Recognition of Abnormal Gait Based on Vibration Signal

The recognition of abnormal gait is of great help to the health care of the elderly. Existing related research mainly uses image acquisition equipment or wearable equipment to obtain relevant feature information for identification. Most of these methods are invasive or have high operational requirements for users. This paper studies and realizes a system prototype based on the detection of abnormal gait and fall based on foot vibration signals as the source of identification. This paper first designs a multi-sensor cooperative signal acquisition method to achieve a large range of signal acquisition, and separate the effective part from it as the active element. After the collected active elements are denoised, an improved dynamic time warping algorithm (DTW) is used to calculate the abnormal index representing the difference between active elements, then the abnormal index is classified by the K nearest neighbor (KNN) algorithm, and the inferred value that initially characterizes the user's gait is obtained. The inferred value is further processed by hidden Markov model (HMM) to identify the user's gait. The experimental results show that the method proposed in this paper can effectively identify abnormal gait in different gait modes, with an accuracy of 96% in a stable environment, and an accuracy of 94% in an environment with unstable floors. © 2023 Chinese Institute of Electronics. All rights reserved.