Smart Safe with Biometric & Voice Control

This smart safe project reimagines traditional lock-and-key security with modern biometric and voice-controlled authentication. The system provides multiple unlock methods while maintaining high security through fingerprint verification and loudness-based intrusion detection.

The integration of custom electronics, 3D-printed enclosure, and sophisticated software creates a fully functional smart home security device. The project showcases end-to-end product development from circuit design to mechanical enclosure fabrication.

The safe enclosure was designed in CAD software and 3D-printed using PLA material. The design incorporates mounting points for the solenoid lock, fingerprint sensor, microphone module, LCD display, and Arduino controller. Wall thickness was optimized for strength while minimizing print time and material usage.

The locking mechanism uses a 12V push-pull solenoid controlling a sliding bolt that engages with the door frame. The door hinge design allows smooth opening with minimal friction while maintaining alignment for reliable locking. Cable management channels route all wiring internally for a clean appearance.

The enclosure features a front panel with cutouts for the fingerprint sensor and LCD display, positioned for ergonomic access. Mounting brackets allow wall installation or freestanding placement. The design permits easy access to internal electronics for maintenance and upgrades.

The control system centers on an Arduino Uno microcontroller interfacing with multiple sensors and actuators. The fingerprint sensor (R307 or similar) communicates via UART serial protocol, storing up to 127 fingerprint templates in onboard flash memory.

The voice recognition module uses onboard speech recognition chips for command detection, operating independently before triggering the Arduino. The loudness detector circuit uses an electret microphone with op-amp amplification, providing analog output to the Arduino's ADC for threshold detection.

The solenoid lock requires 12V at 1A, powered through a MOSFET driver controlled by a digital output pin. A flyback diode protects against inductive kickback. The LCD display (16x2 character) uses I2C communication to minimize pin usage. A 12V power supply feeds a buck converter providing regulated 5V for the Arduino and sensors.

The circuit includes visual feedback LEDs (red/green) indicating lock status and authentication results. A buzzer provides audio alerts for successful/failed unlock attempts and intrusion detection.

The Arduino firmware implements a state machine managing the safe's operational modes: locked, unlocking, unlocked, and alarm. The main loop polls sensors and updates the display while handling user inputs.

Fingerprint authentication uses the Adafruit Fingerprint library for sensor communication. The enrollment mode stores new fingerprints, while verification mode compares scans against stored templates. Match confidence thresholds can be adjusted for security/convenience tradeoff.

Voice command processing listens for predefined wake words like "Open Safe" followed by a passphrase. The software implements debouncing and confirmation sequences to prevent accidental unlocks. Multiple failed authentication attempts trigger a lockout period, logging events to EEPROM.

The loudness detector continuously samples the microphone input, comparing against a dynamic threshold. Sustained loud noises trigger the alarm mode, activating the buzzer and sending notifications. The system uses rolling average filters to reduce false alarms from ambient noise.

Additional features include auto-lock timers, low battery warnings, and a manual emergency override accessible through a hidden button sequence on the LCD.

The completed safe successfully authenticates users via fingerprint with 95%+ accuracy under normal conditions. Voice recognition achieves reliable operation in quiet environments, with occasional false rejections in noisy settings. The solenoid lock engages/disengages reliably over 500+ test cycles.

Loudness detection effectively identifies intrusion attempts (hitting, prying) while filtering normal ambient noise. The 3D-printed enclosure proved durable through drop tests from 1 meter height with no structural failures. Battery life (using 12V 2Ah lead-acid battery) sustains 2+ weeks of standby operation.

The project demonstrates practical IoT security applications and received positive feedback as a portfolio piece. Future enhancements could include WiFi connectivity for remote monitoring, camera integration for visual verification, and mobile app control for access logging and management.

**Secondary video** showcasing the 3D design and assembly: https://youtu.be/1QEjZAZJ12g