TANIKU – Smart Plant Stress Monitoring System Based on Soil Condition
Where AI meets Agriculture—right from your home garden.
🪴 “A healthier plant starts with knowing when it’s stressed.”
🌐 Domain of Impact
TANIKU is an innovative system designed to transform how people care for plants at home by combining IoT, TinyML, and fault-tolerant embedded systems. It helps users detect plant stress in real time—before visible symptoms appear—by monitoring key environmental parameters such as soil moisture, temperature, and humidity.
The system empowers urban gardeners with:
- 🔌 Embedded AI: On-device machine learning for fast and offline stress classification
- 📱 Mobile Dashboard: Real-time data and history tracking via Blynk app
- 📊 Edge Analytics: Lightweight CNN model running on ESP32 to analyze sensor data instantly
- 🌱 Smart Farming: Making precision agriculture accessible—even for home pots and balconies
By enabling early detection and remote monitoring, TANIKU supports sustainable plant care and helps prevent damage due to late intervention.
👩💻 Meet the Team Behind TANIKU
A multidisciplinary student team from Universitas Brawijaya, blending passion and skills in embedded AI, electronics, and cloud systems:
❗ The Real-World Problem
🌿 Home plants often suffer silently—stress symptoms are subtle and go unnoticed until it’s too late.
🕒 Most plant owners don’t know when their plant needs help.
📉 There’s no easy-to-use, real-time system for early stress detection.
⚠️ Result? Overwatering, underwatering, and wasted resources.
🎯 Our Mission
Make smart plant care accessible for everyone.
✅ Real-time monitoring of soil moisture, temperature, and humidity
🧠 TinyML model on ESP32 to classify plant stress: Healthy, Moderate Stress, and High Stress
📲 Easy access via OLED display and Blynk mobile dashboard
🔁 Dual-ESP32 system to ensure fault-tolerant monitoring
💡 How TANIKU Works?
- 📥 Sensor Input – Soil moisture (Capacitive v1.2), DHT11 for air temperature and humidity
- ⚙️ TinyML Inference – CNN model trained on real-world plant data
- 📊 Local Display – OLED shows stress status & sensor values
- 📡 Cloud Integration – Blynk dashboard for mobile access
- 🧯 Fault Recovery – Backup ESP32 takes over if primary fails
🧰 Component Stack
🧩 System Architecture
- Input Layer: Soil Moisture & DHT Sensors
- Processing: CNN-based TinyML model on ESP32
- Output: OLED Display & Blynk Cloud
- Redundancy: Backup ESP32 activated via sensor fault detection
🖼️ Redundancy Workflow: Hot Standby Mechanism
To ensure continuous operation, the system employs a hot standby configuration with two ESP32 microcontrollers. The backup unit mirrors sensor inputs and seamlessly takes over when the main unit fails.
☁️ Cloud Monitoring
- 📲 Users can view data in real-time via Blynk
- 🕓 Historical log access allows users to monitor trends
- 📉 Helps users analyze and respond to stress patterns better
- 🔄 Synced with embedded TinyML inference on-device
🎬 See It In Action
📽️ Watch Demo Video Here –>https://bit.ly/Taniku_7
What you’ll see:
- Real-time monitoring of live plants
- Stress classification shown on OLED + mobile app
- ESP32 switching from main to backup during simulated sensor failure
✅ Results & Impact
✔️ Model Accuracy: 84.69% (train), 85.12% (validation)
✔️ Tested on 15 plants across 3 locations
✔️ System runs offline thanks to on-device AI
✔️ Fault-tolerant architecture proven effective in real-world tests
💬 Project repository!
🌐 GitHub: Smart Plant Stress Monitoring System Based on Soil Condition
🔗 LinkedIn:
Muhammad Zaki Nur Ali
Hasna Najla Latifa
Afra Naima Meryandha
Arfian Faiq Hanafi
Septhian Dian Siswoyo
📸 TANIKU in the Field
Because great ideas bloom with teamwork
