探討不同慣性感測器位置與運動模式對負荷量與攝氧量之關係

Abstract

前言：過高的運動負荷量可能導致疲勞累積並增加受傷風險。過去有多種運動負荷量測方法，但受到儀器昂貴、場地限制等影響。若可以使用便捷的穿戴式裝置進行監測，則可以幫助教練或選手即時量測運動負荷。目的： 運用感測器量化不同運動模式及不同部位之運動員負荷量。探討運動負荷量與生理參數之間的關係，並建立推估攝氧量的回歸方程式，亦探究不同感測器位置間運動員負荷量的轉換關係。方法： 實驗招募24名健康成年人，進行三種不同運動模式 (跑步、折返跑、籃球) 的測驗。將慣性感測器黏貼至手腕、上臂、肩膀、質心與腳踝，並配戴氣體分析儀與心率帶收集數據。研究採用二因子混合設計變異數分析不同部位感測器與不同運動模式的差異；皮爾森積差相關分析各參數與攝氧量的關係；逐步回歸分析法找出推估攝氧量的最佳方程式。結果：研究結果顯示不同位置感測器與不同運動模式所量測之運動員負荷量彼此間有顯著差異，並存在交互作用。各項運動模式下腳踝的運動員負荷量顯著高於其他部位，而折返跑與籃球在各部位量測的運動員負荷量顯著高於跑步。腳踝的運動員負荷量與攝氧量在各項運動模式下達到中至高度相關。逐步回歸分析表明，腳踝運動員負荷量是推估攝氧量的最佳單一變項，建立的模型具中等解釋力 (R2 =.40∼.55)。不同部位運動員負荷量之間存在中度至高度相關性，並建立了可達中等解釋力 (R2=.40∼.69) 的轉換關係模型。結論： 本研究量化在不同情境下的運動負荷特性，建立以腳踝負荷量推估攝氧量的模型，並提供了不同部位負荷量的轉換關係。此發現有助於教練與運動員更精確地評估訓練負荷，優化訓練效益並降低傷害風險。Background: Training load can lead to fatigue and increased injury risk. Traditional exercise load measurement methods are often limited by expensive equipment, complex procedures, and environmental constraints, making real-time application in practical settings challenging. Wearable inertial sensors offer a convenient solution for immediate load monitoring. Objectives: This study aimed to quantify Player Load using inertial sensors across different exercise modalities and body placements. It also investigated the relationship between Player Load and physiological parameters, established regression equations for oxygen consumption estimation, and explored conversion relationships between Player Load values from different sensor locations. Methods: Twenty-four healthy adults participated in three distinct exercise tests (running, shuttle run, basketball). Inertial sensors were attached to the participants' wrist, upper arm, shoulder, sacrum, and ankle, with simultaneous data collection from a gas analyzer and heart rate monitor. A two-way mixed design ANOVA examined differences in sensor data across body sites and exercise modalities. Pearson's correlation analysis assessed relationships between parameters and oxygen consumption, and stepwise regression identified the best equation for oxygen consumption estimation. Results: Sensor placement and exercise modality significantly influenced PlayerLoad, with a notable interaction. Ankle Player Load was consistently highest across all modalities, while Player Load in shuttle run and basketball was significantly higher than in running across all sites. Ankle Player Load exhibited a high correlation with oxygen consumption in basketball and moderate correlations in running and shuttle run. Stepwise regression identified ankle Player Load as the optimal single predictor for oxygen consumption, with models showing moderate explanatory power (R2 =.40∼.55). Furthermore, moderate to high correlations existed between Player Load values from different body sites, enabling conversion models with moderate explanatory power (R2 =.40∼.69). Conclusion: This study successfully quantified IMU-derived exercise load characteristics in various contexts, developed oxygen consumption estimation models based on ankle Player Load, and provided conversion relationships for Player Load across different body sites. These findings offer convenient, economical, and practical tools for exercise monitoring, aiding coaches and athletes in precise load assessment, optimizing training efficacy, and mitigating injury risk, particularly in practical settings where expensive equipment is often inaccessible.