de Ruiter Cornelis J, van Oeveren Ben, Francke Agnieta, Zijlstra Patrick, van Dieen Jaap H
Faculty of Behavioural and Movement Sciences, Research Institute MOVE, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands.
PLoS One. 2016 Sep 20;11(9):e0163023. doi: 10.1371/journal.pone.0163023. eCollection 2016.
The number of validation studies of commercially available foot pods that provide estimates of running speed is limited and these studies have been conducted under laboratory conditions. Moreover, internal data handling and algorithms used to derive speed from these pods are proprietary and thereby unclear. The present study investigates the use of foot contact time (CT) for running speed estimations, which potentially can be used in addition to the global positioning system (GPS) in situations where GPS performance is limited. CT was measured with tri axial inertial sensors attached to the feet of 14 runners, during natural over ground outdoor running, under optimized conditions for GPS. The individual relationships between running speed and CT were established during short runs at different speeds on two days. These relations were subsequently used to predict instantaneous speed during a straight line 4 km run with a single turning point halfway. Stopwatch derived speed, measured for each of 32 consecutive 125m intervals during the 4 km runs, was used as reference. Individual speed-CT relations were strong (r2 >0.96 for all trials) and consistent between days. During the 4km runs, median error (ranges) in predicted speed from CT 2.5% (5.2) was higher (P<0.05) than for GPS 1.6% (0.8). However, around the turning point and during the first and last 125m interval, error for GPS-speed increased to 5.0% (4.5) and became greater (P<0.05) than the error predicted from CT: 2.7% (4.4). Small speed fluctuations during 4km runs were adequately monitored with both methods: CT and GPS respectively explained 85% and 73% of the total speed variance during 4km runs. In conclusion, running speed estimates bases on speed-CT relations, have acceptable accuracy and could serve to backup or substitute for GPS during tarmac running on flat terrain whenever GPS performance is limited.
能够提供跑步速度估算值的市售足部传感器的验证研究数量有限，且这些研究都是在实验室条件下进行的。此外，用于从这些传感器获取速度的内部数据处理和算法是专有的，因此并不明确。本研究探讨了利用足部接触时间（CT）来估算跑步速度，在全球定位系统（GPS）性能受限的情况下，CT有可能作为GPS的补充手段。在自然地面户外跑步且GPS处于优化条件下时，通过附着在14名跑步者足部的三轴惯性传感器测量CT。在两天内，通过不同速度的短跑来建立跑步速度与CT之间的个体关系。随后，利用这些关系预测在中途有一个转折点的4公里直线跑步过程中的瞬时速度。将4公里跑步过程中，对连续32个125米间隔分别测量的秒表得出的速度作为参考。个体速度与CT的关系很强（所有试验的r2>0.96），且在不同日期之间保持一致。在4公里跑步过程中，由CT预测速度的中位数误差（范围）为2.5%（5.2），高于GPS的1.6%（0.8）（P<0.05）。然而，在转折点附近以及第一个和最后一个125米间隔期间，GPS速度的误差增加到5.0%（4.5），且变得比CT预测的误差2.7%（4.4）更大（P<0.05）。两种方法（CT和GPS）都能充分监测4公里跑步过程中的小速度波动：CT和GPS分别解释了4公里跑步过程中总速度变化的85%和73%。总之，基于速度与CT关系的跑步速度估算具有可接受的准确性，并且在平坦地形的柏油路面跑步时，只要GPS性能受限，可用于备份或替代GPS。
