Engineering Concepts

Skittle Sorter Final

For this project, we were prompted to create a skittle sorter that would be able to sort through a full bag of skittle. After doing research, we decided to make a bigger model that was meant to have two sensors and two bottom servos, so the machine would run twice as fast. Due to time constraints we only able to use one of each, but the color sensor still works. The process of the skittle sorting is the skittles dropping through the initial funnel, which are then singularized and ran under the scanner. Once the skittle is scanned the bottom servo is told to move to a certain spot where the color is located. The skittle then drops through and into it's bin.

Skittle Sorter Singularizer Prototype

To start we used Onshape, to create the dimensions of each piece. Afterwards, we prototyped our funnel, and spinning function. The pieces include a funnel, top layer, skittle catching layer, bottom layerOnce all the pieces were assembled we attached them to the 360 servo using a combination of hot glue and string and used a dowl to attach the pieces.

Final Face Recognition Model

The goal of this model was to recognize me and my classmate's faces through the facial recognition model that we created and to have abve an 80% accuracy on the predictions. We collected the data used in this model through taking multiple pictures of the students everytime we came into class, so the model would have more information to base predictions off of. The model was initially trained using training images and labels to learn the data and then uses a different set of images to test itself and make sure that it can recognize different people's faces with high accuracy. The accuracy comes from repeatedly testing the model through epochs and as the machine goes through more epochs which is a system that goes through the data and it adapts and increases its accuracy. I got my training accuracy to 89% and my testing too 85%, training has a higher accuracy because it has more data to go off of. Face Recognition Model

Prototype Phase 1B (RGB sensor)

We uploaded the code that we got from Adfruit's Github and uploaded the necessary libraries. Then, we tested the RGB colors with different colored pens and pencils. (and my hand as shown in the video)

IMG_1097.MOV

Get Servos Moving - GitHub Link

Wiring Table

Wiring Labels

IMG_2921.MOV

Video of it moving

IMG-0813.MOV

Robot is Moving!!

This is a video of my individually functioning robot motors. the things included in my robot are two motors, two battery packs, and pico board, and a breakout board.

Coding Table

This is an organizational table to help code.

Github

We uploaded our code into github. To do this, we saved our code on the code.py file. afterwards we uploaded everything into a new repository named "two_motors_working".

Wiring table

IMG-1345.MOV

Robot Traversing a path

For this project, I started off by making adjustments to my robot to decrease slipping. I used batteries to weigh down each side and taped the wheels to increase the friction . Once I figured out that problem, I used previus code to get my code to go forward, turn left, forward, turn left, forward, turn left, forward, turn right, forward, and finally turn right again.