Research on inspiratory muscle training action recognition based on passive RFID tags and hybrid deep learning

Abstract

Chronic obstructive pulmonary disease (COPD) significantly impacts the quality of life for affected individuals. Inspiratory muscle training emerges as a crucial component in the rehabilitative care of COPD patients. A newly developed respiratory status sensing system is unveiled in an article, utilizing passive radio frequency identification (RFID) positioning technology for non-invasive and remote monitoring of COPD patient activities, achieving recognition accuracy 95.6% in classification tasks. A smart sensing suit was designed, integrating liquid metal tags and passive RFID technology for real-time data collection. At the core of this approach lies the utilization of a CNN-LSTM-AM hybrid deep learning model, precisely identifying respiratory movement patterns and stages. Results from experiments underscore the exceptional effectiveness of the model, particularly in terms of its design and learning capabilities, all substantiated by rigorous performance evaluations. The model’s outstanding performance is attributed to its use of convolutional neural networks (CNNs) for feature extraction, long short-term memory (LSTM) networks for time-series data analysis, and attention mechanisms (AMs) for focusing on essential features. These components collectively enhance the accuracy of identifying respiratory patterns. This study introduces an innovative remote sensing method for COPD patient rehabilitation, aiming to establish a new clinical standard for home-based applications.