mmSense: Multi-Person Detection and Identification via mmWave Sensing

Abstract

In recent years, millimeter-wave (mmWave) is becoming a significant component of the next-generation wireless communication due to its up to 7 Gbps transmission rate. In addition to the communication benefits, the unique sensing feature of mmWave attracts more attention. Nowadays, the services of human detection and identification are needed in numerous application scenarios, such as smart home and smart industry. The RF-based sensing techniques, especially WiFi-based, are widely utilized in human detection and identification. However, these work either require humans to carry devices or cannot detect and identify multiple people simultaneously. In this paper, we propose mmSense, a device-free multi-person detection and identification framework, which exploits the unique mmWave sensing features. First, we utilize the properties of directionality, impenetrability, and reflection of 60 GHz signal for objects to fingerprint the environments. Based on the generated environment fingerprints with and without human presence, mmSense can detect and localize the presence of multiple people simultaneously via the LSTM-based classification model. Moreover, we propose a novel approach to use humans' outline profile and vital signs to identify multiple people by using 60 GHz reflected signals of the human body. We conduct extensive experiments to demonstrate the low-cost and effectiveness of our approach.