建立自行車騎乘之三維全身肌肉骨骼系統模型

Abstract

隨著環保意識抬頭，節能減碳和健康議題日漸受到重視，腳踏車運動 蔚為風潮。有研究指出，騎腳踏車可降低糖尿病、高血壓以及心血管疾病 風險，故政府近年來不斷地推廣此項運動。然而，腳踏車運動常發生騎乘 傷害。但鑒於過去研究多以忽略髖關節內收外展自由度的二維模型進行實 驗分析或模擬，雖可量化肌肉活化情形、關節角度以及關節受力，但缺乏 實際騎乘時三維動作的真實性。因此本研究目的在於建構出三維全身性肌 肉骨骼系統模型，由現有二維自行車騎乘肌肉骨骼系統模型加入髖關節內 收外展自由度，並透過肢段質量、肢段長度量測等客製化流程以提供客製 化常見慢性運動傷害(例如髕骨股骨疼痛症狀)的評估工具與自行車設計模 擬平台。本研究分為三個階段：第一階段利用肌肉骨骼系統軟體建立三維 模型，並針對騎乘情況進行標記點組設計評估及改良；第二階段針對第一 階段改良設計過之標記點組進行自行車尺寸、人體計測參數、肢段質量、 肌電訊號等量測(實驗)，並利用實驗動作擷取參數進行肌肉骨骼系統驅動並 做逆動力學分析(模型)；第三階段調整模型曲柄中心阻抗以模擬合力矩感測 器訊號，並將實驗之肌電訊號與模型分析出來的肌肉活動情形進行比較。 本研究結果發現，量測到的肌肉活化趨勢和文獻結果大致相符。而模擬的 肌肉活化情形在模型經過曲柄中心力矩阻抗的修正後，和實驗量測之肌電 訊號相比，以股二頭肌和脛前肌有較高的相似性，且受測者間存有差異性。 外側廣肌、股二頭肌和脛前肌在肌肉活化的時間點上，模型分析的結果和 實驗情形較為相近。而模型分析肌肉活化結束時間點，則是股二頭肌較相 近於實驗結果。肌肉活化時間的部分，模型分析的肌肉活化時間相較於實 驗的情況，時間上明顯較短，且外側廣肌和脛前肌的模型分析結果，在不 同的騎乘條件下有顯著差異(外側廣肌：p=0.03，股二頭肌：p=0.004)。本研 究成功建立出三維全身肌肉骨骼系統模型，並針對其進行評估。可用於未 來客製化自行車騎乘評估以及自行車設計。

Following the rise of environmental consciousness, issues related to energy conservation and healthcare have already drawn numerous attentions. Since cycling has been proven by many studies to be feasible to reduce the risk of diabetes, high blood pressure and other cardiovascular diseases, it has become a popular exercise and been promoted by government utmostly. However, cycling injuries are common during cycling process. Former studies usually used two-dimensional model, e.g. without hip abductioin/abduction, to study cycling movements and potential injury mechanisms. Lack of hip abduction/adduction movements during cycling may not provide enough informaiton on real cycling. Therefore, the objective of this research was to develop a three-dimensional (3D) full-body musculoskeletal model for evaluating individualized cycling movements and injury risk, such as patellofemoral syndromes, and for designing new bikes. The study consists of three majoy stages: First, establishing a musculoskeletal model using a musculoskeletal simulation software and designing a marker set for cycling movement measurements; Second, acquiring human motion data, crank torque and electromyography (EMG) signals during cycling, and measuring individual anthropometry for model analysis; Third, modifying established musculoskeletal model by comparing model-predicted pedaling torque and muscle activities with experimental data of pedaling torque and EMG signals. Results showed that the EMG signals are similar to analyzed muscle activation on specific muscles. Besides, High similarity between EMG and model-predicted muscle activities was observed in tibialis anterior (TA) and biceps femoris (BF) muscles with variations between subjects, after matching the model crank torque with experiemntal measurements. Model-predicted vastus lateralis (VL), BF and TA muscle activation timing were closed to experimetnal results, while model-predicted deactivation timing of BF agreed with experimental data.. As for the proportion of activation duration in a cycle¸ model-predicted results were significantly shorter than experimental data with the highest statistical differences occurred in VL and TA (VL: p=0.03, TA: p=0.004). In conclusion, this thesis presents a three-dimensional muscluskeletal model, validated by experiment, and procedures for individualized cycling evaluation. The platform could be used to evaluate individual cycling performance and injury risk, and be employed for bike design.