1) Introduction



    • Statement of Purpose

The goal of our project is to at the very least build a primitive machine that can distinguish a singular note produced by a variety of individuals. The potential end goal is for the machine to recognize a person’s voice for security reasons. Examples include an ATM or a safe. I think this takes a unique approach to security that has not been explored in depth as compared to things like face recognition or fingerprint. It also has the potential to be far more convenient than fingerprint scanners as it is a hands-free solution to unlocking something.

    • Background Research

      Most of the research we have done has been related to how we can take a sound and break it down in order to identify the source of the sound. We have decided the first step towards this we will take is to break down a human produced note and figure out how to decipher between two possible sources which it is. This would be accomplished by taking the sound’s fourier transform and looking for any outstanding frequencies that we could use to identify which of the sources it is. This would then be linked to a switch that would turn on an LED based on if a certain source has been identified. We can then build this to be a general voice based security system.

  1. Design Details


  2. Parts
    -Arduino

-Stepper Motor

-Battery

-External Microphone?


  1. Possible Challenges

    -learning how to effectively use Matlab/Python to perform Fourier Transforms

-identifying unique features of a person’s voice 

-comparing and matching two different sounds

  1. References

    [1]"Fourier transform and speech recognition", Sail.usc.edu, 2019. [Online]. Available: https://sail.usc.edu/~lgoldste/Ling285/Slides/Lect20_handout.pdf. [Accessed: 28- Sep- 2019].

[2]"Fourier Series- MATLAB & Simulink", Mathworks.com, 2019. [Online]. Available: https://www.mathworks.com/help/curvefit/fourier.html. [Accessed: 28- Sep- 2019].

[3]2019. [Online]. Available: https://www.academia.edu/39070382/Voice_recognition_via_Fourier_Transform. [Accessed: 28- Sep- 2019].

[4]"Fourier Transforms- MATLAB & Simulink", Mathworks.com, 2019. [Online]. Available: https://www.mathworks.com/help/matlab/math/fourier-transforms.html. [Accessed: 28- Sep- 2019].

Attachments:

Comments:

To be clear, you're final device will light up an LED based on who it thinks if currently talking? If so, I think that's a fun project. I would encourage you guys to work on the microphone and sampling aspect of the project first, as you can prototype it on a breadboard fairly easily and it'll most likely be the most challenging portion.

Posted by chorn4 at Oct 03, 2019 12:44