Real-Time Human Activity Recognition for Elderly VR Training with Body Area Networks

Over the past decade, the number of elderly population is growing enormously and the need of elderly care become more and more important. However, the elderly struggle with cognitive decline. The virtual reality (VR) training of daily living is crucial for improving elderly cognitive functioning. The effective body-tracking devices are key subsystem to immerse the elderly in the VR environment. The body-worn inertial sensor is the bridge to connect physical and virtual environments due to several advantages, such as wearable, light weight, low power consumption and personal privacy. We propose an efficient immersive training method to integrate a VR simulation system with a body area network (BAN). With the customized deep neural network algorithm, the body-worn inertial sensors are capable to recognize the activities of participants and avoid mismatched actions. Moreover, we utilize the neural networks to provide greater access to physical actions of the VR real-time training environment. In this study, we develop and implement a quaternion based deep neural network algorithm for human activity recognition (HAR) and further share the experience on the VR application that has the potential to fulfil immersive VR system on HAR.