🏠 Dynamic Personalized Smart Home (DPSH)

A distributed, IoT-based smart home system designed to simulate adaptive environments using MQTT, Node-RED, and MongoDB Atlas — deployed across AWS EC2 instances for scalability and fault tolerance.

---

🌐 Overview

DPSH (Dynamic Personalized Smart Home) integrates multiple simulated IoT sensors, cloud-based data processing, and real-time dashboards to create a personalized, scalable smart home environment.

Each simulated room (e.g., hallway, living room, bedrooms) streams sensor data such as temperature, light, motion, and facial recognition to an EMQX MQTT broker hosted on AWS.
A controller service processes face detections, performs mood inference, and retrieves personalized environment settings from MongoDB Atlas.
The Node-RED interface orchestrates data flow, applies decision logic, and visualizes real-time metrics through an interactive dashboard.

---

🧩 System Architecture

The system is composed of the following main components:

Component	Description
Sensor Simulators	Node.js scripts publishing temperature, light, motion, and face data for each room.
MQTT Broker (EMQX)	Dockerized MQTT broker hosted on AWS EC2, load-balanced via AWS NLB.
Controller EC2	Runs faceMLSim.js and MoodAggregator to process user moods and environment logic.
MongoDB Atlas	Stores user identity, mood preferences, and environmental settings.
Node-RED EC2	Subscribes to MQTT topics, applies logic flows, and drives dashboard visualization.
PM2 Process Manager	Ensures fault-tolerant Node.js process management and automatic restarts.

---

⚙️ Features

• 🌡️ Real-Time Sensor Simulation – Multi-room environmental data streams.
• 💡 Adaptive Environment Control – Lights and temperature adjust automatically based on mood.
• 🧠 Face-Based Personalization – Facial recognition simulates mood-based preferences.
• ☁️ AWS Cloud Deployment – EMQX, Node-RED, and controllers deployed on EC2 behind a Network Load Balancer.
• 📊 Node-RED Dashboard – Interactive visualization for rooms, users, and system state.
• 🧩 Scalable Modular Design – Add rooms or devices by editing configuration JSON files.

---

🏗️ Project Structure

ProjectSmartHome/
│
├── src/
│   ├── sensors/
│   │   ├── sims/
│   │   │   ├── temp.sim.js
│   │   │   ├── motion.sim.js
│   │   │   ├── light.sim.js
│   │   │   ├── facecam.sim.js
│   │   ├── bed1.json
│   │   ├── bed2.json
│   │   ├── hallway.json
│   │   └── living.json
│   │
│   ├── controllers/
│   │   ├── faceMLSim.js
│   │   ├── moodAggregator.js
│   │   └── utils/
│   │
│   └── .env
│
├── docker-compose.yml
├── package.json
├── README.md
└── docs/
├── Report.pdf
└── images/

---

🚀 Deployment Steps

1. Clone the repository

git clone https://github.com/<your-username>/DynamicPersonalizedSmartHome.git
cd DynamicPersonalizedSmartHome

2. Install dependencies

npm install

3. Configure environment variables

Create a .env file with:

BROKER_URL=mqtt://<NLB-DNS>:1883
MONGO_URL=mongodb+srv://<user>:<pass>@cluster.mongodb.net/smarthome
DB_NAME=smarthome
FACE_COLL=faces
MOOD_COLL=moods

4. Run sensor simulators

pm2 start src/sensors/simRunner.js --name sensors

5. Start controller services

pm2 start src/controllers/faceMLSim.js --name faceML
pm2 start src/controllers/moodAggregator.js --name moodAgg

6. Launch Node-RED

node-red

Access the dashboard at:

http://<NodeRED-EC2-PublicDNS>:1880/ui

---

📡 MQTT Topics

Topic	Description
homeA/<room>/temp	Temperature readings
homeA/<room>/motion	Motion detection
homeA/<room>/light	Light status
homeA/<room>/faceData	Simulated face detection
homeA/<room>/mood/out	Aggregated mood output
homeA/<room>/cmd	Control commands to actuators

---

🧰 Technologies Used

Layer	Tools
Simulation & Logic	Node.js, MQTT.js
Messaging & Communication	EMQX MQTT Broker
Database	MongoDB Atlas
Cloud Infrastructure	AWS EC2, AWS NLB
Visualization	Node-RED
Process Management	PM2

---

🧪 Testing & Validation

Test Case	Description	Result
Motion + Low Light	Light turns on automatically	✅ Pass
Idle > 30s	Auto light off	✅ Pass
Face Detection	Triggers mood lookup	✅ Pass
Multi-user Mood	Aggregation works	✅ Pass
MQTT Latency	Under 2 seconds	✅ Pass

---

📈 Scalability and Future Work

• Integrate real facial recognition models (e.g., OpenCV + TensorFlow Lite).
• Expand to real IoT devices using ESP32 / Raspberry Pi nodes.
• Add a central web dashboard for remote monitoring.
• Migrate to containerized microservices using AWS ECS or Kubernetes.

---

👤 Author

Sunain Mushtaq 📚 Computer Science Student – Deakin University 📧 [your.email@example.com] 🔗 LinkedIn

---

📝 License

This project is licensed under the MIT License – feel free to modify and build upon it for educational or research purposes.