Title:RapidPD: Rapid Human and Pet Presence Detection System for Smart Vehicles via Wi-Fi

Abstract:Heatstroke and life threatening incidents resulting from the retention of children and animals in vehicles pose a critical global safety issue. Current presence detection solutions often require specialized hardware or suffer from detection delays that do not meet safety standards. To tackle this issue, by re-modeling channel state information (CSI) with theoretical analysis of path propagation, this study introduces RapidPD, an innovative system utilizing CSI in subcarrier dimension to detect the presence of humans and pets in vehicles. The system models the impact of motion on CSI and introduces motion statistics in subcarrier dimension using a multi-layer autocorrelation method to quantify environmental changes. RapidPD is implemented using commercial Wi-Fi chipsets and tested in real vehicle environments with data collected from 10 living organisms. Experimental results demonstrate that RapidPD achieves a detection accuracy of 99.05% and a true positive rate of 99.32% within a 1-second time window at a low sampling rate of 20 Hz. These findings represent a significant advancement in vehicle safety and provide a foundation for the widespread adoption of presence detection systems.