Enabling non-invasive and real-time human-machine interactions based on wireless sensing and fog computing

In the era of Industry 4.0, human plays an important role in the design, installation, updating, and maintenance of the intelligent manufacturing system. To facilitate natural and convenient interactions between humans and machines, we need to develop advanced human-machine interaction technologies. In this paper, we propose a novel gesture recognition system by integrating the advantages of Doppler radar-based wireless sensing and fog computing, which is able to facilitate non-invasive and real-time human-machine interactions. We first collect and preprocess the dual channel Doppler information (i.e., I and Q signals), and then adopt a threshold detection method to extract gesture segments. Afterwards, we propose a two-stage classification method to recognize human gestures. We implement the system in real-world environments and recruit volunteers for performance evaluation. Experimental results show that our system can achieve accurate gesture recognition with in less than 1 s. Particularly, the average accuracy for motion detection and gesture recognition is 98.6% and 96.4%, respectively.