WiFi Enhances Human Activity Detection in the Metaverse

Researchers in Singapore have created a novel technique for monitoring human movements through WiFi signals, enabling the most precise translation of physical actions into virtual reality.

Jianfei Yang, Shijie Tang, Yuecong Xu, and Yunjiao Zhou, graduate students at Nanyang Technological University of Singapore, under the mentorship of Lihua Xie, a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), introduced a prototype device for human activity recognition (HAR) that captures body movements in environments with restricted visibility.

To accurately monitor human movements, the researchers utilized a WiFi signal that can be tailored to sense an individual’s heartbeat and respiration. This technique can navigate additional barriers such as walls or limited sightlines. Radio signals akin to WiFi data transmission and reception are employed to identify objects within a space.

The research preprint indicates that the device holds the potential to introduce significant advancements across various sectors, including healthcare and security surveillance. Nonetheless, the team primarily envisions the device’s application in projecting physical body movements into a Metaverse environment.

Existing methods for capturing physical movements rely on sensors attached to devices, cameras, or a combination of both. However, these approaches often assume a restricted range of action under specific conditions, such as inadequate lighting, rendering them ineffective. The method devised by the researchers, based on WiFi signals, aims to eliminate the limitations of current Metaverse devices and facilitate the projection of intricate physical activities into virtual realms.

Despite the benefits of the WiFi-based HAR technique, it does present one drawback. Calibrating the equipment necessitates the accumulation of a substantial volume of data, which can be achieved using artificial intelligence (AI) models, although training these models is a meticulous and demanding process.

The researchers suggested employing MaskFi, an AI training model that utilizes unlabeled videos to analyze WiFi tracking data. This strategy enables the AI model to be trained progressively, beginning with a small dataset and gradually broadening its scope. Notably, MaskFi achieves an accuracy rate of approximately 97% when processing data.

A recent analysis indicated that the global market for Metaverse initiatives in healthcare could approach nearly $500 billion within the next decade.

Сообщение WiFi Helps Human Activity Recognition in Metaverse появились сначала на CoinsPaid Media.