A PoE-based watering system using Arduino and a local Linux server for reliable, internet-independent operation, enhanced with supplementary ChatGPT powered AI-assisted watering schedule. Completed in 36 hours, our 3-member team won the Best Hardware Award for creating a safe, convenient, plug-and-play solution.
The Arduino module features an capacitive moisture sensor, a magnetic water level switch, and a NPN transistor based pump-driver circuit. The sensors ensures the safe operation of the pump while the plants is watered with the precisely set moisture value. Between the raspberry-pi Linux server and the modules, the system features an managed PoE network switch to reduce power cable clutter while the access point provides a network connection to the android app. Providing secondary support to users, the app utilizes ChatGPT along with Pl@ntnet API for plant care recommendations. As a challenge to increase system's practicality and security, our team are currently taking the next steps to implement an more integrated voice-recognition AI assistant feature while the VLAN implementation prevents direct access to the modules.