使用穿戴式裝置放置於手腕評估活動強度

Abstract

前言：身體活動能有效促進身心健康，其強度量化方式多元。腕帶式穿戴裝置或許可成為統一評估工具，但各強度指標在不同運動型態中的表現不一，難以單一指標全面評估。若需涵蓋多元運動型態，仍要開發進階複合指標，並克服硬體限制與演算法簡化等問題，方能充分發揮應用價值。目的：驗證：(1) 腕帶式裝置簡易演算法之可行性；(2) 不同運動情境下適用的強度指標。方法：招募24名健康成年人，使用局部定位系統感測器置於手腕及肩膀處分析速度、PlayerLoad/min (PL/min)、合加速度等數據，智慧手錶配戴於手腕處收取步頻數據，實驗進行跑步、敏捷與跳躍等3種測試，經資料處理成7種不同長度之時間區間，統計分析將以 (1) 斯皮爾曼相關性分析，各測試中合加速度均值 (MA)& 合加速度峰值 (PA) & 步頻 & PL/min與速度之相關性。(2) 使用平均數及信賴區間分級各測試中MA & PA & PL/min於各運動情境的強度區間。結果：(1) 跑步測試中MA與PL/min在所有時間區間均與速度高度相關 (ρ > .9)；PA在1、3、5、10、20秒區間呈高度相關；步頻於1秒區間與速度高度相關。(2) 敏捷測試中MA與PL/min除1秒外所有時間區間均呈高度相關 (ρ > .7)；PA所有時間區間皆呈中度相關。(3) MA、PA在所有時間區間呈中度相關，PL/min則在30與60秒時間區間呈現高度相關 (ρ > .7)。結論：本研究結果證實 MA 與 PL/min 於 60 秒區間可有效評估跑步-變向情境之強度，PL/min 於 60 秒區間則可評估跳躍情境之強度，研究結果可作為腕帶式穿戴式裝置演算法優化依據。

Introduction: Physical activity is essential for promoting physical and mental well-being, and various methods exist to quantify activity intensity. Among them, wrist-worn wearable devices show promise as a unified tool for monitoring physical activity. However, the effectiveness of individual intensity indicators varies across different activity types, making it difficult to rely on a single indicator for comprehensive assessment. To evaluate diverse forms of exercise, advanced composite metrics are required, and it is necessary to overcome the hardware limitations and simplify the algorithms of wearable devices to enhance their practical utility. Purpose: This study aimed to (1) examine the feasibility of simple algorithms using wrist-based devices, and (2) identify the most appropriate intensity indicators for different exercise contexts. Methods: Twenty-four healthy adults were recruited. Local positioning system (LPS) sensors were placed on the wrist and shoulder to collect data on speed, PlayerLoad per minute (PL/min), and absolute acceleration. A smartwatch worn on the wrist recorded step frequency. Participants underwent three types of exercise tests: running, agility, and jumping. The collected data were segmented into seven epochs of varying lengths. Statistical analysis included: (1) Spearman correlation analysis to examine the relationship between speed and the following indicators—mean absolute acceleration (MA), peak absolute acceleration (PA), step frequency, and PL/min; (2) Classification of intensity levels in each exercise context based on mean and confidence epochs values of the indicators. Result: (1) In running test, both MA and PL/min show a high correlation with speed across all time epochs (ρ> .9). The PA showed a high correlation with speed at the 1-, 3-, 5-, 10-, and 20-second epochs. Step frequency was highly correlated with speed only at the 1-second epoch. (2) In the agility test, the MA and PL/min were highly correlated with speed across all time epochs except for the 1-second epoch (ρ > .7). PA showed a moderate correlation across all epochs. (3) The mean and PA demonstrated a moderate correlation across all epochs, while PL/min showed a high correlation specifically at the 30- and 60-second epochs (ρ> .7). Conclusion: The main findings of this study are that MA and PL/min measured over a 60-second epoch can effectively assess the intensity of running or running-with-change-of-direction scenarios. While PL/min over a 60-second epoch can assess the intensity of jump scenarios. These results may serve as a basis for optimizing wrist-based devices algorithms.