Researchers at Nanyang Technological University in Singapore have introduced an innovative approach to track human movement for the development of the metaverse. Unlike traditional methods using sensors or cameras, this new method leverages WiFi signals for real-time monitoring.
Traditional tracking methods have limitations, such as capturing information at only one point of the human body or struggling with low-light environments. WiFi sensing, similar to radar, offers a promising alternative by detecting objects in space and capturing various human activities, including heartbeats, breathing patterns, and movement through obstacles like walls.
However, integrating WiFi tracking with artificial intelligence (AI) models presents challenges in training due to the extensive data required. The team introduced “MaskFi,” which utilizes unsupervised learning to train AI models more efficiently. This approach reduces the need for extensive data labeling and focuses on improving the accuracy of predictions.
The MaskFi system achieved approximately 97% accuracy in initial benchmarks, indicating its potential to enable a new modality in the metaverse: providing real-world representations in real-time. This advancement could significantly enhance the immersive experience and capabilities of the metaverse for users worldwide.