This project presents a Wi-Fi-enabled fire alarm system built using two ESP-01S (ESP8266) modules communicating over the MQTT protocol through the HiveMQ public cloud broker. The system is designed to detect early signs of fire by monitoring both temperature and combustible gas levels in real time.

Node 1 is equipped with a DS18B20 digital temperature sensor and a buzzer. It continuously monitors the ambient temperature and automatically triggers the buzzer when the reading exceeds 35°C, serving as a heat-based fire alert.

Node 2 is equipped with an MQ-9 gas sensor and an LED. It detects the presence of combustible gases or carbon monoxide in the environment and activates the LED when gas is detected.

Both nodes transmit sensor data wirelessly to a Node-RED dashboard, which serves as the central monitoring interface. The dashboard displays live temperature readings, gas detection status, and node connectivity. It also allows the user to remotely control Node 1's buzzer and Node 2's LED.

This fire alarm system is suited for homes, small offices, server rooms, laboratories, and any enclosed space where early detection of heat or combustible gas is critical. Key use cases include:

• Home Fire Safety — Monitors kitchen or utility areas for abnormal temperature rises or gas leaks from LPG stoves and detects early warning signs before a fire develops.
• Office and Workplace Safety — Provides continuous background monitoring of temperature and air quality, alerting staff remotely through the Node-RED dashboard without requiring physical presence near the sensors.
• Server Room Monitoring — Detects overheating conditions caused by equipment failure and sends immediate alerts to operators before hardware damage or fire occurs.
• IoT and MQTT Learning Platform — Serves as a hands-on project for students and hobbyists learning Wi-Fi-based IoT communication, MQTT publish-subscribe architecture, and Node-RED dashboard development.
• Remote Monitoring via Smartphone — The Node-RED dashboard is accessible from any browser on any device, allowing users to monitor sensor readings and control alarm outputs remotely from anywhere using Ngrok.

Gather everything below before you start. This is your checklist — names, models, and versions only.

Here is what each component does and why it is part of this project.

The ESP-01S serves as the main controller for each node. It is built around the ESP8266 SoC which integrates a 32-bit microprocessor and a full TCP/IP networking stack, allowing it to connect to Wi-Fi and communicate with the HiveMQ broker independently using the MQTT protocol.

The DS18B20 is a digital temperature sensor that uses the 1-Wire protocol to transmit readings through a single data line. It measures temperatures from -55°C to +125°C with up to 12-bit resolution. In this project, it monitors ambient temperature in Node 1 and automatically triggers the buzzer when the reading exceeds 35°C.

The MQ-9 is a metal oxide semiconductor gas sensor that detects combustible gases such as CO, CH4, and LPG by measuring changes in resistance of a heated tin dioxide element. It requires 5V and a 20 to 30 second warm-up after power-on. In this project, Node 2 uses its digital threshold output (DO) to detect gas presence and trigger the LED alert.

The active buzzer provides audible alerts for Node 1 only. It contains a built-in oscillator circuit that produces sound simply by applying DC voltage, requiring no PWM signal from the microcontroller. It operates on active LOW logic on GPIO0 and activates automatically when temperature exceeds the alert threshold, or remotely via the Node-RED dashboard.

The LED acts as a visual alert indicator for Node 2. It is connected to GPIO0 through a 220Ω current-limiting resistor and operates on active LOW logic. It blinks every 500ms during a gas alert using non-blocking millis() timing to keep the MQTT connection alive during an active alarm.

MQTT is a lightweight publish-subscribe messaging protocol designed for IoT devices. All messages pass through a central broker which routes them by topic instead of direct device-to-device communication. In this project, both nodes and the Node-RED dashboard connect to the HiveMQ free public broker at broker.hivemq.com on port 1883 for real-time data exchange.

Node-RED is a flow-based, low-code programming tool built on Node.js. It serves as the central monitoring dashboard of this project, displaying live temperature and gas readings from both nodes, showing node connection status, and allowing the user to remotely control Node 1's buzzer and Node 2's LED through MQTT.

The MB102 is a breadboard-compatible voltage regulator module that accepts 6.5V to 12V via DC jack and provides stable 3.3V and 5V output rails. It supplies the 3.3V required by the ESP-01S modules and DS18B20 sensor, and the 5V required by the MQ-9 gas sensor heater for proper operation.

The FT232R is a USB to UART bridge IC by FTDI that allows the Arduino IDE to upload firmware to the ESP-01S over serial communication. It presents itself as a virtual COM port to the computer and is only connected during the flashing stage, then disconnected once the upload is complete.

The pushbutton acts as a manual reset trigger for each ESP-01S node. When pressed, it pulls the RST pin to GND, forcing the module to restart. This is useful during development and testing when a manual reboot is needed without disconnecting the power supply.

Ngrok is a reverse tunneling tool that exposes the locally running Node-RED dashboard to the internet by assigning a public HTTPS URL that forwards traffic to the user's machine. It allows the dashboard to be accessed remotely from any device on any network without requiring port forwarding or a static IP address.

Wire the components to the board using the tables below.

Assembly Instructions

Assembly Instructions

Assembly Instructions

Follow these steps in order. Do not skip any steps — they are required for both ESP-01S nodes to work correctly.

1. Open a web browser and go to the official Arduino website at arduino.cc/en/software.
2. Download the latest version of the Arduino IDE compatible with your operating system.
3. Run the installer after the download is complete.
4. Follow the installation instructions and wait for the setup to finish.
5. Open the Arduino IDE after installation.

1. Open the Arduino IDE.
2. Go to File > Preferences.
3. In the Additional Boards Manager URLs field, add the following link:
   http://arduino.esp8266.com/stable/package_esp8266com_index.json
4. Click OK to save the settings.
5. Go to Tools > Board > Boards Manager.
6. Search for ESP8266.
7. Install the ESP8266 board package by ESP8266 Community.
8. After installation, use exactly these settings in Arduino IDE under Tools:

1. In the Arduino IDE, go to Sketch > Include Library > Manage Libraries.
2. Search and install the following libraries one by one:

1. Open the Node 1 code in the Arduino IDE.
2. Find these two lines near the top of the code:

const char* WIFI_SSID     = "YOUR_WIFI_SSID";
const char* WIFI_PASSWORD = "YOUR_WIFI_PASSWORD";

3. Replace YOUR_WIFI_SSID with your Wi-Fi network name.
4. Replace YOUR_WIFI_PASSWORD with your Wi-Fi password.
5. Repeat the same for the Node 2 code.

1. Connect the FT232R USB to UART module to the Node 1 ESP-01S.
2. Pull GPIO0 to GND before powering on to enter flash mode.
3. Connect the FT232R to your PC via USB.
4. Open the Node 1 code in the Arduino IDE.
5. Go to Tools > Board and select Generic ESP8266 Module.
6. Go to Tools > Port and select the correct COM port.
7. Click the Verify button to compile the code and check for errors.
8. After successful verification, click the Upload button.
9. Wait for the upload to complete.
10. Disconnect GPIO0 from GND and reset the module.
11. Open the Serial Monitor at 115200 baud to verify Wi-Fi and MQTT connection.

1. Disconnect the FT232R from Node 1 and connect it to the Node 2 ESP-01S.
2. Pull GPIO0 to GND before powering on to enter flash mode.
3. Open the Node 2 code in the Arduino IDE.
4. Go to Tools > Board and select Generic ESP8266 Module.
5. Go to Tools > Port and select the correct COM port.
6. Click the Verify button to compile the code and check for errors.
7. After successful verification, click the Upload button.
8. Wait for the upload to complete.
9. Disconnect GPIO0 from GND and reset the module.
10. Open the Serial Monitor at 115200 baud to verify Wi-Fi and MQTT connection.
11. Wait 20 to 30 seconds for the MQ-9 sensor to warm up before testing gas detection.

1. Install Node-RED on your Windows PC by following the instructions at nodered.org.
2. Open a terminal and run node-red to start the server.
3. Open a browser and go to http://localhost:1880.
4. Go to Manage Palette and install the node-red-dashboard package.
5. Add an MQTT In node and configure it to connect to broker.hivemq.com on port 1883. This same broker connection can be reused for all MQTT In and MQTT Out nodes in this flow — you only need to configure it once.
6. Subscribe to home/node1/temperature. In the MQTT In node settings, set Output to "a parsed JSON object" so the payload becomes a usable object instead of raw text.
7. Add a Change node after it that sets msg.payload to msg.payload.tempC, then connect it to a Gauge widget.
8. Subscribe to home/node2/gas, set its MQTT In Output to "a parsed JSON object" as well.
9. Add a Change node after it that sets msg.payload to msg.payload.gas, then connect it to a Text widget.
10. Subscribe to home/node1/status and home/node2/status and connect them directly to status indicator widgets — these are already plain text strings, so no parsing is needed.
11. Add an MQTT Out node with topic home/node1/buzzer, connected to a Switch or Button widget for Node 1's buzzer.
12. Add an MQTT Out node with topic home/node2/led, connected to a separate widget for Node 2's LED.
13. For each of these two button/switch widgets, open its edit panel and set the Payload (On) field to the exact string ON and the Payload (Off) field to the exact string OFF, with the payload type set to string.
14. Click the red Deploy button to activate the flow.
15. Open the dashboard at http://localhost:1880/ui and confirm both nodes appear as online.

Copy each file below into the Arduino IDE. Make sure to update the Wi-Fi credentials before uploading as described in Step 4 of the Software Setup section.

Here is what each part of the code does. Read this after uploading to understand how the system works.

1. Initialize serial communication, buzzer pin, and DS18B20 sensor on startup.
2. Connect to Wi-Fi and establish MQTT broker connection.
3. Subscribe to the buzzer control topic for remote commands.
4. Every 5 seconds, read the temperature from the DS18B20 sensor.
5. Publish the temperature as a JSON payload to the broker with retain flag.
6. If temperature exceeds 35°C, activate the buzzer automatically.
7. If a remote ON/OFF command is received, control the buzzer accordingly.
8. If MQTT connection drops, automatically reconnect and re-subscribe.

1. Initialize serial communication and LED pin on startup.
2. Connect to Wi-Fi and establish MQTT broker connection.
3. Subscribe to the LED control topic for remote commands.
4. Every 2 seconds, read the digital output of the MQ-9 gas sensor.
5. Publish the gas status as a JSON payload to the broker with retain flag.
6. If gas is detected, activate the gasAlertActive flag and blink the LED every 500ms.
7. If air returns to normal, clear the alert and turn off the LED.
8. If a remote LED ON/OFF command is received, control the LED accordingly.
9. If MQTT connection drops, automatically reconnect and re-subscribe.

After uploading, verify each of the following to confirm the system is working correctly.

Open the Serial Monitor at 115200 baud after uploading. Both nodes should print a connected message followed by the assigned IP address. If the connection fails after 20 retries, the module will automatically restart.

After Wi-Fi connects, the Serial Monitor should show a message confirming connection to the HiveMQ broker. The node will then publish "online" to its status topic and subscribe to its control topics. If the broker connection fails, the node retries up to 5 times before restarting.

The Serial Monitor for Node 1 should show temperature readings published every 5 seconds in the format {"tempC":27.50,"tempF":81.50}. If the DS18B20 is disconnected or not detected, the monitor will show a sensor_error message instead of temperature data.

To test the automatic temperature alert, carefully apply gentle heat near the DS18B20 sensor (such as holding it briefly between your fingers) until the reading exceeds 35°C. The buzzer should activate immediately. Remove the heat source and the buzzer should stop once the temperature drops back below 35°C.

After the 20 to 30 second warm-up period, the Serial Monitor for Node 2 should show gas readings published every 2 seconds showing {"gas":"normal"}. To test detection, briefly expose the MQ-9 sensor to a combustible gas source such as a lighter (without igniting it). The LED should start blinking and the Serial Monitor should show the gas alert message.

Open the Node-RED dashboard at http://localhost:1880/ui. Both nodes should appear as online in the status indicators. The temperature gauge for Node 1 should update every 5 seconds and the gas status text for Node 2 should update every 2 seconds. Test the remote control buttons by clicking the ON/OFF buttons for Node 1's buzzer and Node 2's LED and confirming the physical devices respond.

The video below shows the complete working system. Use this to verify your output matches the expected behavior.

The system successfully demonstrates two autonomous sensor nodes independently monitoring environmental hazards and reporting to a unified cloud dashboard with minimal latency — typically under two seconds from sensor event to dashboard update.

The DS18B20's digital 1-Wire protocol proved far more robust than analog alternatives, delivering consistent readings without calibration. The MQ-9's digital threshold output simplified gas detection to a single digitalRead(), though the 20 to 30 second warm-up must be accounted for in deployment.

The non-blocking millis() pattern in Node 2's handleAlert() is a critical design decision: using delay() for blink and beep timings would block mqtt.loop() and cause the connection to drop during an alert — the worst possible moment to lose connectivity. The Last Will and Testament mechanism proved its value during testing: pulling power from a node immediately caused the dashboard status indicator to flip to "offline" within the broker's keep-alive window.

HiveMQ's free public broker is ideal for prototyping but unsuitable for production — any device that knows your topic strings can publish or subscribe. Moving to a private broker with TLS and authentication is the first production concern.

• Add a local OLED display (SSD1306 via I2C) on each node so current readings are visible without network access.
• Introduce SMS or push notification alerts via Twilio or Pushover Node-RED nodes so alerts reach the operator even when the dashboard is not open.
• Add a DHT22 to Node 1 to also report relative humidity alongside temperature.

• Espressif Systems — ESP8266 Technical Reference Manual
• Espressif — ESP8266 Arduino Core Board Package Documentation
• Maxim Integrated (Analog Devices) — DS18B20 Digital Thermometer Datasheet
• Zhengzhou Winsen Electronics — MQ-9B Toxic Gas Sensor Datasheet
• FTDI — FT232R USB UART IC Datasheet
• Nick O'Leary — PubSubClient Arduino Library Documentation
• Miles Burton — DallasTemperature Arduino Library Documentation
• HiveMQ — Public MQTT Broker Documentation
• HiveMQ — Web MQTT Client for Testing
• Node-RED Project — Official Documentation
• Node-RED Dashboard — node-red-dashboard Package Documentation
• Ngrok — Official Documentation
• Arduino — ESP8266WiFi Library Reference
• Canva — Online Design Tool, Used for Hardware Schematic and Wiring Diagram
• Paul Stoffregen — OneWire Arduino Library Documentation

• McRaven B. Batapa
• Joshua T. Colendres