Gait-based identification for elderly users in wearable healthcare systems

The increasing scope of sensitive personal information that is collected and stored in wearable healthcare devices includes physical, physiological, and daily activities, which makes the security of these devices very essential. Gait-based identity recognition is an emerging technology, which is increasingly used for the access control of wearable devices, due to its outstanding performance. However, gait-based identity recognition of elderly users is more challenging than that of young adults, due to significant intra-subject gait fluctuation, which becomes more pronounced with user age. This study introduces a gait-based identity recognition method used for the access control of elderly people-centred wearable healthcare devices, which alleviates the intra-subject gait fluctuation problem and provides a significant recognition rate improvement, as compared to available methods. Firstly, a gait template synthesis method is proposed to reduce the intra-subject gait fluctuation of elderly users. Then, an arbitration-based score level fusion method is defined to improve the recognition accuracy. Finally, the proposed method feasibility is verified using a public dataset containing acceleration signals from three IMUs worn by 64 elderly users with the age range from 50 to 79 years. The experimental results obtained prove that the average recognition rate of the proposed method reaches 96.7%. This makes the proposed method quite lucrative for the robust gait-based identification of elderly users of wearable healthcare devices.