IMU Based Human Activity Recognition

IMU-based human activity recognition (HAR) aims to identify human actions using data from inertial measurement units (accelerometers and gyroscopes) embedded in wearable devices. Current research focuses on improving accuracy and robustness through advanced deep learning architectures like Transformers and novel neural networks (e.g., Kolmogorov-Arnold Networks), as well as exploring multi-modal fusion with other sensor data (e.g., ambient light sensors) and leveraging large language models for activity interpretation. This field is significant for its potential applications in healthcare monitoring, sports performance analysis, and human-computer interaction, driving advancements in both algorithm design and data annotation techniques.