Description:
- Senior care
- Rehabilitation and physical therapy
- Personal fitness and life-logging
Abstract
To combat the signal degradation seen in typical radio wave-based wearable sensor networks, researchers at USC have proposed a novel wireless system based instead on magnetic induction (MI). A set of transmitter coils and receiver antennas are mounted throughout the body. Any subsequent motion changes the distance of the transmitters relative to the receivers which in turn affects the corresponding forward voltage gain between the pairs. These measurements are then integrated with machine learning techniques to model a wide range of human motions with an 88% success rate.
Benefits
- Effective: Signals are not attenuated in vicinity of human body
- Accurate: Up to 88% success rate at identifying complex human motion
Market Application
Human activity recognition (HAR) systems, which detect human motion and behaviors in complex, real-world settings, are used in applications ranging from rehabilitation to personal fitness and senior care. Though several HAR platforms exist, the most common and attractive are based on wearable sensors (such as accelerometers, gyroscopes, and magnetometers) that attach directly to users and wirelessly transmit data to processing nodes. Conventional state-of-the-art wireless sensor networks utilize radio wave propagation for signal transmission. Such approaches, however, are not ideal for the human body as it is a lossy medium that significantly attenuates radio waves.
Publications
Golestani, Negar, and Mahta Moghaddam. "Theoretical modeling and analysis of magnetic induction communication in wireless body area networks (WBANs)." IEEE Journal of Electromagnetics, RF and Microwaves in Medicine and Biology 2.1 (2018): 48-55. https://doi.org/10.1109/JERM.2018.2810603
Stage of Development
- The MI HAR system has been evaluated on generated synthetic motion and found to have an 88% success rate
- Available for licensing