E-Nose Project: Detecting and Classifying Smells Using MQ Sensors and Machine Learning

---

Table of Contents

1. Introduction
2. Project Overview
3. Hardware Requirements
4. Pinout Map
5. Software Requirements
6. Setup Instructions
7. How It Works
8. Python (Qt) Interface Features
9. Arduino Code
10. Training the Model
11. Real-Time Inference
12. Adding New Smells
13. Troubleshooting
14. Future Enhancements
15. License
16. Folder Structure

---

Introduction

The E-Nose Project is a system designed to detect and classify smells using an array of MQ gas sensors and a machine learning model. The system consists of:

• A microcontroller (MCU) to read sensor data.
• A Python (Qt) interface to record data, train a model, and perform real-time inference.

This project is ideal for applications such as:

• Environmental monitoring.
• Industrial safety.
• Food quality control.
• Home safety (e.g., gas leak detection).

---

Project Overview

The E-Nose system works as follows:

1. Sensor Data Acquisition:
   • The MCU reads data from the MQ sensors and sends it to the computer via serial communication.
2. Data Recording and Filtering:
   • The Python (Qt) interface records 150 data points per smell and filters them down to 25 points to minimize errors.
3. Model Training:
   • The recorded data is used to train an Artificial Neural Network (ANN) model.
4. Real-Time Inference:
   • The trained model is used to classify smells in real-time based on sensor data.

---

Hardware Requirements

1. Microcontroller (MCU):
   • Any MCU with at least 1 analog pin (e.g., Arduino Nano, Arduino Mega, ESP32, STM32, etc.).
   • If using more than 1 cluster, ensure the MCU has enough digital pins for multiplexer control.
2. Sensors:
   • MQ-2: LPG, propane, methane, hydrogen, smoke.
   • MQ-3: Alcohol, ethanol, benzene.
   • MQ-4: Methane, natural gas.
   • MQ-6: LPG, butane, propane.
   • MQ-135: Ammonia, benzene, alcohol, smoke, CO₂.
   • MQ-136: Hydrogen sulfide (H₂S).
   • MQ-138: Benzene, toluene, alcohol, acetone, propane, formaldehyde.
   • MQ-9: Carbon monoxide (CO), flammable gases.
3. Multiplexers (Optional):
   • Use CD74HC4067 (16-channel analog multiplexer) if the MCU has limited analog pins.
4. Connections:
   • Connect each MQ sensor to the multiplexer (if used) or directly to the MCU’s analog pins.
   • Power the sensors using the 5V and GND pins of the MCU.

---

Pinout Map

For Arduino Mega

Multiplexer	Control Pins (S0-S3)	Analog Pin	Sensors Connected
MUX 1	D2, D3, D4, D5	A0	1 × MQ-2, 1 × MQ-3, 1 × MQ-4, 1 × MQ-6, 1 × MQ-135, 1 × MQ-136, 1 × MQ-138, 1 × MQ-9
MUX 2	D6, D7, D8, D9	A1	1 × MQ-2, 1 × MQ-3, 1 × MQ-4, 1 × MQ-6, 1 × MQ-135, 1 × MQ-136, 1 × MQ-138, 1 × MQ-9
MUX 3	D10, D11, D12, D13	A2	1 × MQ-2, 1 × MQ-3, 1 × MQ-4, 1 × MQ-6, 1 × MQ-135, 1 × MQ-136, 1 × MQ-138, 1 × MQ-9
MUX 4	D14, D15, D16, D17	A3	1 × MQ-2, 1 × MQ-3, 1 × MQ-4, 1 × MQ-6, 1 × MQ-135, 1 × MQ-136, 1 × MQ-138, 1 × MQ-9

For Other MCUs

• Use the same structure, but adjust the pin numbers based on your MCU’s pinout.

---

Software Requirements

1. Arduino IDE:
   • To upload the code to the MCU.
2. Python 3.x:
   • To run the Python (Qt) interface.
3. Python Libraries:
   • Install the required libraries using:

     pip install pyserial numpy pandas scikit-learn PyQt5 matplotlib

---

Setup Instructions

1. Upload Arduino Code:
   • Open the Arduino IDE.
   • Copy and paste the provided Arduino code.
   • Upload the code to the MCU.
2. Run Python (Qt) Interface:
   • Open the Python script on your computer.
   • Run the script using:

     python enose_gui.py

3. Connect MCU to Computer:
   • Connect the MCU to the computer via USB.
   • Ensure the correct COM port is selected in the Python script.

---

How It Works

1. Sensor Data Acquisition:
   • The MCU reads data from the MQ sensors and sends it to the computer.
2. Data Recording:
   • The Python interface records 150 data points for each smell.
3. Data Filtering:
   • The recorded data is filtered down to 25 points to minimize errors.
4. Model Training:
   • The filtered data is used to train an ANN model.
5. Real-Time Inference:
   • The trained model classifies smells in real-time based on sensor data.

---

Python (Qt) Interface Features

1. Add New Smell Library:
   • Record and save data for new smells.
2. Train Model:
   • Train an ANN model using the recorded data.
3. Import/Export Model:
   • Save and load trained models.
4. Real-Time Inference:
   • Classify smells in real-time.
5. Data Logging:
   • Log sensor data and predictions.
6. Real-Time Graph:
   • Visualize sensor data in real-time.

---

Arduino Code

The Arduino code reads data from the MQ sensors and sends it to the computer via serial communication. Refer to the provided Arduino code for details.

---

Training the Model

1. Record Data:
   • Use the "Add New Smell Library" feature to record data for each smell.
2. Train Model:
   • Click the "Train Model" button to train the ANN model.
3. Evaluate Model:
   • The model's accuracy is displayed after training.

---

Real-Time Inference

1. Start Logging:
   • Click the "Start Logging" button to begin real-time inference.
2. View Predictions:
   • The predicted class is displayed in the GUI.

---

Adding New Smells

1. Click "Add New Smell Library":
   • Enter the name of the new smell.
2. Record Data:
   • The system will record 150 data points for the new smell.
3. Save Data:
   • Save the filtered data to a CSV file.

---

Troubleshooting

1. No Data Received:
   • Check the COM port and ensure the MCU is connected.
2. Low Accuracy:
   • Ensure sufficient data is recorded for each smell.
3. Sensor Errors:
   • Calibrate the sensors and check connections.

---

Future Enhancements

1. Advanced Filtering:
   • Implement more sophisticated filtering techniques.
2. Data Visualization:
   • Add real-time plotting of sensor data.
3. Mobile App:
   • Develop a mobile app for remote monitoring.

---

License

This project is licensed under the MIT License. Feel free to use, modify, and distribute it as needed.

---

Folder Structure

The project is organized as follows:

- Readme.md
- enose_gui.py
- dummy_generator (folder)
     - dummy_generator.ino
- sensors_data (folder)
    - single (folder)
        - single.ino
    - cluster (folder)
        - cluster.ino

• Readme.md: This file, containing the project documentation.
• enose_gui.py: The Python (Qt) interface for data recording, model training, and real-time inference.
• dummy_generator: Contains the dummy_generator.ino file, which simulates sensor data for testing purposes.
• sensors_data: Contains Arduino code for reading sensor data.
  • single: Contains single.ino, which reads data from multiple sensors (1 sensor per type) without multiplexer.
  • cluster: Contains cluster.ino, which reads data from multiple sensors using a multiplexer.

---