Smart Dual Lock

Inspiration

We saw the 2 step verificaton systems in place like Microsoft where you need to put in the password and get a code from your email. So we decided we could implement something similar for door locks.

What it does

The camera looks at your face and sees if it is recognized. When someone puts in the correct pin and their face is recognized, the lock will open. If they put in the correct pin but their face is not recognized, the lock will not open. Similarly, if their face is recognized but the pin is incorrect, the lock will not open.

How we built it

We used a Raspberry Pi 4 with the Pi Camera 2 module to control the logic of the smart door lock system. To train the system, we took pictures of the person and ran a script to train the facial recognition model. We then paired this with a hardware control script that determines whether the lock will unlock or remain locked based on the facial input. Additionally, we used two motors, which required a relay module due to their power incompatibility with the Raspberry Pi. For the outer casing, we chose Eucalyptus wood, which we cut with a laser cutter using CAD software. We also 3D printed gears and applied gear reduction techniques to increase torque, allowing us to use a less powerful motor.

Challenges we ran into

• laser cutting the wood
• 3d printing the gears
• larning freeCad for wood cutting
• relay module had litte documentation
• number pad matrix was flawed, numbers not logged correctly
• had to switch uor plan from 2 boxes (inside and outside) to one box

Accomplishments that we're proud of

As a group, we've collaboratively used our skills in different areas and were able to come together to work on a common goal. We experimented and stepped into unfamiliar waters and were able to complete a finished project in our first hackathon. We taught ourselves how to use completely unfamiliar hardware, software (CAD, LaserCutting, 3D printing), and program the hardware with little documentation and simply trial and error.

What we learned

We learned a great deal about hardware integration, including the challenges of combining facial recognition with physical control systems. We also gained hands-on experience with the full process of product development, from conceptualization to prototyping and testing. Additionally, we became proficient in tools like FreeCAD, laser cutting, and 3D printing, which were completely new to us. The project reinforced the importance of flexibility and creative problem-solving when working with new technologies.

What's next for Smart Dual Lock

We will implement a TripleLock System, where there's also a physical key slot. It will override the digital software. We could also use FPGAs. By incorporating FPGAs, we could offload specific tasks, such as image processing and motor control, from the Raspberry Pi, improving the overall speed and reliability of the system. Additionally, the parallel processing capabilities of FPGAs would allow for faster decision-making, making the system more responsive in real-time scenarios. We aim to use FPGAs to create a more scalable, high-performance version of our smart dual lock system. Right now it's in one box but a finished product would have a box for outside and inside. The outside box would house the numberpad, key slot, camera, sensors. The inside box would have the motor, pi, and batteries.

Built With

• opencv
• python
• raspberry-pi