對位控制，碩士論文，國立臺北科技大學電機工程系研究所，臺北，2013。

[4] C. Zhu, Q. Cheng, and W. Sheng, “Human Intention Recognition in Smart Assisted Living Systems Using A Hierarchical Hidden Markov Model,” IEEE International Conference on Automation Science and Engineering, 2008, pp.253-258.

[5] 洪兆欣，以軌跡辨識為基礎之手勢識別系統，碩士論文，國立中央大學資訊工程研究所，中壢，2006。

[6] C.-F. Lai, S.-Y. Chang, H.-C. Chao and Y.-M. Huang, “Detection of cognitive injured body region using multiple triaxial accelerometers for elderly falling,” IEEE Sensors Journal, 2011, pp.763-770.

[7] 鍾哲民，加速度動作辨識系統之研究及應用，碩士論文，國立成功大學工程科學系，台南，2008。

國際替代計量

膝關節復健動作之即時辨識嵌入式系統

全文下載

主題瀏覽