IoT-Based Automated Greenhouse Monitoring & Climate Control by AGROBROS
AppCon2024 Top 20 Finalist
AgroSphere is an IoT-based system designed, implemented, and validated for automated monitoring and internal climate control within a prototype greenhouse environment. It aims to demonstrate the feasibility of applying IoT technology to enhance precision agriculture concepts, improve resource efficiency, and bolster food security strategies, particularly by showcasing a functional and user-friendly system adaptable for resource-constrained environments. The system integrates hardware (ESP32, sensors, actuators), cloud services (Firebase), and a mobile application (Android) for comprehensive greenhouse management.
- Real-time Monitoring: Continuous tracking of Temperature, Humidity, CO₂, Light Intensity, Water Level, Water Temperature, pH, and TDS.
- Remote Control: Manual control of Fan, Mist System, and Ventilation via the Android app.
- Automated Climate Control: ESP32 automatically manages actuators based on configurable thresholds (fetched from Firebase) and sensor readings, with hysteresis to prevent rapid cycling.
- Dual Control Modes: Seamless switching between Manual and Automatic control.
- Preset Growth Profiles: Allows selection of predefined sensor range settings (e.g., Seedling, Flowering) or manual configuration.
- Historical Data Logging & Visualization: Sensor data logged to Firebase with timestamps, visualized in charts on the mobile app.
- Offline Resilience:
- ESP32 continues automatic control using last known settings during network outages.
- Historical sensor data is buffered locally on ESP32's SPIFFS filesystem when offline and uploaded upon reconnection.
- Mobile app utilizes multi-level caching (memory, SharedPreferences) for data persistence.
- Alert Notifications: Mobile app notifications triggered when sensor readings exceed safe thresholds (with cooldown).
- User Authentication: Secure login via Email/Password or Google Sign-In using Firebase Authentication.
- User Preferences: Customizable app theme (Light/Dark/System) and dynamic color options (Android 12+).
- System Information: In-app display of version, authors, and privacy policy/terms of service.
AgroSphere employs a multi-tier architecture:
- Hardware Tier (ESP32):
- Acquires data from various sensors (DHT11/22, MQ135, BH1750, DS18B20, PH4502C, TDS, Float Switch).
- Controls actuators (Fan Relay, Mist Relay, Servo Motor).
- Executes automatic climate control logic.
- Communicates securely with Firebase via WiFi (HTTPS).
- Handles offline data buffering using SPIFFS.
- Cloud Tier (Firebase):
- Realtime Database: Stores current sensor data, actuator states, sensor range settings, and historical data logs. Provides real-time data synchronization.
- Authentication: Manages user authentication and authorization.
- Mobile Application Tier (Android):
- UI (Jetpack Compose): Provides user interface for monitoring, control, and configuration using Material 3 design principles.
- ViewModels (Kotlin): Expose state to the UI and handle user interactions.
- Repositories (Kotlin): Abstract data sources (Firebase, local cache), implement caching logic, and manage data synchronization.
- Models (Kotlin): Data classes representing sensor data, actuator states, etc.
- Local Storage: Room Database (for notifications), SharedPreferences (for preferences and caching).
- Hardware: ESP32-WROOM, DHT22, MQ135, BH1750, DS18B20, PH4502C, DFRobot TDS Sensor, Float Switch, SG90 Servo, Relays, 5V Fan, 5V Atomizer.
- Firmware: C++ (Arduino Framework), FirebaseClient library, ArduinoJson, Sensor-specific libraries.
- Cloud: Google Firebase (Realtime Database, Authentication).
- Mobile App: Android (Kotlin), Jetpack Compose, Material 3, MVVM Architecture, Coroutines, StateFlow, ViewModel, LiveData (potentially via
collectAsState), Room, Retrofit (if used, though Firebase SDK is primary), MPAndroidChart (via Compose wrapper), Google Sign-In. - Communication: WiFi, HTTPS, MQTT (via Firebase library).
- Microcontroller: ESP32-WROOM-32E
- Sensors:
- Temperature & Humidity: DHT22
- Carbon Dioxide: MQ135 (Analog)
- Light Intensity: BH1750 (I2C)
- Water Temperature: DS18B20 (OneWire)
- pH Level: PH4502C (Analog)
- Total Dissolved Solids (TDS): DFRobot Analog TDS Sensor
- Water Level: Float Switch (Digital)
- Actuators:
- Fan: 5V DC Fan controlled via Relay
- Mist System: 5V Atomizer controlled via Relay
- Ventilation: SG90 Servo Motor
- Setup: Initializes WiFi, Firebase connection (with authentication), sensors, and actuators.
- Sensor Reading: Periodically reads data from all connected sensors, applying smoothing or calibration where necessary.
- Actuator Control:
- Implements
controlActuatorAuto()logic based on fetched thresholds and current sensor readings. - Includes hysteresis for temperature and humidity control.
- Includes safety cycle for mist pump (e.g., 2 min ON / 5 min OFF).
- Responds to manual control commands fetched from Firebase.
- Implements
- Firebase Communication:
- Uploads current sensor data (
/sensorData). - Uploads historical sensor data (
/sensorHistory) with timestamps. - Fetches actuator states (
/actuatorStates) and sensor ranges (/sensorRanges). - Handles authentication token refresh (
JWT.loop()).
- Uploads current sensor data (
- Offline Handling:
- Detects network/internet unavailability.
- Saves historical data to SPIFFS (
/offline_history.jsonl) when offline. - Uploads buffered data upon reconnection (
uploadOfflineHistory()). - Uses last known ranges/settings for automatic control when offline.
- Time Management: Uses estimated timestamps based on last successful Firebase sync or uptime approximation.
- Architecture: Follows MVVM (Model-View-ViewModel).
- UI: Built entirely with Jetpack Compose and Material 3 components.
DashboardScreen: Displays current sensor values, actuator controls, preset selection.MonitorScreen: Shows historical data using charts (MPAndroidChart wrapper).NotificationsScreen: Lists historical alerts fetched from Room DB.PreferenceScreen: Allows theme selection, dynamic color toggle, account info, sign out.LoginScreen,SplashScreen.
- ViewModels: (
SensorDataViewModel,ActuatorDataViewModel,SensorRangeDataViewModel) Manage UI state, interact with Repositories, handle user actions using Coroutines and StateFlow. - Repositories: (
SensorDataRepository,ActuatorDataRepository,SensorRangeDataRepository,NotificationRepository) Abstract data sources, manage Firebase listeners, implement multi-level caching (memory/SharedPreferences), interact with Room DB (for notifications). - Utilities: Helper functions for authentication (
AuthHelper), Firebase setup (FirebaseHelper), notifications (NotificationHelper), preferences (PreferencesManager), date formatting, etc.
- Realtime Database: Acts as the central hub for data synchronization.
/sensorData: Holds the latest readings from the ESP32./sensorHistory: Stores timestamped historical records (from ESP32, including offline buffer uploads)./actuatorStates: Reflects the current state of each actuator and the control mode (auto/manual). Updated by both ESP32 (auto mode) and App (manual mode)./sensorRanges: Stores the user-defined or preset temperature/humidity thresholds used by the ESP32 in auto mode.
- Authentication: Secures access to the database and identifies users. Supports Email/Password and Google Sign-In.
- Sensor Network: ~2.07s average update interval, ~1.51s average end-to-end latency (sensor-to-app).
- Actuator Response: <1.2s average for manual control (app-to-actuator), <0.2s average for automatic control (ESP32 internal).
- Reliability: Successfully handled simulated power and network outages, recovering and synchronizing data. Maintained local control during outages.
- Data Optimization: Multi-level caching in the app resulted in an estimated ~8.7% reduction in Firebase data transfer.
- Usability: High task completion rates (>93%) and positive user feedback (>85% satisfaction) in testing.
Copyright (c) 2025 AgroBros
Charles Bryan Fabian, Jerson Sumalinog, Paul John Vicente
This project is provided strictly for demonstration and educational viewing purposes only.
Important Restrictions:
You may not copy, modify, distribute, or use this code for any purpose (commercial or otherwise) without explicit written permission from the authors.