1. Human Motion Recognition Based on Passive Wireless Backscatter

We use the reflected signal from the passive tag to enable recognition of human motions, which can count the repetitive motion of human body. We design an adaptive counting algorithm, and use the dynamic time wrap (DTW) method to stretch the wireless signal on the time series for comparing. The counting accuracy has nothing to do with the experimental environment due to the adaptive algorithm, the average count error is controlled within 5%, and by using the periodicity of repetitive motions,we can further recognize the types of motions with an accuracy more than 95%.
Compared to other RFID devices, our system can use the widespread Wi-Fi signal, which can be applied to various scenes. Besides, it is different from the radar system, which needn’t to take the initiative to launch the signal. We first use the reflection signal from tag for human perception attempt, and the working range of backscattering signals is relatively short, which reduces the intra-system interference significantly. This enables our system to support parallel motion recognition for multiple persons.


  • Address1: Rm.712, Third Electric Building, West Campus
    Address2: Rm.119, Technology Building, West Campus
    School of Computer Science and Technology
    University of Science and Technology of China
    Hefei, Anhui 230027, China