利用穿戴式裝置評估使用者於運動中之體能狀態

Abstract

由於近年人們認知到健康的重要性，各式各樣的主題路跑及馬拉松活動成為 了人們熱愛的活動，不僅可以促進身體的健康，還能與朋友們維繫感情，並從運動 中獲得成就感。但許多路跑造成的運動傷害卻在無形之中悄悄發生，人們經常忽略 了體能的負荷狀態，而導致熱衰竭、脫水、肌肉痙攣、膝蓋疼痛、扭傷及拉傷等。 因此，如何協助人們在享受路跑之餘，避免運動傷害的發生成為了一個重要的課題。 許多知名大廠也積極參與運動相關設備的研究及產品開發，推出各式各樣的穿戴 式設備， 但產品主要針對如何激勵使用者運動， 將運動結合遊戲化因子 (Gamification)，並給予虛擬分數(Rewards)，讓好友間彼此鼓勵及競爭，卻忽略了 運動安全上的問題。本篇論文提出一個藉由穿戴式裝置評估跑者體能的系統，跑者 只需穿上此款於腳底裝上壓力感測器(Force-sensor resistor)之穿戴式裝置，即可開 始進行跑步運動。系統會將跑者剛開始跑步之一段時間的壓力特徵值數據做為體 能良好之參考依據，並持續紀錄壓力特徵值，再經由如SVM 及LIBLINEAR 機器 學習之方式分類及評估體能消耗，給予即時的體能剩餘量參考資訊，在跑者體能即 將耗盡時，主動提醒跑者做適當休息，以免造成自身運動傷害。實驗結果顯示，基 於機器學習之評估方法能符合跑者之自行體能評估值，且跑者能在跑步結束後，獲 知整體體能消耗趨勢，進而了解自身體能衰退情形並於跑步時有效分配體能。

Recently, due to the prevalence of the long-distance running and Marathon, lots of featured running events have been held. Participating in a running event can not only improve our health but also bond with our friends. Nevertheless, there are lots of injury that may occur when people enjoy in the running events, such as heart exhaustion, muscle spasm, knee injuries, dehydration and strain. Therefore, how to avoid injury from running becomes an important issue that is worthy of be discussed. In this thesis, we propose a system which can use wearable devices to evaluate the stamina. System will records the normalized pressure features collected during the beginning of the run as the energetic reference, then classify and evaluate the stamina through machine learning methods, such as SVM and LIBLINEAR. The system can give runners the information of stamina in real-time and alarm the runners when their stamina is about to run out. This way, runners can have appropriate rest before they get injured. The experimental results shown that the evaluation method based on machine learning is in accordance with the runner’s feeling, also, runners will obtain the trend of stamina consuming when they finish running.