Researchers from Singapore developed a new method to track human movements using WiFi signals, allowing for the most detailed projection of physical movements into virtual reality.
Jianfei Yang, Shijie Tang, Yuecong Xu, and Yunjiao Zhou, graduate students from Nanyang Technological University of Singapore, under the guidance of Lihua Xie, Fellow of the Institute of Electrical and Electronics Engineers (IEEE), presented a prototype device for human activity recognition (HAR) and capturing body movements in an environment with limited visibility.
To effectively track human movements, the researchers used a WiFi signal that can be customized to detect a person’s heartbeat and breathing. This method can overcome additional obstacles such as walls or limited visibility. Radio signals similar to WiFi data sending and receiving signals are used to detect objects in space.
According to the research preprint, the device has the potential to bring significant innovations in various fields, such as healthcare and security monitoring. However, the team views the device’s primary use as projecting physical body movements into a Metaverse space.
Current methods of capturing physical movements use sensors on devices, cameras, or a combination of both. But this approach assumes a limited range of action in certain conditions, such as the need for more necessary lighting, making them inoperable. The method developed by the scientists based on WiFi signals can rid modern Metaverse devices of all existing drawbacks and project complex physical activity into virtual space.
Despite all the advantages of using the WiFi-based HAR method, it has one disadvantage. Tuning the equipment requires collecting a huge amount of data, which is possible using artificial intelligence (AI) models, but training them is a painstaking and challenging task.
The researchers proposed using MaskFi, an AI training model based on unlabeled videos, to process WiFi tracking data. This approach allows the AI model to be trained incrementally, starting with a small data set and gradually expanding its range. It’s worth noting that MaskFi archives an accuracy of approximately 97% when processing data.
A recent study showed that the size of the global market for Metaverse projects in healthcare could reach nearly $500 billion within ten years.
