Machine learning integrated 3D printed smart numpad for user recognition using piezo-tribo mechano-electric transducer

Real-time smart user identification systems provide enhanced security by capturing user-specific keystroke dynamics rather than relying solely on conventional passwords or card-based authentication. In recent times, mechano-electric transducer-based smart systems have been developed for different sensing, recognition, and authentication. In this work, a mechano-electric-transducer-based smart numpad was developed using triboelectric and piezoelectric nanogenerators to acquire detailed data for user authentication. The piezoelectric component was prepared using BCST (0.3Ba 0.7 Ca 0.3 TiO 3 –0.7BaSn 0.12 Ti 0.88 O 3 ) ceramics optimized through controlled ball milling and thermal-assisted sintering, resulting in multiphase coexistence, enhanced mechano-electric and functional properties. The PDMS/BCST composite film was prepared to investigate its triboelectric and piezoelectric-based mechanical-electrical response. The optimized mechano-electric transducer is integrated into a 3D-printed smart numpad capable of generating multi-signal responses. For accuracy, smart user identification systems deployed a machine-learning model trained with 1200 datasets to achieve an accuracy of 90.67%.