A Novel Gait Detection Algorithm Based On Wireless Inertial Sensors

Yueming Gao,Ziqin Jiang,Wenshu Ni,Zeljka Lucev Vasic,Mario Cifrek,Min Du,Mang I. Vai,Sio Hang Pun

DOI: https://doi.org/10.1007/978-981-10-4166-2_45

2017-01-01

Abstract:Gait event detection has been widely implemented in realtime gait monitoring devices, orthoses and FES system. Certainly, the latency and the accuracy of the gait even detection under diversities of gait are crucial. However, due to the high detection accuracy usually comes with high time-delay, it is somewhat hard to find a trade-off between high accuracy and low latency. Therefore, this paper presents a real-time algorithm based on wireless inertial sensor placed on the shank for gait-even detection. It combines the use of the cycle-extremum and the updating threshold method to detected the heel-strike (HS), as the minimum of the flexion/extension angle, the toe-off (TO), as minimum of the angular velocity and the mid-swing (MS), as maximum of the angular velocity. The angle and angular velocity were collected from 2 subjects who imitated the patient that suffered from drop-foot for different degrees to validate the algorithm against the wireless inertial measurement system. The results showed that the proposed method achieved comparable levels of accuracy and significant lower detection delays compared with other published methods.