Singapore Scientists Advance Metaverse Tracking with Wi-Fi

Scientists at Nanyang Technological University (NTU) in Singapore have developed a method for tracking human movement and activity in metaverse environments using Wi-Fi signals.
 
The research promises to overcome limitations faced by current tracking modalities that rely on body-worn sensors or external cameras. These methods struggle with obstacles and poor lighting conditions and only provide data from specific points on the body.
 
The NTU team’s solution impacts the ability of Wi-Fi signals to penetrate walls and detect minute movements. The data is then fed into AI, which interprets the signals to model full body motion and activities.
 
A key obstacle faced by prior efforts was the need for large labeled datasets to train the AI models. To resolve this, the researchers pioneered an unsupervised learning technique named “MaskFi.”
 
MaskFi allows models to be trained with less data and then iteratively refined until a very high level of accuracy is reached. In testing, the system achieved approximately 97% accuracy in relevant experiments.


​Source: Crypto Times