Biomedical Signals & Systems Group, University of Twente, Enschede, The Netherlands.
Med Eng Phys. 2010 May;32(4):287-97. doi: 10.1016/j.medengphy.2009.10.014. Epub 2010 Feb 11.
An Inertial Gait Phase Detection system was developed to replace heel switches and footswitches currently being used for the triggering of drop foot stimulators. A series of four algorithms utilising accelerometers and gyroscopes individually and in combination were tested and initial results are shown. Sensors were positioned on the outside of the upper shank. Tests were performed on data gathered from a subject, sufferer of stroke, implanted with a drop foot stimulator and triggered with the current trigger, the heel switch. Data tested includes a variety of activities representing everyday life. Flat surface walking, rough terrain and carpet walking show 100% detection and the ability of the algorithms to ignore non-gait events such as weight shifts. Timing analysis is performed against the current triggering method, the heel switch. After evaluating the heel switch timing against a reference system, namely the Vicon 370 marker and force plates system. Initial results show a close correlation between the current trigger detection and the inertial sensor based triggering algorithms. Algorithms were tested for stairs up and stairs down. Best results are observed for algorithms using gyroscope data. Algorithms were designed using threshold techniques for lowest possible computational load and with least possible sensor components to minimize power requirements and to allow for potential future implantation of sensor system.
我们开发了一种惯性步态相位检测系统，以替代目前用于触发足下垂刺激器的脚跟开关和脚踏开关。我们测试了一系列单独和组合使用加速度计和陀螺仪的四个算法，并展示了初步结果。传感器被放置在上小腿的外侧。测试是在一位患有中风并植入足下垂刺激器的患者身上采集的数据上进行的，该刺激器使用当前的触发装置，即脚跟开关进行触发。测试的数据包括各种代表日常生活的活动。在平坦表面行走、不平坦地形和地毯行走中，检测率达到了 100%，并且算法能够忽略非步态事件，如重心转移。与当前的触发方法（脚跟开关）进行了定时分析。在对参考系统（即 Vicon 370 标记和力板系统）进行评估脚跟开关定时后，初始结果表明当前触发检测与基于惯性传感器的触发算法之间具有密切的相关性。我们还针对上下楼梯进行了算法测试。使用陀螺仪数据的算法观察到了最佳的结果。算法是使用阈值技术设计的，以实现最低的计算负载和最小的传感器组件，从而最小化功率需求，并为未来潜在的传感器系统植入做好准备。
