ECE 110/120 Honors Lab Section : Quadrocopter with facial recognition

Created by Artamonov, Maxim on Sep 25, 2019

  1. Introduction

    1. Statement of purpose

Create a quadrocopter that would perform face recognition and lock onto the desired person, following him/her around. Facial recognition finds more and more applications every day. For example one of the application could be during a search after a natural disaster when the rescue teams are looking for the victims.

    1. Background research and Personal Engagement:

We want to work on this project because it incorporates both electrical and computer engineering in a very neat way, and since our group includes people from both majors we decided that this would make a great and interesting project.

  1. Design details

    1. Block diagram

    1. System overview

The camera gathers visual information and sends a signal to Raspberry Pi which processes the input and sends out a signal to Arduino which directly controls the motors via ESC since an Arduino is not rated for enough power to make motors work directly.

  1. Parts

4x brushless motors

4x ESC

8x motor propellers (in case some break)

1x camera module

1x Raspberry Pi (maybe something else as the main processor)

Wires

1x Arduino

1x Sabertooth

  1. Possible challenges

Building a functional quadrocopter does not seem like a big challenge and there are many tutorials online. The part that could be hard is the programming associated with facial recognition and turning the input from a camera into the signal for the motor ESCs.

  1. References

There are a bunch of resources online on the matter, here are some of them:

https://www.instructables.com/id/Ultimate-Intelligent-Fully-Automatic-Drone-Robot-w/

https://www.youtube.com/watch?v=Hu3p5ZR_i5s

https://www.oreilly.com/ideas/how-to-build-an-autonomous-voice-controlled-face-recognizing-drone-for-200

Comments:

This project is outside the scope of what you can likely accomplish in one semester. Additionally, we typically do not allow flying projects.

Rather than a drone, could you try implementing a similar method on a car on the ground? Having a car follow up, perhaps for the purpose of carrying heavy items or luggage, would use similar skills but avoid the complexity of flying
Posted by chorn4 at Oct 03, 2019 12:46

We would discourage a drone project, aside from the very particular edge case in which the entire team has extensive pre-existing experience with drone design and construction. Unless the entire team is highly experienced in drone design, please try and find an alternative non-flying project.
Posted by fns2 at Oct 03, 2019 18:22

Approved, please be responsible with your flying project. Know that this is the first time we've ever let someone do a flying project, and this is only because your team has hands-on direct experience with building and flying quadcopters. Please be responsible and safe.
Posted by fns2 at Oct 03, 2019 18:44