Inspiration

The idea for Sator came from seeing how difficult it is for smaller or developing regions to access precise, real-time agricultural insights. Expensive sensor kits and proprietary data systems keep precision farming locked behind corporate barriers. We wanted to create something modular, open, and adaptable — a rover and dashboard system that could make environmental data more accessible to everyone.

What it does

Sator is an AI-integrated rover platform that collects soil and environmental data (moisture, pH, temperature, and NPK) and transmits it securely to a live dashboard. The system features an Eden AI assistant, which summarizes readings, flags anomalies, and helps users interpret changes over time. The dashboard allows teams to switch between sensor types, monitor live feeds, and review historical data trends.

How we built it

We built the front end with React (Vite) and Tailwind CSS, with a design-first workflow prototyped in Figma. The rover control and feed pages communicate through simulated sensor endpoints built in Node.js. We integrated Gemini AI for natural-language explanations and are developing ElevenLabs text-to-speech for field voice output. Version control and deployment are managed through GitHub, with collaboration between hardware and AI teams on the same repo. We built a WebSocket interface so our phone controller could send drive commands and control the drop-down mechanisms. Using our custom interface module, the R4 drove the four DC drivetrain motors and forwarded all up/down mechanism commands over to Dynamixel's OpenRB-150. Powered by an external 5V battery pack, the RB150 then generated the PWM signals for the two MG90S servos that lowered and raised the soil-moisture probe and the NPK+pH sensor, giving us a reliable motion even when the drivetrain motors were under heavy load.

Challenges we ran into

Establishing reliable communication between local sensor data and the web dashboard

Integrating Gemini’s API (authentication and model versioning)

Managing design consistency across multiple pages

Balancing visual polish with functional backend connections

Servos we were using for drop down mechanism only had a 180 deg rotation and was stalling due to running multiple motors on the Arduino R4.

Accomplishments that we’re proud of

Fully functional multi-page dashboard with modular navigation

Real-time rover feed and AI-driven data summaries

A unified visual identity — minimal, scientific, and accessible

Smooth cross-collaboration between front-end, AI, and robotics teams

What we learned

Building Sator taught us the value of modularity — both in hardware and software. By designing flexible systems that connect through standard interfaces, we created a framework adaptable to future missions, sensors, and datasets.