Inspiration
everyone in our group has had plants die due to neglect. This made us think of a way we could remove the chore of keeping you plants alive by developing a system that will automatically water your plants as well as give you live updates on your plants health.
What it does
Water Me is fundamentally an Arduino that can send sensor data to a website. This website then decides when to water your plants and what their health is using soil moisture level, temperature, and humidity. We also challenged ourselves to add plant detection where you can upload an image of your plants and it will tell you what it is.
How we built it
We 3d printed the casing for our automatic water dispenser. We then placed an Arduino inside the casing and attached the sensor to the outside of the box a Battery back is connected to the pump and another batter powers the ESP32. A phone hotspot acts as a router that connects the users device and the ESP32(s).
Challenges we ran into
The relay control for the water pump was not working correctly so we substituted the water pump for an led light. Connecting the users device (my laptop in this case) to the ESP32 took some effort. We had problems making sure that the device could be accessed and send back sensor data while the pump is running.
Accomplishments that we're proud of
The website UI looks much better than we thought we could achieve. The LLM that we trained was able to detect some of the objects correctly.
What we learned
We learned how to wire an ESP32 board correctly and get the board to communicate with a database/server though a WIFI connection. We learned how to make a basic CAD model. We learned the basics of machine learning and how to train an LLM to do image detection to some degree. We learned how to incorporate Gemini and Claude into our workflow to more effectively get ideas working.
What's next for Water Me
We wish to continue working on the LLM and continue to iterate and improve upon the CAD design for the watering device.
Log in or sign up for Devpost to join the conversation.