In densely populated cities, trash often accumulates on the sides of roads. We wanted to make a robot that would be able to clean trash off of the streets. This would have an especially useful applications in public places, like parks. In Boston, there is a high number of sharps (needles, etc.) reported in gardens and common pathways, sometimes leading to child or pet injury.
What it does
trash-E is an autonomous robot that can receive commands through wifi, be driven manually, and has spaces for stepper motors to attach and actuate a gripping device
trash-E has a pair of eyes (a camera) that uses a machine-vision algorithm to identify people, so that it can drive towards them.
How we built it
We used an 80-20 base with attached DC motors to make trash-E move. It is controled using an Arduino and Adafruit Huzza breakout board.
Challenges we ran into
The biggest challenge we faced was with out stepper motors. The motors we had did not produce enough torque to move our robot forward, so we had to scrap our current plan and try to used DC motors to make our robot move.
We created scripts to automate a training classification algorithm, but the data was not clean enough for it to work out
We had many challenges with integration, after our original plans did not plan out. We did learn a lot about rapid prototyping, and adapted by re-scoping our project!
Accomplishments that we're proud of
Amy learned how to use stepper motors for the first time, and made them wireless by integrating them with a wifi board (also a first)
Ever learned that laser cutting isn't good for prototyping a first pass of a complicated system. Cardboard is.
Kyle learned that too, and that traditional rapid prototyping is not that rapid (especially at a hackathon.
Emma learned how to train and implement openCV algorithms, and created a training set of her own, before pivoting and learning to use someone else's library.