A research-grade platform for Precision Agriculture and Automated Plant Cultivation.
AMiGA (Automated Modular Irrigation & Growth Assistant) is an open-source, NASA-affiliated prototype developed at the Autonomy Research Center for STEAHM (ARCS). It represents the cutting edge of Controlled Environment Agriculture (CEA), bridging the gap between academic research and sustainable, hyper-efficient food production pipelines.
By unifying modular hardware with a robust full-stack software environment, AMiGA allows researchers, agronomists, and developers to closely monitor, simulate, and automate plant growth conditions natively.
Our platform goes beyond standard automated watering setups. It's a cohesive, multi-node architecture designed to handle rigorous data collection and real-world farm integration:
- 🛡️ Autonomous Control Loop: Closed-loop feedback mechanisms that perfectly balance irrigation, diurnal light scheduling, and complex nutrient delivery without human intervention.
- 📊 True Soil Analytics: Continuous 7-in-1 soil chemistry tracking (incorporating NPK levels, pH, Electrical Conductivity, Temperature, and Moisture indexing).
- 🌬️ Atmospheric Sensing: High-precision environmental tracking using I2C protocols (SCD41 for CO2/Temp/Humidity and TSL2561 for precise Luminosity).
- ⚖️ Precision Telemetry: Real-time load-cell tracking to measure evapotranspiration rates, nutrient uptake weight, and final harvest biomass.
- 🧠 Edge Computer Vision: Jetson Orin-powered hardware-accelerated image pipeline handling IR moisture heatmaps and ultra-high-resolution timelapses.
- 💻 Dev-First Simulation: Full hardware simulation layer (
start_simulate.sh) that allows developers to write code, test algorithms, and build the UI on Windows/Mac without needing access to a physical Raspberry Pi or Jetson Orin.
AMiGA is designed with scalability and high modularity in mind.
graph TD
A[React/Vite Dashboard] -- REST API --> B[FastAPI Backend]
B -- Background Tasks --> C[Grow Scheduler & Telemetry]
B -- lgpio Drivers --> D[Raspberry Pi / Native Controller]
D -- I2C / Modbus --> E[Sensors & Actuators]
E --> F[Peristaltic Pumps / SSR Lighting / Soil Array / Load Cell]
G[NVIDIA Jetson Orin] -. Local LAN .-> B
G -- High-Perf Vision --> H[Moisture Heatmaps & Timelapses]
I[Raspberry Pi 4 Scraper] -. OCR Data Sync .-> B
-
Backend (Python / FastAPI)
Residing in/backend, the core logic operates as a high-performance REST API. It handles hardware abstraction via specialized driver classes, manages persistent local state (api/,data/), and orchestrates thegrow_scheduler. It asynchronously processes telemetry to dedicated persistent CSV layers for research reproducibility. -
Frontend (React / Vite)
Residing in/frontend, the dashboard is a responsive web application built with React, Vite, and Tailwind CSS. It provides a single-pane-of-glass view for researchers, allowing instant manual overrides, rule configuration for moisture thresholds, and live data visualization of soil health and system metrics. -
Edge Computing Integrations
To offload high-compute tasks, the AMiGA architecture spans out to edge devices:- Jetson Orin (
/orin/scripts): Captures heavy analytical visual data such as thermal/IR moisture mapping and processes long-term timelapse generation via GStreamer offloading. - Pi4 Telemetry Scraper (
/pi4): Serves as a dedicated node to interactively execute OCR and scrape data from closed-ecosystem apps (e.g., proprietary equipment like Vivosun smart environments).
- Jetson Orin (
The physical manifestation of AMiGA leverages premium agricultural components controlled by robust drivers:
- 💧 Irrigation: Peristaltic pumps driven precisely by TMC2209 stepper drivers to dose micro-amounts of nutrients.
- ☀️ Lighting: High-intensity photosynthetic active radiation (PAR) grow lights, controlled safely via solid-state relays (SSR).
- 🌡️ Atmospherics: SCD41 sensors (CO2/Temp/Hum) and TSL2561 (Luminosity).
- 🌱 Soil Intelligence: Hybrid setup utilizing an Analog Moisture Array alongside a 7-in-1 Modbus NPK/pH/EC soil intelligence sensor.
- ⚖️ Measurement: Dedicated load cells configured for continuous weight sampling to correlate irrigation input with plant mass growth.
AMiGA is engineered for low friction. It provides auto-configured deployment scripts across all major operating systems.
Depending on your host machine, run the setup script to instantly pull requirements, set up virtual environments (.venv), and initialize local data stores:
- Windows: Run
.\install_dependencies.bat - Linux/Pi: Run
./install_dependencies.sh - macOS: Run
./install_dependencies_mac.sh
| Operating Mode | Environment | Command | Description |
|---|---|---|---|
| Simulation | Windows | .\start_simulate.bat |
Boots UI and mock backend hardware layers. |
| Simulation | Linux / macOS | ./start_simulate.sh |
Boots UI and mock backend hardware layers. |
| Native Prod | Raspberry Pi / Hardware | ./start.sh |
Runs actual GPIO drivers & real hardware |
| Scale Service | Raspberry Pi | ./start_scale_hw.sh |
Standalone precision weight management |
AMiGA/
├── backend/ # FastAPI core, hardware classes, async schedulers, and unit tests
├── frontend/ # React + Vite web dashboard application
├── orin/ # NVIDIA Jetson Orin specialized computer vision and thermal scripts
├── pi4/ # Telemetry scraping and OCR integration scripts
├── docs/ # Detailed technical documentation and hardware integration guides
├── data/ # Persistent local storage across simulation and native modes
├── kratky/ # Auxiliary legacy modes (hydroponic routines)
└── legacy/ # Deprecated experimental concepts & tests
For researchers, contributors, and operators looking to understand the inner workings, explore our extended technical documents:
docs/system_overview.md- Comprehensive architectural and software structural breakdown.docs/hardware_integration.md- Specifics on hardware wiring, sensor calibration, and component datasheets.docs/hardware/orin_vision.md- Details pertaining specifically to machine vision implementation.
AMiGA is developed actively under the FOODI initiative — Facilitating Overcoming Obstacles to the Development and Integration of Modern Technologies for Controlled Environment Agriculture.
This core mission is focused on securing humanity's food independence through advanced STEM research, sustainable agricultural architectures, and data-driven cultivation.