步幅分割的挑战及其在步态障碍中的应用

Challenges of Stride Segmentation and Their Implementation for Impaired Gait.

作者信息

Bobic Vladislava N, Djuric-Jovieic Milica D, Radovanovic Saa M, Dragaevic Nataa T, Kostic Vladimir S, Popovic Mirjana B

出版信息

Annu Int Conf IEEE Eng Med Biol Soc. 2018 Jul;2018:2284-2287. doi: 10.1109/EMBC.2018.8512836.

DOI:10.1109/EMBC.2018.8512836

Abstract

Stride segmentation represents important but challenging part of the gait analysis. Different methods and sensor systems have been proposed for detection of markers for segmentation of gait sequences. This task is often performed with wearable sensors comprising force sensors and/or inertial sensors. In this paper, we have compared four different methods for stride segmentation based on signals collected from force sensing resistors, accelerometers and gyro sensors. The results were evaluated on 15 healthy and 15 patients with Parkinson's disease, and expressed in terms of number of imprecisely, missed or wrongly detected gait events, as well as temporal absolute error. It was established that the methods using the inertial data, provide results with up to 12% higher error rate comparing to detection from force sensing resistors.

摘要

步幅分割是步态分析中重要但具有挑战性的部分。已经提出了不同的方法和传感器系统来检测用于步态序列分割的标记。这项任务通常使用包括力传感器和/或惯性传感器的可穿戴传感器来执行。在本文中，我们基于从力敏电阻器、加速度计和陀螺仪传感器收集的信号，比较了四种不同的步幅分割方法。对15名健康人和15名帕金森病患者的结果进行了评估，并以不精确、遗漏或错误检测的步态事件数量以及时间绝对误差来表示。结果表明，与用力敏电阻器进行检测相比，使用惯性数据的方法提供的结果错误率高出12%。