Inspiration

We were inspired by our own experiences of untimely delivery. There were many moments in our everyday life when we wanted to deliver something important to our classmates or needed something we forgot in our dorms. With a lack of time, we realized having a campus delivery robot would be very helpful in students' everyday lives.

What it does

Our car delivers small items. It can detect and stay within lines, identify and categorize objects, and stop before obstacles to prevent crashes.

How we built it

We used OpenCV, TensorFlow, CAD, Python, ultrasonic sensors, a camera, and a Raspberry Pi to build the vehicle.

Challenges we ran into

Uneven Motors: Made steering difficult as usually left and right turns of the same degree would simply require flipped motor speeds. However, due to hardware imbalances, the proper motor speeds were much harder to determine.

Camera Height: Camera mount height on the car was too low, making it hard to see the lines on both sides at once. This decreased the accuracy of our line capturing and steering.

OpenCV: It was quite hard to actually get running on the Raspberry Pi at first as many of the dependencies were not updated or available anymore. In addition, getting the camera to work on the Pi with OpenCV was a challenge.

Accomplishments that we're proud of

Our group was not very familiar with OpenCV and TensorFlow. It was a new experience for all of us because we never really participated in a large hackathon. Our accomplishments include programming a successful product which can detect and stay within lanes as well as identify objects.

What we learned

We gained more experience with OpenCV, TensorFlow, Python, and designing an innovative solution to a problem. We also learned more about teamwork and having effective communication both verbally and through GitHub.

What's next for CampusBot

CampusBot will become more refined and tested on a bigger scale. We will work on making its turns smoother with the lanes, have applications of the object identification, and allow it to travel across campus.

Built With

Share this project:

Updates