Smart Domotic Plant Pot
IoT system for automatic and intelligent management of potted plants, featuring sensors, automatic irrigation, web dashboard, local Machine Learning models and a conversational AI assistant. Commercially distributable on Raspberry Pi 5.
Smart domotic plant pot designed to be installed on Raspberry Pi 5 with real sensors (soil moisture, temperature, light), irrigation pump and lighting LED. The application is currently in testing phase in simulation mode (realistic virtual pot) before deployment on the final hardware. KEY FEATURES: • Setup wizard with 36 preset plant profiles (aromatic, indoor, edible, flower, succulent) and automatic threshold calculation based on species + position + season. • User dashboard identical in simulation or real hardware mode: real-time sensor visualization, historical charts, manual pump/LED control, automatic/manual mode. • Hardware Twin: animated SVG virtual pot with reactive plant (6 health stages), preset scenarios (drought, rain, 1000× vacation), sensor fault injection, time-acceleration to test weeks in minutes. • Plant Brain (local ML, scikit-learn BSD-3): moisture forecast at +1h/+3h/+6h/+12h, sensor anomaly detection (Isolation Forest), root-cause diagnostics, threshold auto-tuning on historical data, predictive anticipated irrigation. • Plant Coach AI (local LLM via Ollama, MIT): conversational chat in Italian for gardening assistance, weekly insight generation, zero cloud, zero API keys, zero recurring costs. • MQTT integration with Home Assistant (auto-discovery). • Email notifications on critical thresholds. • Complete logging system, alerts, bcrypt authentication + rate limiting. • Git auto-update, systemd service with auto-restart. • 100% open source software stack with permissive licenses (BSD/MIT/Apache), commercially distributable without source release obligation. Since May 13, 2026 the system was online in simulation mode (public demo no longer active), with shareable PDF user manual. Migration to Raspberry Pi 5 with real hardware is planned as a subsequent phase.
Tecnologie: Python 3, Flask, SQLite, scikit-learn, Ollama, Phi-3.5 Mini, WeasyPrint, paho-mqtt, bcrypt, Raspberry Pi 5, gpiozero, Home Assistant, systemd, Nginx