Team
- E/21/087, Shihara Dewagedara, e21087@eng.pdn.ac.lk
- E/21/138, Fikry M.N.M., e21138@eng.pdn.ac.lk
- E/21/302, Sahandi Perera, e21302@eng.pdn.ac.lk
- E/21/452, Zaid M.R.M., e21452@eng.pdn.ac.lk
Table of Contents
- Introduction
- Overall System Architecture
- Data Flow Architecture
- Hardware Design
- Software Architecture & Stack
- Project Timeline
- Testing & Validation
- Detailed Budget
- Conclusion & Future Work
- Links
Introduction
Landslides remain a significant threat in hilly regions of Sri Lanka, often triggered by intense monsoon rainfall. SlideSense is an integrated IoT solution designed for real-time monitoring of high-risk slopes.
The system moves beyond simple data logging by incorporating:
- Redundant Communication: Utilizing both Cloud (AWS/Firebase) and Local (Gateway) paths.
- Edge Intelligence: Real-time threshold analysis on the ESP32 to trigger immediate local sirens.
- Resilience: Designed to function even when external cellular networks fail during heavy storms.
Overall System Architecture
SlideSense follows a multi-layer IoT architecture:
- Perception Layer: Distributed sensor nodes (ESP32) collect soil moisture, volumetric water content, and tilt data.
- Network Layer: Data is transmitted via MQTT protocol to a central Raspberry Pi gateway and subsequently to the AWS IoT Core.
- Application Layer: A React-based web dashboard and Firebase Cloud Messaging (FCM) provide localized alerts and historical data visualization for authorities.
Data Flow Architecture
The system ensures reliable and redundant data transmission:
- Sensors collect environmental data
- ESP32 performs edge analysis
- Data transmitted via LoRa / SIM900A
- Cloud (AWS/Firebase) processes & stores data
- Alerts triggered via Dashboard / SMS / FCM
Core Components
- ESP32 WROOM 32U
- Capacitive Soil Moisture Sensors (x4)
- Tipping Bucket Rain Gauge
- High Sensitivity Microphone Sensor
- LoRa RA-02 SX1278 Module
- SIM900A GSM Module
- 20W Solar Panel
- MPPT Charge Controller
- 3.7V Li-Po Battery
Power System
- MPPT Solar Charging
- 3.3V Regulation
- Deep Sleep Power Optimization
Software Architecture & Stack
Firmware
- Arduino Framework
- FreeRTOS Task Management
- MQTT Communication
- Deep Sleep Mode
Backend & Cloud
- AWS IoT Core / Firebase
- Mosquitto MQTT Broker
- Node.js Gateway (Optional)
- Firebase Cloud Messaging (Alerts)
Frontend
- React.js Dashboard
- Real-time Data Visualization
- Alert Monitoring Panel
🛠 Software Stack Diagram
Project Timeline
The project was executed in four structured milestones:
- Milestone 01 – Proposal & Planning
- Milestone 02 – Hardware Setup & Testing
- Milestone 03 – Working Prototype
- Milestone 04 – Final Product & Documentation
Testing & Validation
Hardware Testing
- Soil moisture calibration (dry vs saturated soil)
- Rain gauge pulse verification
- Battery discharge & deep sleep measurement
Connectivity Testing
- MQTT latency: < 2 seconds
- GSM fallback verification
- LoRa range testing in open field
Failover Testing
- Local gateway broadcast when WAN disconnected
- Alert triggering under simulated rainfall conditions
Detailed Budget
| Item | Quantity | Unit Cost (LKR) | Total (LKR) |
|---|---|---|---|
| 20W Solar Panel | 1 | 3,350 | 3,350 |
| Li-Po Battery | 1 | 1,185 | 1,185 |
| MPPT Controller | 1 | 1,450 | 1,450 |
| Voltage Regulator | 1 | 120 | 120 |
| ESP32 | 1 | 1,860 | 1,860 |
| Soil Moisture Sensor | 4 | 290 | 1,160 |
| Tipping Bucket | 1 | 4,000 | 4,000 |
| LoRa Module | 1 | 1,500 | 1,500 |
| Microphone Sensor | 1 | 200 | 200 |
| SIM900A | 1 | 1,450 | 1,450 |
| Total Cost | 16,275 LKR |
Conclusion & Future Work
SlideSense demonstrates a cost-effective, scalable, and resilient landslide monitoring system integrating edge intelligence with cloud-based alerts.
Future Enhancements
- LoRaWAN integration for extended mountainous coverage
- Machine Learning (LSTM-based rainfall prediction)
- IP67-rated rugged enclosure
- Mobile application for public warning