AI in IoT with Python - Rishan Solutions

The combination of Artificial Intelligence (AI) and the Internet of Things (IoT) creates smart, autonomous, and predictive systems. AI-powered IoT solutions use machine learning, deep learning, and analytics to make real-time decisions, optimize performance, and detect anomalies.

1. Why Combine AI with IoT?

Predictive Maintenance – AI detects potential failures in IoT devices.
Anomaly Detection – AI spots unusual behavior in IoT sensor data.
Autonomous Decision-Making – AI processes IoT data without human intervention.
Energy Optimization – AI adjusts power consumption based on real-time usage.
Real-Time Analytics – AI extracts insights from IoT sensor data.

2. AI in IoT Use Cases

Smart Homes – AI-powered assistants adjust lights, temperature, and security.
Industrial Automation – AI predicts machine failures and optimizes operations.
Healthcare – AI analyzes patient vitals from wearable devices.
Smart Cities – AI optimizes traffic, waste management, and power grids.
Agriculture – AI analyzes soil moisture and predicts crop yields.

3. AI + IoT Architecture

IoT Device Layer – Sensors collect real-time data.

Edge Computing Layer – AI processes data on IoT devices before sending it to the cloud.

Cloud Processing Layer – AI models analyze data for predictions and automation.

Application Layer – Dashboards visualize insights for users.

4. Collecting IoT Data for AI Processing

IoT devices collect data using sensors and send it to cloud platforms like AWS, Azure, or Firebase.

Example: Collecting Sensor Data from IoT Devices

import random
import time

def get_sensor_data():
    return {"temperature": random.uniform(20, 35), "humidity": random.uniform(30, 70)}

while True:
    data = get_sensor_data()
    print("Collected Sensor Data:", data)
    time.sleep(5)

This simulates IoT sensors collecting temperature and humidity data.

5. Storing IoT Data for AI Analysis

IoT data is stored in databases like MongoDB, Firebase, or InfluxDB for real-time processing.

Example: Storing IoT Data in MongoDB

pip install pymongo

import pymongo

client = pymongo.MongoClient("mongodb://localhost:27017/")
db = client["iot_db"]
collection = db["sensor_data"]

data = {"temperature": 28.5, "humidity": 65}
collection.insert_one(data)
print("Data stored in MongoDB")

This saves IoT sensor data to a MongoDB database.

6. Processing IoT Data with AI

Once data is collected, we can apply Machine Learning (ML) models for prediction and automation.

Example: Training an AI Model on IoT Data

We use Scikit-Learn to train a regression model that predicts temperature based on humidity.

pip install pandas scikit-learn

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression

# Sample IoT data
data = {"temperature": [22, 24, 26, 28, 30, 32, 34], "humidity": [40, 45, 50, 55, 60, 65, 70]}
df = pd.DataFrame(data)

X = df[["humidity"]]
y = df["temperature"]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = LinearRegression()
model.fit(X_train, y_train)

humidity_input = [[50]]
predicted_temperature = model.predict(humidity_input)
print("Predicted Temperature:", predicted_temperature[0])

This trains an AI model to predict temperature based on humidity.

7. Real-Time AI Predictions on IoT Data

We can use AI on live IoT data to trigger alerts or automate actions.

Example: AI Model Predicting IoT Data in Real-Time

import random

def predict_temperature(humidity):
    return model.predict([[humidity]])[0]

while True:
    humidity = random.uniform(30, 70)
    predicted_temp = predict_temperature(humidity)
    print(f"Humidity: {humidity:.2f}% → Predicted Temperature: {predicted_temp:.2f}°C")
    
    if predicted_temp > 30:
        print("Alert: High temperature detected!")
    
    time.sleep(5)

This runs real-time AI predictions on IoT data.

8. Deploying AI Models on IoT Edge Devices

Instead of sending data to the cloud, we can run AI models on edge devices like Raspberry Pi, Jetson Nano, or ESP32 to reduce latency.

Deploying AI Model on Raspberry Pi

Train the AI model on a powerful machine.
Convert the model to TensorFlow Lite for edge deployment.
Load the model on Raspberry Pi to process IoT data locally.

pip install tensorflow

import tensorflow as tf

model.save("iot_model.h5")

This saves the AI model for deployment on an IoT edge device.

9. AI-Based IoT Automation

AI can automate IoT actions based on real-time data analysis.

Example: AI Automatically Adjusting a Smart AC System

def adjust_ac_temperature(predicted_temp):
    if predicted_temp > 30:
        print("Turning AC ON")
    else:
        print("Turning AC OFF")

humidity = 60
predicted_temp = predict_temperature(humidity)
adjust_ac_temperature(predicted_temp)

This turns the AC on/off based on AI predictions.

10. AI-Based IoT Anomaly Detection

AI can detect faults, intrusions, or unusual activity in IoT devices.

Example: Detecting Anomalies in IoT Sensor Data

pip install pyod

from pyod.models.iforest import IForest

X_train = [[20], [21], [22], [23], [80]]  # 80 is an anomaly
model = IForest()
model.fit(X_train)

new_data = [[24], [85]]
predictions = model.predict(new_data)  # 1 means anomaly, 0 means normal
print("Anomaly Detection Results:", predictions)

This detects unusual sensor data values.

11. Cloud AI for IoT

We can use cloud AI services for IoT analytics:
🔹 AWS SageMaker – Trains ML models on IoT data.
🔹 Google Cloud AI – Uses AI to analyze IoT data streams.
🔹 Azure Machine Learning – Runs AI models on cloud IoT data