Class 9 - Advances in Microcontroller-Based System Design

Objective:
To dive deeper into the recent advancements in microcontroller technology, focusing on IoT
integration, AI capabilities, and their role in modern applications. The session will provide
practical knowledge alongside hands-on activities and real-world case studies.

1. Introduction to Microcontroller Advancements

Overview

 Start by explaining microcontrollers as embedded systems with a processor, memory, and

peripherals integrated onto a single chip.
 Emphasize the need for advancements:
o Demand for connected, efficient, and intelligent devices.
o Growth in IoT, AI, automation, and green energy sectors.

Key Discussion Points

1. How Traditional Microcontrollers Work:

o Example: Basic Arduino controlling LED or sensors.
2. Shifting to Advanced Systems:
o From standalone systems to interconnected smart systems.
o Example: ESP32 with built-in Wi-Fi and BLE for IoT.

2. Trends in Microcontroller Technology

IoT Integration

 Definition of IoT: Internet of Things connects devices to collect and exchange data.
 Microcontroller Role in IoT:
o Connectivity (Wi-Fi, BLE, ZigBee, LoRa).
o Real-time processing of sensor data.
o Communication with cloud platforms (AWS IoT, Google Firebase).

Example:

 ESP32: IoT-enabled microcontroller with dual-core processor, Wi-Fi, and BLE.

 Applications:
o Smart home (light automation).
 o Wearables (fitness trackers).

AI Integration

 Microcontrollers with AI Capabilities:

o Running AI models directly on the microcontroller (edge AI).
o Example: TensorFlow Lite on Arduino Portenta H7.
 Use Cases:
o Object Detection: AI-powered camera using STM32 microcontroller.
o Predictive Maintenance: Monitoring industrial machines for faults.

Case Study: AI-powered microcontroller detecting vehicle traffic patterns.

Improved Power Efficiency

 Ultra-low-power microcontrollers for battery-operated devices.

 Example: ARM Cortex-M0+ for wearables.
 Techniques: Sleep modes, power optimization, efficient clock cycles.

Enhanced Security

 Challenges in IoT: Vulnerabilities in connected devices.

 Solutions: Hardware-based encryption and secure boot processes.
o Example: Microchip SAM L10/L11 with TrustZone technology.

Diagram: Show an IoT microcontroller with encryption in communication (e.g., HTTPS).

3. Modern Applications of Microcontrollers
Applications and Examples

Application Role of Microcontroller Example

IoT Systems Process sensor data, communicate with cloud ESP8266 in smart lighting
AI-Powered Arduino Portenta in object
Edge AI for local decision-making
Devices detection
Healthcare
Monitoring vitals, automating drug delivery STM32 in a pulse oximeter
Devices
ATmega328 in solar
Renewable Energy Managing solar power and energy distribution
inverter
Autonomous Controlling sensors, actuators, and decision- STM32 in adaptive cruise
Vehicles making in real-time control
Here’s a detailed explanation of each example

1. IoT Systems: Smart Lighting using ESP8266

Description:

 IoT (Internet of Things): A system where devices communicate over the internet to
perform specific tasks.
 Smart Lighting: Uses IoT to automatically adjust lights based on environmental
conditions or user preferences.

Role of ESP8266:

 ESP8266 is a low-cost Wi-Fi-enabled microcontroller that connects the lighting system to

the internet.
 It processes data from sensors like light sensors or motion detectors.
 It communicates with a cloud platform (e.g., AWS IoT, Blynk) to enable remote control
and automation.

Example Workflow:

1. A light sensor detects ambient light levels.

2. The ESP8266 reads this data.
3. If the light level is low, the ESP8266 turns on the light via a relay module.
4. Users can control the lights remotely through a mobile app or a web dashboard.

Advantages:

 Energy-efficient.
 Provides remote control and scheduling features.

2. AI-Powered Devices: Object Detection using Arduino Portenta

Description:

 Edge AI: Running artificial intelligence algorithms directly on the device

(microcontroller), without needing cloud processing.
 Object Detection: Identifying objects in an image or video stream.

Role of Arduino Portenta:

  Arduino Portenta is a high-performance microcontroller that supports TensorFlow Lite,
enabling AI model deployment on the edge.
 A camera module connected to Portenta captures images.
 The microcontroller runs an AI model to detect objects like vehicles, people, or specific
items.

Example Workflow:

1. A camera captures an image.

2. The AI model (e.g., YOLO or MobileNet) processes the image on the Portenta.
3. Detected objects are classified, and the result (e.g., "Car detected") is displayed or used to
trigger an action.

Applications:

 Security systems (e.g., detecting intruders).

 Industrial automation (e.g., quality control in manufacturing).

Advantages:

 Real-time processing without internet dependency.

 Reduced latency and better privacy.

What is TensorFlow Lite

TensorFlow Lite (TFLite) is a lightweight version of TensorFlow designed for running machine
learning models on edge devices like microcontrollers, mobile phones, and embedded systems. It
enables AI inference with low latency and minimal power consumption, making it ideal for real-
time applications.

🔹 Key Features of TensorFlow Lite

✅ Optimized for Edge AI – Runs on low-power devices like Arduino Portenta, Raspberry Pi,
ESP32, and smartphones
✅ Supports Quantization – Reduces model size and improves efficiency without much loss in
accuracy
✅ Fast Inference – Performs AI tasks quickly with minimal resource usage
✅ Cross-Platform Support – Works on Android, iOS, Linux, Windows, and embedded systems
✅ Pre-trained Model Compatibility – Deploys models trained in TensorFlow with conversion to
TFLite format

🔹 Common Applications
🔹 Image Recognition – Identify objects, faces, or text using a camera
🔹 Speech & Gesture Recognition – Process voice commands or hand movements
🔹 Anomaly Detection – Detect unusual patterns in industrial machines
🔹 Predictive Maintenance – Forecast failures in IoT devices
3. Healthcare Devices: Pulse Oximeter using STM32

Description:

 Pulse Oximeter: A device that measures the oxygen saturation (SpO₂) level in the blood
and pulse rate.
 Widely used in hospitals and homes for health monitoring.

Role of STM32:

 STM32 microcontroller processes signals from sensors like photoplethysmography (PPG)

and calculates SpO₂ and pulse rate.
 It powers the display for showing the readings and manages power efficiency for long
battery life.

Example Workflow:

1. LEDs (infrared and red) shine light through the user’s finger.
2. A photodetector captures the light intensity after it passes through the blood vessels.
3. The STM32 processes the data using algorithms to compute SpO₂ and pulse rate.
4. Results are displayed on an LCD or sent to a mobile app via Bluetooth.

Advantages:

 Portable and battery-efficient.

 Provides critical health metrics in real-time.

4. Renewable Energy: Solar Inverter using ATmega328

Description:

 Solar Inverter: Converts DC power generated by solar panels into AC power for
household appliances.
 Ensures optimal energy management in renewable energy systems.

Role of ATmega328:

 Reads input from solar panels (e.g., voltage and current).

 Controls the inverter circuit to convert DC to AC.
 Monitors battery status and implements maximum power point tracking (MPPT) for
efficiency.

Example Workflow:
 1. ATmega328 reads solar panel data (e.g., voltage, current).
2. It adjusts the duty cycle of a PWM signal to maximize power output using MPPT.
3. Converts DC to AC using an H-bridge inverter circuit.
4. Monitors battery charge and switches to the grid if necessary.

Advantages:

 Increases energy efficiency.

 Reduces dependency on non-renewable energy sources.

5. Autonomous Vehicles: Adaptive Cruise Control using STM32

Description:

 Adaptive Cruise Control: Automatically adjusts a vehicle's speed based on traffic

conditions, maintaining a safe distance from the vehicle ahead.
 A key feature in autonomous and semi-autonomous vehicles.

Role of STM32:

 Processes data from sensors like radar, LIDAR, or cameras.

 Runs algorithms to calculate the distance from the vehicle ahead.
 Sends control signals to adjust the vehicle’s throttle and brake systems.

Example Workflow:

1. A radar sensor measures the distance to the vehicle ahead.

2. STM32 calculates the speed difference and determines the required speed adjustment.
3. If the vehicle ahead slows down, STM32 sends a signal to apply the brakes.
4. Once the road is clear, STM32 restores the set cruising speed.

Advantages:

 Improves safety by preventing collisions.

 Reduces driver fatigue in long-distance travel.
Comparison Table

Example Use
Application Microcontroller Key Role
Case
Processes sensor data, connects to the
IoT Systems ESP8266 Smart lighting
internet
AI-Powered Arduino
Runs AI models on the edge Object detection
Devices Portenta
Healthcare Monitors health metrics and processes
STM32 Pulse oximeter
Devices sensor signals
Renewable Controls and optimizes energy
ATmega328 Solar inverter
Energy conversion
Autonomous Manages sensors and actuators for real- Adaptive cruise
STM32
Vehicles time decision-making control

Exam Questions with Detailed Answers

Question 1:

Explain how microcontrollers can integrate with AI for edge computing. Provide an example
of a microcontroller and a use case to support your answer.

Answer:

Microcontrollers and AI for Edge Computing: Microcontrollers with AI capabilities enable the
processing of AI models locally, without relying on cloud services. This is called Edge AI, which
minimizes latency, reduces power consumption, and ensures data privacy. Modern
microcontrollers, such as Arduino Portenta H7, STM32, and ESP32, support frameworks like
TensorFlow Lite for Microcontrollers, enabling the deployment of pre-trained AI models.

Example:

 Microcontroller: Arduino Portenta H7.

 Use Case: Object detection for security systems.
 How it works:
o A camera connected to the Arduino Portenta captures images.
o The microcontroller runs a lightweight AI model (e.g., YOLO or MobileNet) to
detect objects like people, vehicles, or animals.
o Once an object is detected, it triggers alarms or notifications.
Real-World Scenario: Consider a wildlife monitoring system where Arduino Portenta H7 is
installed in a forest. The microcontroller processes live camera feed and identifies specific animals.
If an endangered species is detected, the system sends a real-time alert to wildlife conservationists.
This solution avoids the need for constant internet connectivity and processes data on-site.

Question 2:

Describe the role of AI-powered microcontrollers in predictive maintenance. Provide a

detailed example of a use case in an industrial setting.

Answer:

Predictive Maintenance Using AI-powered Microcontrollers: Predictive maintenance

leverages AI to monitor the health of industrial machines, predict faults before they occur, and
schedule maintenance accordingly. AI models, when deployed on microcontrollers, analyze sensor
data in real time and detect anomalies or patterns indicating equipment failure.

Example:

 Microcontroller: STM32 microcontroller.

 Use Case: Monitoring the vibration levels of industrial motors.
 How it works:
1. Data Collection: Vibration sensors (accelerometers) are connected to the STM32
microcontroller.
2. AI Analysis: An AI model pre-trained to identify abnormal vibration patterns runs
on the microcontroller.
3. Fault Prediction: When the vibration exceeds normal thresholds, the system
predicts potential motor failure and raises an alert.

Real-World Scenario: In a manufacturing plant, an STM32-based predictive maintenance system

is deployed on multiple conveyor belt motors. The system continuously analyzes motor vibrations.
If one motor shows signs of wear and tear (e.g., excessive vibration), the system notifies
technicians, preventing costly downtime and repairs.

Question 3:

What are the advantages of using AI-powered microcontrollers for traffic monitoring?
Illustrate your answer with a case study.
Answer:

Advantages of AI-powered Microcontrollers for Traffic Monitoring:

1. Real-time Processing: Processes data directly on-site with minimal latency.

2. Energy Efficiency: Optimized for low-power operations in remote or urban environments.
3. Data Privacy: Keeps sensitive traffic data local, reducing cybersecurity risks.

Case Study:

 System: AI-powered STM32 microcontroller for traffic monitoring.

 Objective: Detect and analyze vehicle traffic patterns to optimize urban traffic flow.
 Implementation:
1. A camera connected to the STM32 microcontroller captures live footage of a busy
intersection.
2. An AI model, trained to count vehicles and classify them (e.g., car, truck, bike), is
deployed on the microcontroller using TensorFlow Lite.
3. The system identifies traffic congestion and sends data to city traffic management
for dynamic traffic light adjustment.

Outcome: This system improves traffic flow by dynamically adjusting signal timings based on
real-time traffic patterns. Additionally, the AI-powered microcontroller operates independently,
ensuring consistent performance even in areas with limited internet connectivity.

Summary of Key Points for Students:

 Microcontroller AI Integration: AI models like TensorFlow Lite for Microcontrollers

enable advanced edge computing capabilities.
 Applications: Object detection, predictive maintenance, traffic monitoring, etc.
 Examples: Arduino Portenta H7 for object detection, STM32 for vibration analysis in
motors, and traffic pattern monitoring.

Edge Computing refers to a computing paradigm where data processing and analysis are
performed close to the source of data generation, rather than relying on a centralized cloud server.
In edge computing, computation happens on devices like microcontrollers, edge devices, or local
gateways, ensuring faster processing, reduced latency, and greater privacy.

Key Features of Edge Computing:

1. Local Data Processing:

o Instead of sending data to a distant cloud server, the data is processed locally at or
near the device collecting it.
 o Example: A microcontroller processes temperature readings from a sensor directly
on-site.
2. Low Latency:
o Reduces the time needed to process and respond to data.
o Example: A self-driving car detects obstacles in real-time without relying on the
cloud.
3. Improved Data Privacy:
o Sensitive data remains local, minimizing the risk of exposure during transmission.
o Example: A wearable health monitor processes heart rate locally instead of
uploading data to the cloud.
4. Bandwidth Savings:
o Only processed or important data is sent to the cloud, reducing the load on
networks.
o Example: A security camera streams video to the cloud only when motion is
detected.
5. Increased Reliability:
o Local processing ensures the system works even when the internet connection is
unavailable.
o Example: An edge AI-powered traffic light adjusts signal timing even during
network outages.

Example of Edge Computing in AI:

 Scenario: Object detection with a microcontroller.

 Device: Arduino Portenta H7 running a TensorFlow Lite model.
 Operation:
o A camera captures live images.
o The microcontroller processes the image locally to identify objects (e.g., cars,
pedestrians).
o The detected objects trigger local actions, like turning on alarms or sending
selective data to the cloud.

Why Use Edge Computing?

1. Faster Response: Real-time decision-making without depending on cloud processing.

2. Cost Efficiency: Reduces expenses associated with cloud storage and data transfer.
3. Autonomous Functionality: Works in remote areas where internet connectivity is
unreliable.
Edge computing is particularly beneficial for IoT systems, AI-powered devices, and industrial
automation, where speed, privacy, and efficiency are critical.

Comparison of Arduino Portenta H7, STM32, and ESP32

Below is a detailed comparison of these popular microcontrollers based on their features,

performance, and applications.

Example Applications

Microcontroller Example Applications

Arduino Portenta AI-based image recognition, industrial IoT systems, real-time predictive
H7 maintenance.
Motor control systems, medical devices (e.g., ECG monitors), advanced
STM32
automotive systems.
Smart lighting, home automation (e.g., Google Home clones), wireless
ESP32
sensor networks for agriculture.

Here’s a general and simplified comparison of Arduino Portenta H7, STM32, and ESP32 that’s
easier to understand for beginners:

Aspect Arduino Portenta H7 STM32 ESP32

A luxury sports car— A budget-friendly
What it’s A reliable sedan—
powerful, high-tech, and scooter—simple,
like versatile and flexible
expensive affordable, and effective
Cutting-edge projects like Affordable and fun IoT
Industrial machines,
Best For AI, Machine Learning, or projects like smart lights
robots, and appliances
industrial IoT or gadgets
Super fast and powerful forMedium to fast, Fast enough for small to
Speed
advanced tasks depending on the model medium projects
High—usually for serious Medium—can fit both Low—great for budget-
Price
professional projects hobby and pro needs friendly projects
Can be tricky for
A bit harder to learn but Beginner-friendly, with
Complexity beginners, but very
extremely capable many tutorials available
customizable
Lots of options like Wi-Fi, Depends on the model, Built-in Wi-Fi and
Connectivity Bluetooth, LoRa, and but can include Ethernet Bluetooth make it super
Ethernet or USB handy
Moderate power use,
Uses a lot of power because Low power, ideal for
Energy Use optimized for industrial
it’s so powerful battery projects
setups
Size of Ideal for big, ambitious Can handle medium to Perfect for small to
Projects projects large projects medium projects
 Aspect Arduino Portenta H7 STM32 ESP32
Example - Robots and automation - Smart home devices
- AI-powered cameras
Uses systems like smart bulbs
- Advanced traffic control - Medical devices like - Wireless weather
systems heart monitors stations
- Predictive maintenance in - Affordable hobbyist
- High-tech vehicles
factories projects

Real-Life Analogy

1. Arduino Portenta H7
Imagine you want to build a high-tech robot that can recognize faces, talk, and work in a
factory. You’d pick Arduino Portenta H7 because it’s like a futuristic robot brain—
powerful but costly.
2. STM32:
Let’s say you’re making a smart washing machine or an industrial conveyor belt system.
STM32 is like a sturdy worker who’s very adaptable, reliable, and good at many tasks.
3. ESP32:
Suppose you’re making a smart light bulb for your home or a simple wireless weather
monitor. ESP32 is like an affordable gadget that’s easy to use and gets the job done without
breaking the bank.

In general:

 If you’re doing big and advanced projects, use Arduino Portenta H7.
 For customized industrial systems, pick STM32.
 If you’re a beginner or on a budget, ESP32 is the go-to choice!

Here’s a simplified explanation of the differences between Wi-Fi, Bluetooth, LoRa, and Ethernet,
including when you might use each one:

What It’s Power

Technology Range Speed Best Use Cases
Like Usage
- Streaming data
Wireless Medium (up to
High (10- Medium - IoT devices that need
Wi-Fi internet 50-100 meters
100+ Mbps) to high fast communication
connection indoors)
- Home automation
- Wearables
Short-range Short (up to 10- Medium (1- Very low (smartwatches, fitness
Bluetooth
wireless link 30 meters) 3 Mbps) to low trackers)
- Wireless
 What It’s Power
Technology Range Speed Best Use Cases
Like Usage
speakers/headphones
- Local device pairing
- Remote sensors
Long-range, (agriculture, weather
Very long (up to Very low
LoRa low-speed Very low stations)
10+ kilometers) (<1 Mbps)
signal - Smart cities (meter
readings)
- Industrial equipment
Physical cable - Devices that require
Wired Very high
length (100+ stable and fast data
Ethernet network (100 Mbps High
meters with transfer
connection to 1 Gbps+)
boosters) - Office/home wired
internet

Explanation by Key Features

1. Wi-Fi
o
Range: Covers most homes or offices but weakens through walls.
o
Speed: Excellent for streaming or transferring large data files.
o
Power Usage: Uses more energy compared to others, not ideal for battery-powered
devices.
o Use Case: Smart TVs, IoT cameras, or any device requiring internet access.
2. Bluetooth
o Range: Works well within a room or two but struggles beyond that.
o Speed: Decent for sending files, music, or commands.
o Power Usage: Very low, great for small, portable, battery-operated gadgets.
o Use Case: Connecting fitness trackers, smartwatches, or wireless headphones.
3. LoRa (Long Range)
o Range: Fantastic for rural areas or cities, even through obstacles.
o Speed: Very slow—meant for sending small packets like sensor readings.
o Power Usage: Extremely efficient, so devices can run on batteries for years.
o Use Case: Agriculture (monitoring soil moisture), smart meters, or tracking
wildlife.
4. Ethernet
o Range: Limited by cable length but very reliable for connected devices.
o Speed: Super-fast and stable; perfect for streaming or transferring huge files.
o Power Usage: High since it powers both the device and cable connection.
o Use Case: Servers, gaming setups, or factory automation where reliability matters.
Simple Analogy

 Wi-Fi: Like wireless broadband at home—fast and reliable but only works within a specific
distance.
 Bluetooth: Like whispering to a friend close by—great for nearby communication but
limited range.
 LoRa: Like sending a slow but reliable message across a long distance with a carrier
pigeon.
 Ethernet: Like having a direct line of communication—fast, steady, and secure, but you’re
tethered to the cable.

Which One to Choose?

 If your device needs fast data transfer and internet access, go for Wi-Fi.
 For short-range, low-power connectivity, choose Bluetooth.
 When you need long-range communication with minimal data, use LoRa.
 For stable, fast, and wired setups, Ethernet is your best option.

Cloud Computing - Simplified Explanation

Cloud computing means storing and accessing data and applications over the internet instead
of using your computer's hard drive. It’s like renting someone else’s powerful computer system to
do your work.

Key Features of Cloud Computing

1. Access Anywhere
o As long as you have the internet, you can access your data or software from any
device.
2. Pay-as-You-Go
o You only pay for what you use, like paying for electricity or water.
3. No Physical Storage
o Your data is stored on remote servers (the "cloud") rather than on your own devices.
Simple Analogy

Imagine you need a car occasionally. Instead of buying one, you rent a car when you need it.
Similarly, instead of owning your own massive computer servers, you use the "cloud" services
when you need them.

How Does It Work?

1. Your data is stored in remote servers owned by cloud providers like Amazon, Google, or
Microsoft.
2. You can use the internet to access tools, run software, or process data on those servers.
3. These servers handle everything, so you don’t need to worry about maintenance or updates.

Examples in Real Life

1. Google Drive – Store and share files online.

2. Netflix – Stream movies and shows (data stored in the cloud).
3. Amazon Web Services (AWS) – Hosts websites, apps, or even virtual servers for
businesses.

Benefits of Cloud Computing

1. Cost-Effective: No need to buy expensive hardware or maintain it.

2. Scalability: Increase or decrease resources based on demand.
o Example: More storage during peak usage, less during off-peak.
3. Flexibility: Work from anywhere with an internet connection.
4. Backup and Security: Data is safely stored and backed up on secure servers.

Types of Cloud Computing

1. Public Cloud: Shared by many users, like renting a shared apartment (e.g., AWS, Google
Cloud).
2. Private Cloud: Dedicated to one user or company, like owning your own house.
3. Hybrid Cloud: A mix of public and private cloud, combining flexibility and security.
In Short

Cloud computing is a modern way to store, manage, and process data online instead of relying
on your own computer. It’s widely used in businesses, apps, and everyday life because it’s
convenient, cost-effective, and scalable.

Lightweight AI Models: An Overview

Lightweight AI models are designed to run efficiently on devices with limited computing power,
such as microcontrollers, smartphones, or edge devices. These models are optimized for speed,
low memory usage, and minimal energy consumption, making them ideal for real-time
applications in resource-constrained environments.

Popular Lightweight AI Models

1. YOLO (You Only Look Once)

o Purpose: Real-time object detection and classification.
o Key Feature: Processes the entire image at once rather than scanning regions,
making it fast and efficient.
o Variants: YOLOv4-tiny and YOLOv5-nano are lightweight versions for edge
devices.
o Example Use Case:
 Detecting cars, pedestrians, and road signs in a smart traffic system using
an NVIDIA Jetson Nano.
2. MobileNet
o Purpose: Image classification and object recognition.
o Key Feature: Uses depthwise separable convolutions to reduce computation
without sacrificing accuracy.
o Variants: MobileNetV2 and MobileNetV3 are further optimized versions.
o Example Use Case:
 Running image recognition on smartphones for augmented reality (AR)
applications.

Characteristics of Lightweight AI Models

1. Compact Architecture:
o Designed with fewer parameters and layers to reduce size.
2. Optimized Computation:
o Efficient operations like depthwise separable convolutions (used in MobileNet).
3. Lower Accuracy Trade-Off:
o Slight reduction in accuracy compared to larger models, but acceptable for many
real-world applications.
 4. Cross-Platform Compatibility:
o Can run on CPUs, GPUs, and even specialized hardware like TPUs or
microcontrollers.

How Lightweight Models Work

1. Data Input: Receive an image or sensor data.

2. Feature Extraction: Extract key features using efficient neural network layers.
3. Inference: Predict outcomes like detecting objects, recognizing faces, or classifying
images.
4. Output: Return results (e.g., bounding boxes for objects or category labels).

Comparison: YOLO vs. MobileNet

Aspect YOLO MobileNet

Primary Use Object Detection Image Classification
Speed Faster, real-time detection Efficient, not as fast as YOLO
Complexity Slightly higher Simpler and more compact
Accuracy High for detection tasks High for classification tasks
Example Detecting vehicles on roads Recognizing plant species

Real-Life Applications of Lightweight AI Models

1. Smart Cameras:
o YOLO on edge devices for real-time surveillance.
2. Mobile Apps:
o MobileNet for facial recognition and filters in apps like Snapchat.
3. IoT Devices:
o Lightweight AI models embedded in smart home devices for detecting motion or
recognizing objects.
4. Wearables:
o MobileNet for fitness tracking or gesture recognition in smartwatches.

Advantages of Lightweight Models

1. Efficient Performance: Run smoothly on devices with limited hardware.

2. Low Power Consumption: Ideal for battery-powered devices.
 3. Real-Time Operation: Provide immediate results for time-critical tasks.

In Simple Terms

Think of lightweight AI models as compact and energy-efficient versions of full-sized AI models.

They may not have the power of a big computer, but they’re perfectly suited for smaller, real-time
tasks like identifying objects, recognizing faces, or controlling smart devices.