Eyes-Free, Touch-Free, and Worry-Free: Repurposing Wireless Chargers for Safer In-Car Gesture Interaction

As vehicles evolve into digital living spaces, the demand for seamless human-vehicle interaction has surged. However, traditional interaction modalities impose critical safety and usability burdens. Touchscreens require visual attention (eyes-on), compromising driving safety, while voice assistants, although eyes-free, suffer from inherent latency, high cognitive load, and the intrusive disruption of cabin audio and social conversations. Mid-air gestures offer a promising shortcut for silent, immediate, and eyes-free control. Yet, enabling robust gesture sensing remains a challenge. Vision-based solutions can raise privacy and illumination concerns in some in-cabin deployments; millimeter-wave radar often increases hardware cost and integration complexity; and acoustic sensing can be affected by cabin noise and multipath interference in the in-cabin environment. To bridge this gap, we present MagHarp , a novel system that repurposes the ubiquitous in-car wireless charging pad for active gesture recognition. We observe that the human hand acts as a capacitive bridge within the near-field region, subtly modulating the charging signals. By treating the existing Wireless Power Transfer (WPT) hardware as a near-field electromagnetic sensor, we avoid modifying the charger or smartphone. MagHarp captures these minute impedance perturbations without modifying the transmitter. We design a robust pipeline featuring a lightweight CNN-BiLSTM network that helps distinguish driver gestures from vehicle-related disturbances and environmental interference. Extensive experiments across 5 chargers and 4 phones demonstrate that MagHarp achieves an average gesture recognition accuracy of 97.6% on a 3,500-sample dataset, effective up to 20 cm (recommended within 15 cm for robustness).