Walking Posture Classification via Acoustic Analysis and Convolutional Neural Network
Yuanying Qu, Xinheng Wang · 2022
Research on activities of daily living (ADL) con-tinues to attract scientists due to the prospection. Walking is a unique biological characteristic which is an indispensable activity. The related research has potential applications in the vast fields or scenes confronted in daily life. Examples include human-computer interaction, behaviour assessment, emergency search and rescue, and healthcare. This paper proposes walking posture classification based on acoustic analysis and the lightweight convolutional neural network. The findings indicate that the classification accuracy can reach 93 %.