Personalized Human Activity Recognition Based on Integrated Wearable Sensor and Transfer Learning

Abstract

Human activity recognition (HAR) based on the wearable device has attracted more attention from researchers with sensor technology development in recent years. However, personalized HAR requires high accuracy of recognition, while maintaining the model’s generalization capability is a major challenge in this field. This paper designed a compact wireless wearable sensor node, which combines an air pressure sensor and inertial measurement unit (IMU) to provide multi-modal information for HAR model training. To solve personalized recognition of user activities, we propose a new transfer learning algorithm, which is a joint probability domain adaptive method with improved pseudo-labels (IPL-JPDA). This method adds the improved pseudo-label strategy to the JPDA algorithm to avoid cumulative errors due to inaccurate initial pseudo-labels. In order to verify our equipment and method, we use the newly designed sensor node to collect seven daily activities of 7 subjects. Nine different HAR models are trained by traditional machine learning and transfer learning methods. The experimental results show that the multi-modal data improve the accuracy of the HAR system. The IPL-JPDA algorithm proposed in this paper has the best performance among five HAR models, and the average recognition accuracy of different subjects is 93.2%.

From the Laboratory to the Field: IMU-Based Shot and Pass Detection in Football Training and Game Scenarios Using Deep Learning

“…In other application areas, transfer learning for activity recognition is already a vivid field of research [ 62 ]. In addition, it was recently applied to develop personalized HAR models [ 12 ].…”

“…In contrast to conventional methods which require hand-crafted, domain-specific features, deep learning models automatically extract abstract features from sensor signals [ 11 ]. However, the training of neural networks generally requires large labelled datasets to achieve satisfactory performance explaining its rare use for sensor-based performance analysis in field-sports [ 6 , 12 ].…”

Human Activity Recognition with Deep Learning: Methods, Progress & Possibilities

“…The use of air pressure sensor data and the inertial measuring unit for improving the HAR framework is indicated by Fu et al [83] (IMU). This HAR model has a performance of close to 2% higher than other models that are not applicable to sensors.…”

ANTON: Activity Recognition-Based Smart Home Control System

“…The extension of fall detection is the detection of the main activities of the human body. Zhongzheng et al [32] used air pressure sensors and inertial measurement units as the main tools to provide data for model training for activity recognition. To improve the generalization ability of the model, the author uses joint probability domain adaptive method with improved pseudo-labels (IPL-JPDA) for transfer learning.…”