Raspberry Pi

Raspberry Pi Models for IoT

Protocol Selection Guide

Gateway Capabilities

The Raspberry Pi excels as a gateway because it can:

1. Protocol translation: Convert BLE, Zigbee, or LoRa data into MQTT or HTTP for cloud delivery
2. Edge filtering: Reduce data volume by filtering noise and aggregating readings before transmission
3. Local buffering: Store data in SQLite or InfluxDB during network outages for later sync
4. Security enforcement: Run TLS/SSL, implement access control, and manage device certificates
5. Edge ML: Run lightweight inference models (TensorFlow Lite, ONNX) to detect anomalies locally

Problem Statement

Build a system that reads temperature, humidity, and barometric pressure every 30 seconds, stores data locally, and publishes readings to an MQTT broker for remote monitoring.

Step 1: Hardware Setup

Components needed:

• Raspberry Pi 4B (or newer)
• BME280 sensor breakout board (I2C)
• Jumper wires (4: VCC, GND, SDA, SCL)
• 5V power supply

Wiring (I2C connection):

Step 2: Software Setup

# Enable I2C interface
sudo raspi-config nonint do_i2c 0

# Install dependencies
pip3 install adafruit-circuitpython-bme280 paho-mqtt

# Verify sensor is detected (should show address 0x76 or 0x77)
i2cdetect -y 1

Step 3: Python Application

import time
import json
import board
import adafruit_bme280.advanced as adafruit_bme280
import paho.mqtt.client as mqtt

# Initialize I2C sensor
i2c = board.I2C()
bme280 = adafruit_bme280.Adafruit_BME280_I2C(i2c, address=0x76)

# Configure MQTT
client = mqtt.Client(client_id="rpi-env-monitor")
client.tls_set()  # Enable TLS for security
client.username_pw_set("iot_user", "secure_password")
client.connect("mqtt.example.com", 8883)
client.loop_start()

INTERVAL = 30  # seconds

try:
    while True:
        reading = {
            "temperature_c": round(bme280.temperature, 2),
            "humidity_pct": round(bme280.relative_humidity, 2),
            "pressure_hpa": round(bme280.pressure, 2),
            "timestamp": time.time()
        }

        # Publish to MQTT
        client.publish(
            "home/livingroom/environment",
            json.dumps(reading),
            qos=1  # At-least-once delivery
        )
        print(f"Published: {reading}")
        time.sleep(INTERVAL)

except KeyboardInterrupt:
    client.loop_stop()
    client.disconnect()

Step 4: Run as a System Service

# /etc/systemd/system/env-monitor.service
[Unit]
Description=IoT Environment Monitor
After=network-online.target
Wants=network-online.target

[Service]
ExecStart=/usr/bin/python3 /home/pi/env_monitor.py
Restart=always
RestartSec=10
User=pi

[Install]
WantedBy=multi-user.target

sudo systemctl enable env-monitor
sudo systemctl start env-monitor

Expected Output

The system publishes JSON payloads every 30 seconds:

{
  "temperature_c": 22.45,
  "humidity_pct": 48.12,
  "pressure_hpa": 1013.25,
  "timestamp": 1706625000.0
}

Key design decisions in this example:

• I2C chosen over SPI because only one sensor is needed and I2C uses fewer wires
• QoS 1 ensures readings are delivered even if the network briefly drops
• systemd service ensures the monitor restarts automatically after power failures
• TLS encryption secures data in transit to the MQTT broker