<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Circuit Digest on Medium]]></title>
        <description><![CDATA[Stories by Circuit Digest on Medium]]></description>
        <link>https://medium.com/@circuitdigest?source=rss-d4da06c4ac74------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*vlkrl-HVCYFcRZCTyIcSEA.png</url>
            <title>Stories by Circuit Digest on Medium</title>
            <link>https://medium.com/@circuitdigest?source=rss-d4da06c4ac74------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 22:21:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@circuitdigest/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Raspberry Pi Object Detection Made Simple]]></title>
            <link>https://circuitdigest.medium.com/raspberry-pi-object-detection-made-simple-7c81e40d48ad?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/7c81e40d48ad</guid>
            <category><![CDATA[ai-projects]]></category>
            <category><![CDATA[raspberry-pi]]></category>
            <category><![CDATA[iot-projects]]></category>
            <category><![CDATA[object-detection]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Tue, 30 Jun 2026 11:39:24 GMT</pubDate>
            <atom:updated>2026-06-30T11:39:24.799Z</atom:updated>
            <content:encoded><![CDATA[<h4>Build an AI-powered vision system with Raspberry Pi and explore real-world object detection</h4><p>Artificial intelligence is no longer limited to cloud servers or expensive hardware. Thanks to affordable single-board computers like the Raspberry Pi, anyone can build intelligent computer vision applications capable of identifying everyday objects in real time. Whether you’re a student, electronics enthusiast, maker, or embedded developer, object detection is one of the most exciting AI projects you can add to your portfolio.</p><p>In this article, we’ll explore how object detection works on a Raspberry Pi, why it has become increasingly popular, and what it takes to build a practical AI vision system. If you’re looking for the complete implementation, including hardware setup, software installation, source code, and deployment, you can find the full step-by-step tutorial on Circuit Digest.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*J974hwniA-Lh19fa_CVmRw.jpeg" /></figure><h3>Why Object Detection Matters</h3><p>Object detection is a branch of computer vision that enables a machine to recognize and locate objects within an image or video stream. Unlike image classification, which only identifies what an image contains, object detection determines both <em>what</em> the object is and <em>where</em> it appears.</p><p>This capability powers countless real-world applications, including:</p><ul><li>Smart surveillance systems</li><li>Autonomous robots</li><li>Traffic monitoring</li><li>Industrial quality inspection</li><li>Retail inventory management</li><li>Smart agriculture</li><li>Wildlife monitoring</li><li>Home automation</li></ul><p>As edge AI continues to evolve, running these models directly on small devices has become more practical than ever.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*9BdXitIIAEOZYkVjMuwyVA.jpeg" /></figure><h3>Why Choose Raspberry Pi?</h3><p>The Raspberry Pi has become one of the most popular platforms for learning embedded systems and artificial intelligence. Despite its compact size and low cost, it offers enough processing power to run lightweight AI inference locally.</p><p>Some advantages include:</p><ul><li>Affordable hardware</li><li>Strong community support</li><li>Extensive software ecosystem</li><li>Compatibility with Python</li><li>Easy camera integration</li><li>GPIO support for automation projects</li></ul><p>Combined with optimized machine learning models, the Raspberry Pi becomes a capable edge AI device that can perform object detection without relying entirely on cloud computing.</p><h3>Understanding the Workflow</h3><p>A typical object detection pipeline follows several stages.</p><p>First, the Raspberry Pi captures frames using a connected camera. Each frame is then processed by a machine learning model trained to recognise various categories of objects.</p><p>The AI model analyses the image and predicts:</p><ul><li>Which objects are present</li><li>Their confidence scores</li><li>Their locations within the image</li></ul><p>Finally, the detected objects are displayed by drawing bounding boxes and labels around them. This entire process repeats continuously, creating real-time object detection.</p><p>Although the concept sounds complex, modern AI frameworks make implementation surprisingly approachable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*HjK-jeJGf-rXaDC2UKkLQg.jpeg" /></figure><h3>The Role of Machine Learning Models</h3><p>Behind every object detection application is a trained neural network.</p><p>Rather than manually programming rules for identifying every object, developers train deep learning models using thousands -or even millions- of labelled images. During training, the model learns patterns such as shapes, colours, textures, and spatial relationships.</p><p>Once trained, the model can recognise similar objects in new images with impressive accuracy.</p><p>Many modern object detection models are optimised specifically for edge devices, making them suitable for hardware with limited computational resources.</p><h3>Why Cloud-Assisted AI Is Becoming Popular</h3><p>Running complex AI models entirely on embedded hardware can sometimes be challenging, especially when balancing speed, memory usage, and power consumption.</p><p>Cloud-assisted AI provides another option.</p><p>In this approach, the edge device captures images while cloud services perform the computationally intensive inference. This enables developers to use more powerful models without requiring high-end embedded hardware.</p><p>The result is often:</p><ul><li>Improved accuracy</li><li>Faster development</li><li>Easier model updates</li><li>Reduced maintenance</li><li>Better scalability</li></ul><p>Depending on the application, developers may choose local inference, cloud inference, or a hybrid approach.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*_Tx2eadYlCUKSGakDR4VUg.jpeg" /></figure><h3>Typical Hardware Requirements</h3><p>Building an object detection system does not require expensive equipment.</p><p>A basic setup generally includes:</p><ul><li>Raspberry Pi</li><li>Raspberry Pi Camera Module or USB camera</li><li>MicroSD card</li><li>Stable power supply</li><li>Internet connection (for cloud-enabled projects)</li><li>Display or remote SSH access</li></ul><p>With these components, you can begin experimenting with real-time computer vision.</p><h3>Software Stack</h3><p>Python remains the preferred language for Raspberry Pi AI projects because of its simplicity and extensive ecosystem.</p><p>A typical software stack includes:</p><ul><li>Raspberry Pi OS</li><li>Python</li><li>OpenCV</li><li>Camera libraries</li><li>AI inference framework</li><li>Supporting Python packages</li></ul><p>These tools simplify image capture, preprocessing, inference, and visualisation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*ozM32CgfAOmJIJiLtWO0cA.jpeg" /></figure><h3>Common Challenges</h3><p>While getting started is relatively easy, several practical challenges often appear during development.</p><h3>Performance</h3><p>Real-time processing demands significant computational resources. Choosing an optimised model helps improve frame rates.</p><h3>Lighting Conditions</h3><p>Poor lighting can reduce detection accuracy. Consistent illumination generally produces better results.</p><h3>Camera Quality</h3><p>Higher-quality cameras provide clearer images, allowing the model to detect objects more reliably.</p><h3>Internet Connectivity</h3><p>Cloud-based systems rely on stable network connections. Slow or inconsistent internet access can increase latency.</p><h3>Resource Management</h3><p>Monitoring CPU usage, memory consumption, and thermal performance is important for maintaining stable operation during extended runtimes.</p><h3>Real-World Applications</h3><p>Object detection has moved well beyond research laboratories.</p><p>Here are a few practical examples where Raspberry Pi-based vision systems can make a difference.</p><h3>Smart Security</h3><p>Detect people entering restricted areas and trigger alerts automatically.</p><h3>Home Automation</h3><p>Recognize specific objects or people and activate connected devices accordingly.</p><h3>Industrial Automation</h3><p>Monitor production lines for missing components or manufacturing defects.</p><h3>Agriculture</h3><p>Identify crops, monitor livestock, and detect plant diseases using camera-based AI systems.</p><h3>Educational Projects</h3><p>Students can gain hands-on experience with artificial intelligence, embedded systems, and computer vision through practical experimentation.</p><h3>Tips for Better Results</h3><p>If you’re building your own object detection project, keep these best practices in mind.</p><p>Use adequate lighting whenever possible.</p><p>Keep the camera stable to minimize motion blur.</p><p>Choose models that balance speed and accuracy according to your hardware.</p><p>Test the system using different backgrounds and object sizes.</p><p>Measure inference speed in addition to detection accuracy, especially for real-time applications.</p><h3>Where to Go Next</h3><p>Object detection is only the beginning of edge AI.</p><p>Once you’re comfortable with the basics, you can explore more advanced topics such as:</p><ul><li>Face recognition</li><li>Human pose estimation</li><li>Gesture recognition</li><li>Image segmentation</li><li>License plate recognition</li><li>Object tracking</li><li>OCR and text detection</li><li>AI-powered robotics</li></ul><p>Each project builds upon the same computer vision fundamentals while introducing more advanced machine learning concepts.</p><h3>Final Thoughts</h3><p>The Raspberry Pi has made artificial intelligence accessible to developers, students, and hobbyists around the world. By combining affordable hardware with modern computer vision techniques, it’s possible to create intelligent systems capable of understanding their surroundings in real time.</p><p>Whether you’re experimenting with AI for the first time or expanding an existing embedded project, object detection offers an excellent introduction to practical machine learning.</p><p>Explore our collection of <a href="https://circuitdigest.com/internet-of-things-iot-projects"><strong>IoT Projects</strong></a><strong> and Tutorials</strong> to build innovative connected devices, then take your skills further with practical <a href="https://circuitdigest.com/ai-projects-and-tutorials"><strong>AI Projects</strong></a><strong> and Tutorials</strong> for real-world intelligent applications.</p><p>If you’d like to build the complete project yourself, including hardware connections, software setup, installation steps, source code, screenshots, and detailed implementation, visit the full tutorial on Circuit Digest. You’ll find everything you need to recreate the project from start to finish.</p><p>If you enjoyed this article, follow me on Medium for more tutorials on <a href="https://circuitdigest.com/simple-raspberry-pi-projects-for-beginners">Raspberry Pi projects</a>, ESP32, IoT, Embedded Systems, Robotics, AI, and <a href="https://circuitdigest.com/electronics-projects">Electronics projects</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7c81e40d48ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build Raspberry Pi Face Detection Without ML Training]]></title>
            <link>https://circuitdigest.medium.com/build-raspberry-pi-face-detection-without-ml-training-17a131a5e99c?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/17a131a5e99c</guid>
            <category><![CDATA[raspberry-pi]]></category>
            <category><![CDATA[ai-projects]]></category>
            <category><![CDATA[iot-projects]]></category>
            <category><![CDATA[facedetection]]></category>
            <category><![CDATA[circuitdigest]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Tue, 30 Jun 2026 08:29:32 GMT</pubDate>
            <atom:updated>2026-06-30T08:29:32.870Z</atom:updated>
            <content:encoded><![CDATA[<p>Create a cloud-powered face detection system using Raspberry Pi, Python, and OpenCV.</p><p>Face detection is one of the most common applications of computer vision. It powers security systems, smart surveillance, attendance monitoring, visitor counting, and many other intelligent applications. Traditionally, building such a system required collecting thousands of images, training machine learning models, and optimizing them to run efficiently on embedded hardware.</p><p>Fortunately, that is no longer the only option.</p><p>In this project, I built a real-time face detection system using a Raspberry Pi, a USB camera, Python, and OpenCV. Instead of performing AI inference locally, every captured image is securely sent to the CircuitDigest Cloud Face Detection API, where the detection is performed. The Raspberry Pi simply captures the image, uploads it, and displays the detection results.</p><p>This cloud-based approach eliminates the need for machine learning expertise while making computer vision projects accessible to beginners, students, educators, and rapid prototype developers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*FnN6OL6jxDc7ZojAhKCbXQ.jpeg" /></figure><h3>What this project does</h3><p>The Raspberry Pi continuously receives images from a USB camera. Each image is compressed into JPEG format and transmitted to the cloud through an HTTPS request. Once processed, the cloud service returns information such as:</p><ul><li>Number of faces detected</li><li>Bounding box coordinates</li><li>Confidence score for each detected face</li></ul><p>The Raspberry Pi then prints these results in real time, allowing the entire system to work without downloading AI models or performing inference locally.</p><h3>Why use cloud-based face detection?</h3><p>Running deep learning models directly on embedded hardware can consume significant processing power and memory. Cloud inference moves the computational workload away from the Raspberry Pi, making the project easier to deploy and maintain.</p><p>Some of the biggest advantages include:</p><ul><li>No model training</li><li>No dataset collection</li><li>No image labeling</li><li>Minimal Raspberry Pi resource usage</li><li>Simple Python implementation</li><li>Quick project setup</li><li>Easy integration into larger IoT systems</li></ul><p>This allows developers to focus on building applications rather than spending weeks training AI models.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*Cyscg6I8rMTLTasH-IsM0w.jpeg" /></figure><h3>Hardware required</h3><p>The hardware setup is intentionally minimal:</p><ul><li>Raspberry Pi</li><li>USB camera</li><li>MicroSD card with Raspberry Pi OS</li><li>Internet connection</li></ul><p>Since all AI processing happens remotely, no dedicated AI accelerator or GPU is required.</p><h3>Software stack</h3><p>The application is built entirely with Python.</p><p>OpenCV handles camera access and image capture, while the Requests library sends captured images securely to the cloud API. The Raspberry Pi processes the returned JSON response and displays the face detection results.</p><p>Because the project relies on standard Python libraries, it is easy to customise and extend for other computer vision applications.</p><h3>How the system works</h3><p>The workflow is straightforward:</p><ol><li>Capture an image from the USB camera.</li><li>Convert the frame into JPEG format.</li><li>Send the image to the cloud API using HTTPS.</li><li>Wait for the AI service to analyse the image.</li><li>Receive the face detection results.</li><li>Display the number of detected faces and confidence values.</li></ol><p>Since the Raspberry Pi only handles image acquisition and communication, even older Raspberry Pi models can participate in <a href="https://circuitdigest.com/ai-projects-and-tutorials"><strong>AI projects</strong>.</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*HutsrkWGKc1VbOdWbiKb-A.jpeg" /></figure><h3>Multiple operating modes</h3><p>One feature I particularly like is the flexibility in how the project can be operated.</p><p>The application supports manual image capture through the keyboard, automatic image capture at fixed intervals, and terminal-based execution over SSH for completely headless deployments. This makes it suitable for desktop testing as well as unattended monitoring systems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*f6nDFRClJ6b93cuN8g8eWw.jpeg" /></figure><h3>Practical applications</h3><p>Although this project serves as an educational example, it can easily become the foundation for more advanced systems.</p><p>Possible use cases include:</p><ul><li>Smart surveillance</li><li>Visitor counting</li><li>Classroom attendance</li><li>Building access monitoring</li><li>Occupancy estimation</li><li>Retail analytics</li><li>Home automation</li><li>Embedded AI demonstrations</li></ul><p>Because detection is handled in the cloud, developers can rapidly prototype new ideas without worrying about model optimization.</p><h3>Things to keep in mind</h3><p>Cloud inference also introduces a few practical considerations.</p><p>A stable internet connection is required because every image is uploaded before processing. Detection speed depends on both network latency and camera quality. Poor lighting conditions may also reduce detection accuracy, just as they do with most computer vision systems.</p><h3>Why this project is beginner-friendly</h3><p>Many Raspberry Pi AI tutorials become complicated because they involve TensorFlow, model conversion, dataset preparation, or hardware acceleration.</p><p>This project removes those barriers.</p><p>If you’re comfortable writing basic Python code and connecting a USB camera, you can build a working face detection system in a relatively short amount of time. That makes it an excellent introduction to cloud-based computer vision and modern IoT development.</p><h3>Final thoughts</h3><p>Cloud-powered computer vision is making embedded AI significantly more accessible. Instead of spending days configuring machine learning frameworks, developers can focus on building useful products and experimenting with new ideas.</p><p>If you’re looking to build connected devices and smart automation systems, explore these <a href="https://circuitdigest.com/internet-of-things-iot-projects">IoT projects</a> with complete tutorials. They cover everything from beginner-friendly builds to advanced Internet of Things applications.</p><p>If you’re interested in creating AI-powered <a href="https://circuitdigest.com/simple-raspberry-pi-projects-for-beginners"><strong>Raspberry Pi projects</strong> </a>without the complexity of model training, this project offers a practical starting point.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=17a131a5e99c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build Modern ESP32 Interfaces with Arduino and LVGL]]></title>
            <link>https://circuitdigest.medium.com/build-modern-esp32-interfaces-with-arduino-and-lvgl-2750d24edfb8?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/2750d24edfb8</guid>
            <category><![CDATA[lvgl]]></category>
            <category><![CDATA[circuit]]></category>
            <category><![CDATA[diy]]></category>
            <category><![CDATA[electronics]]></category>
            <category><![CDATA[arduino-projects]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Sat, 27 Jun 2026 08:30:54 GMT</pubDate>
            <atom:updated>2026-06-27T08:30:54.303Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Create beautiful embedded GUIs with Arduino IDE and LVGL on an ESP32 display, no complex frameworks required.</strong></p><p>Modern embedded devices are no longer limited to plain text or static graphics. From smart home dashboards and wearable devices to industrial control panels, polished graphical user interfaces (GUIs) have become an essential part of embedded product design.</p><p>Fortunately, building these interfaces doesn’t require writing thousands of lines of graphics code anymore. With <strong>LVGL (Light and Versatile Graphics Library)</strong> and the familiar <strong>Arduino IDE</strong>, you can create responsive, animated, and professional-looking user interfaces that run smoothly on ESP32-based development boards.</p><p>In this article, you’ll learn how <a href="https://circuitdigest.com/tutorial/getting-started-with-arduino-lvgl"><strong>Arduino LVGL</strong></a> works, how to set up your development environment, and how to build your first interactive interface using an ESP32 display.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*VUp34MstTJsVw8flkbGk9w.jpeg" /></figure><h3>Why Use LVGL?</h3><p>LVGL is an open-source graphics library designed specifically for embedded systems. It provides ready-made UI components such as buttons, sliders, switches, labels, charts, progress bars, and animations, making it much easier to create modern interfaces.</p><p>Instead of manually drawing every pixel on the screen, developers work with reusable graphical objects that are easier to customize and maintain.</p><p>Some advantages of LVGL include:</p><ul><li>Rich collection of built-in widgets</li><li>Smooth animations and transitions</li><li>Touchscreen support</li><li>Broad compatibility with microcontrollers</li><li>Open-source MIT license</li><li>Support for visual design tools such as SquareLine Studio</li></ul><p>Whether you’re building a hobby project or developing a commercial embedded product, LVGL significantly reduces development time while improving the overall user experience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*x-WW4mH_xhPyMF4TOWI4Hg.jpeg" /></figure><h3>Hardware Used</h3><p>The demonstration uses an ESP32-C3 development board equipped with a 1.28-inch round LCD display and capacitive touch functionality. The display communicates through SPI, making it suitable for responsive graphical applications.</p><p>Although this board is used throughout the project, the same workflow applies to many other ESP32 development boards. The primary difference is updating the display driver and GPIO pin configuration to match your hardware.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*cYrvXPdaAMquC-MqGoHQTw.jpeg" /></figure><h3>Software Requirements</h3><p>Before getting started, install the following tools:</p><ul><li>Arduino IDE 2.x</li><li>ESP32 Board Package</li><li>LVGL library from the Arduino Library Manager</li><li>Display and touch controller drivers for your hardware</li></ul><p>Having a basic understanding of Arduino programming, SPI communication, and GPIO configuration will make the setup process much easier.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*y487xclqZbEaMHMcuX0v9w.jpeg" /></figure><h3>Setting Up LVGL in Arduino IDE</h3><p>Installing LVGL is straightforward.</p><p>Open the Arduino Library Manager, search for <strong>LVGL</strong>, and install the latest stable version. Once installed, your project gains access to an extensive collection of graphical widgets and interface components.</p><p>After installation, you’ll also need the configuration file and display drivers that allow LVGL to communicate with your specific display hardware.</p><p>This separation between graphics and hardware is one of LVGL’s biggest strengths because it allows the same application code to run on different displays with only minimal hardware-specific changes.</p><h3>Creating Your First GUI</h3><p>One of the easiest ways to learn LVGL is by modifying its official widget examples.</p><p>The project demonstrates this approach by starting with a simple button example from the LVGL documentation. Instead of building an interface from scratch, you copy the example code into your Arduino sketch and integrate it with your display initialization code.</p><p>This method allows you to quickly understand how LVGL creates objects, handles events, and updates the screen.</p><p>Once the example runs successfully, you can begin experimenting by changing:</p><ul><li>Button styles</li><li>Colors</li><li>Fonts</li><li>Labels</li><li>Screen backgrounds</li><li>Touch interactions</li></ul><p>Small modifications quickly help you understand the overall structure of an LVGL application.</p><h3>Understanding the Project Structure</h3><p>The project consists of a few important files.</p><p>The main Arduino sketch initializes the ESP32, configures the display, and starts LVGL.</p><p>A separate LVGL configuration file defines memory usage and enables or disables library features.</p><p>Additional display and touchscreen driver files handle communication with the LCD controller and touch controller.</p><p>Keeping these responsibilities separated makes projects easier to manage as they grow in complexity.</p><h3>Uploading the Program</h3><p>After integrating the widget example into your sketch:</p><ul><li>Select the correct ESP32 board.</li><li>Choose the appropriate serial port.</li><li>Compile the project.</li><li>Upload it to the development board.</li></ul><p>If everything is configured correctly, the display will show the same interactive button demonstrated in the LVGL documentation.</p><p>Touching the button immediately triggers the programmed event, demonstrating how easily interactive embedded interfaces can be created.</p><h3>How the Code Works</h3><p>Although the complete source code contains several hardware-specific sections, its overall structure is easy to understand.</p><p>The program first loads the required libraries and initializes the display hardware.</p><p>Next, GPIO pins for the display and touchscreen are configured.</p><p>After the display driver is initialized, LVGL creates a display buffer that stores graphical updates before sending them to the LCD.</p><p>The application then creates graphical objects such as buttons and labels.</p><p>Finally, the main Arduino loop continuously calls the LVGL task handler, allowing animations, touch events, and screen updates to run smoothly.</p><p>This event-driven architecture is one of the reasons LVGL scales well from simple interfaces to large embedded applications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*aRfu2dXVnlJUFIk4pseFew.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*62EZT8GaKpMWJAxYsGT04Q.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*T6WkkpwmRigfRkmSro8X5g.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*vqcXyQlXKUH0s5cnVTt1yA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*WezdeZxnkmjAXZBotdzlFA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*HrLHS5XxEDIqXl7dWvLXSw.jpeg" /></figure><h3>Common Issues During Setup</h3><p>Many beginners encounter similar problems when working with LVGL for the first time.</p><p>A blank screen usually indicates incorrect SPI wiring or a disabled display backlight.</p><p>If touch input doesn’t respond, the touchscreen driver or I²C pin configuration may be incorrect.</p><p>Compilation errors involving the LVGL configuration file often occur because the required configuration file hasn’t been copied into the project.</p><p>Incorrect display colors or mirrored touch coordinates are typically caused by mismatched display settings that can be corrected through the display configuration.</p><p>Fortunately, most of these issues are simple to resolve once the hardware configuration is verified.</p><h3>Why Choose LVGL Instead of Manual Graphics?</h3><p>Traditional graphics libraries require developers to manually draw every line, circle, and text element.</p><p>LVGL abstracts these low-level drawing operations into reusable interface objects.</p><p>Instead of worrying about pixels, developers focus on application logic and user interaction.</p><p>This makes projects easier to maintain while producing interfaces that feel much closer to those found on modern consumer electronics.</p><p>Animations, touch handling, themes, and reusable widgets come built in, allowing developers to spend more time building features rather than graphics infrastructure.</p><h3>Where to Go Next</h3><p>After creating your first button-based interface, you can gradually explore more advanced LVGL features.</p><p>Popular next projects include:</p><ul><li>Smart home control panels</li><li>Weather dashboards</li><li>Wearable interfaces</li><li>IoT monitoring systems</li><li>Industrial HMI displays</li><li>Media controllers</li><li>Sensor dashboards</li></ul><p>As your confidence grows, you can also experiment with visual UI designers like SquareLine Studio to create complete interfaces with drag-and-drop tools before exporting them into Arduino projects.</p><h3>Final Thoughts</h3><p>LVGL has become one of the most powerful graphics libraries available for embedded systems. Combined with the simplicity of Arduino and the performance of the ESP32, it provides an excellent platform for building responsive and visually appealing graphical interfaces.</p><p>Rather than manually drawing graphics, developers can focus on designing intuitive user experiences while LVGL handles rendering, animations, and event management behind the scenes.</p><p>If you’re looking to move beyond basic OLED text displays and create professional embedded interfaces, learning LVGL is an excellent next step.</p><p>.Discover more hands-on embedded systems tutorials and innovative DIY electronics projects at Circuit Digest’s <a href="https://circuitdigest.com/arduino-projects">Arduino Projects</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2750d24edfb8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ESP32 AI Speech-to-Text Made Simple]]></title>
            <link>https://circuitdigest.medium.com/esp32-ai-speech-to-text-made-simple-0b4f7eb930f9?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/0b4f7eb930f9</guid>
            <category><![CDATA[diy-projects]]></category>
            <category><![CDATA[electronics]]></category>
            <category><![CDATA[esp32]]></category>
            <category><![CDATA[diy]]></category>
            <category><![CDATA[text-to-speech]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Sat, 27 Jun 2026 08:22:17 GMT</pubDate>
            <atom:updated>2026-06-27T08:22:17.287Z</atom:updated>
            <content:encoded><![CDATA[<p>Learn how to build an AI-powered speech recognition system using ESP32 and convert voice into text in real time.</p><p>Voice-controlled devices have become an essential part of modern technology. From smart assistants to home automation systems, the ability to convert spoken words into text opens the door to countless innovative applications. While speech recognition once required powerful computers, today’s cloud-based AI services make it possible to build capable voice interfaces using affordable microcontrollers like the ESP32.</p><p>In this project, we’ll build a real-time speech-to-text system using the ESP32 development board, an I2S MEMS microphone, and an AI speech recognition service. The ESP32 captures voice through the digital microphone, sends the audio to a cloud-based speech recognition engine, and receives the transcribed text almost instantly.</p><p>Whether you’re developing a voice-controlled IoT device, experimenting with artificial intelligence, or simply learning how cloud APIs work with embedded systems, this project offers an excellent starting point.</p><h3>Why Use the ESP32 for Speech Recognition?</h3><p>The ESP32 is one of the most popular microcontrollers for IoT development because it combines Wi-Fi, Bluetooth, and impressive processing power at a low cost. Although it cannot perform large-scale speech recognition locally, it excels at collecting audio data and securely sending it to cloud AI services.</p><p>By combining the ESP32 with an external digital microphone and an AI speech recognition platform, you can create responsive voice-enabled applications without requiring expensive hardware.</p><p>Some advantages of this approach include:</p><ul><li>Low-cost hardware</li><li>Built-in Wi-Fi connectivity</li><li>Real-time speech recognition</li><li>Easy integration with cloud AI services</li><li>Scalable for future voice-controlled projects</li></ul><h3>Project Overview</h3><p>The system works in a straightforward sequence.</p><p>First, the ESP32 records audio through an I2S MEMS microphone. The captured audio samples are formatted correctly and transmitted over Wi-Fi to an AI speech recognition service. After processing the voice input, the cloud returns the recognized text, which is then displayed through the Serial Monitor or can be used to trigger actions within your application.</p><p>This architecture allows the ESP32 to focus on audio acquisition while leveraging the computational power of cloud-based machine learning models for accurate transcription.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*eSt8sjW_Slz5LA_ngqFyEQ.jpeg" /></figure><h3>Hardware Required</h3><p>To build this project, you’ll need:</p><ul><li>ESP32 development board</li><li>I2S MEMS microphone module</li><li>USB cable</li><li>Breadboard</li><li>Jumper wires</li><li>Stable Wi-Fi connection</li></ul><p>The simplicity of the hardware makes this project accessible for beginners while remaining powerful enough for advanced experimentation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*VXm5LYJZNTDf5EmCX4CpaQ.jpeg" /></figure><h3>Understanding the I2S Microphone</h3><p>Unlike traditional analog microphones, an I2S MEMS microphone outputs digital audio directly. This eliminates many of the noise and signal conditioning issues associated with analog circuits.</p><p>The ESP32 includes dedicated I2S peripherals, making it well suited for capturing high-quality digital audio. The microphone continuously streams sampled audio, which the ESP32 packages and forwards to the AI service for processing.</p><p>Because the audio is already digital, the overall design becomes cleaner, more reliable, and easier to implement.</p><h3>Setting Up the Development Environment</h3><p>Begin by installing the Arduino IDE or your preferred ESP32 development environment.</p><p>Next, install the ESP32 board package and any required libraries for Wi-Fi communication, HTTP requests, JSON parsing, and I2S audio support.</p><p>Create an account with your preferred speech recognition service and obtain the required API credentials or access token. These credentials allow the ESP32 to securely authenticate while sending recorded audio for transcription.</p><p>Finally, update the program with your Wi-Fi credentials and API information before uploading it to the ESP32.</p><h3>How the Software Works</h3><p>Once powered on, the ESP32 connects to the configured Wi-Fi network.</p><p>After establishing the connection, it initializes the I2S microphone and begins recording audio samples.</p><p>The captured audio is converted into the required format and transmitted to the cloud speech recognition API using an HTTP request. Once processing is complete, the server returns the recognized text.</p><p>The ESP32 receives the response, extracts the transcription, and displays it through the Serial Monitor. From here, the recognized words can also trigger automation routines, control hardware, or interact with other IoT devices.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*Y98ELJ37hdZ_sea9J1WrUg.jpeg" /></figure><h3>Improving Speech Recognition Accuracy</h3><p>Several factors influence transcription quality.</p><p>A quiet recording environment significantly improves recognition accuracy. Using a quality MEMS microphone with proper power supply filtering also reduces unwanted background noise.</p><p>Speaking clearly at a moderate pace helps the AI engine distinguish individual words more effectively. Maintaining a stable Wi-Fi connection further minimizes transmission delays and ensures faster responses.</p><p>Fine-tuning microphone placement and gain settings can also improve overall performance.</p><h3>Practical Applications</h3><p>Speech-to-text technology enables a wide variety of embedded AI applications.</p><p>Some examples include:</p><ul><li>Voice-controlled home automation</li><li>Smart IoT appliances</li><li>Hands-free industrial controls</li><li>Accessibility devices</li><li>Voice-enabled robotics</li><li>Smart classroom projects</li><li>AI-powered assistants</li><li>Custom voice command systems</li></ul><p>These applications demonstrate how speech recognition can make electronic projects more interactive and user-friendly.</p><h3>Challenges to Consider</h3><p>Since speech recognition is processed in the cloud, an active internet connection is required during operation.</p><p>Network latency can introduce slight delays depending on connection quality. Background noise may also affect transcription accuracy if the recording environment is not ideal.</p><p>Despite these limitations, cloud-based speech recognition generally provides significantly higher accuracy than lightweight offline alternatives running directly on microcontrollers.</p><h3>Expanding the Project</h3><p>After successfully converting speech into text, numerous enhancements become possible.</p><p>You can build a complete voice assistant, integrate home automation platforms, control relays using spoken commands, connect to ChatGPT or other large language models, or combine speech recognition with text-to-speech systems for full conversational interfaces.</p><p>The ESP32 serves as an excellent foundation for exploring embedded artificial intelligence and voice-driven IoT applications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*kfUWfA-vuG-OuDRVe82gxg.jpeg" /></figure><h3>Final Thoughts</h3><p>Building an <a href="https://circuitdigest.com/microcontroller-projects/esp32-speech-to-text-using-ai"><strong>ESP32 Speech-to-Text</strong></a><strong> </strong>using AI<strong> </strong>demonstrates how affordable hardware can leverage cloud intelligence to create sophisticated voice-enabled applications. By combining an ESP32, an I2S MEMS microphone, and an AI speech recognition service, developers can add natural voice interaction to a wide range of projects without requiring high-end computing hardware.</p><p>These <a href="https://circuitdigest.com/esp32-projects"><strong>ESP32 projects</strong></a> are an excellent introduction to embedded AI, cloud APIs, and IoT communication, making them valuable for beginners and experienced developers alike.</p><p>If you’d like to see the complete circuit diagram, source code, wiring details, and step-by-step implementation, visit the full project guide on CircuitDigest.</p><p>Explore more hands-on <a href="https://circuitdigest.com/ai-projects-and-tutorials">AI Projects</a> and embedded machine learning projects.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FjG_9HgEOipE%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DjG_9HgEOipE&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FjG_9HgEOipE%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/4e0ed64744b29988e67173b89c7cd389/href">https://medium.com/media/4e0ed64744b29988e67173b89c7cd389/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0b4f7eb930f9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ESP32-CAM Object Detection Using Cloud AI]]></title>
            <link>https://circuitdigest.medium.com/esp32-cam-object-detection-using-cloud-ai-d5ed0b0b240f?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/d5ed0b0b240f</guid>
            <category><![CDATA[diy]]></category>
            <category><![CDATA[electronics]]></category>
            <category><![CDATA[esp32-cam]]></category>
            <category><![CDATA[esp32]]></category>
            <category><![CDATA[diy-projects]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Thu, 25 Jun 2026 13:38:15 GMT</pubDate>
            <atom:updated>2026-06-25T13:38:15.917Z</atom:updated>
            <content:encoded><![CDATA[<p>Build a Real-Time Object Detection System Without Training Any AI Model</p><p>Artificial Intelligence and computer vision are no longer limited to powerful computers and expensive hardware. Today, even a compact and affordable ESP32-CAM module can identify everyday objects such as laptops, phones, books, cups, and more with the help of cloud-based AI services.</p><p>In this <a href="https://circuitdigest.com/microcontroller-projects/esp32-cam-attendance-system-using-circuitdigest-cloud"><strong>ESP32-CAM object detection</strong> </a>project, we’ll build a real-time object detection system using an ESP32-CAM and the CircuitDigest Cloud Object Detection API. Unlike traditional machine learning workflows, this approach does not require collecting datasets, labeling images, training models, or deploying neural networks on the device. The cloud handles all the heavy processing, allowing the ESP32-CAM to focus on image capture and communication.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*csLVjhKylQgNqlfXdW-O8w.jpeg" /></figure><h3>Why Use Cloud-Based Object Detection?</h3><p>Object detection has traditionally been challenging on microcontrollers because image processing and neural network inference require significant computational resources. While edge AI platforms can run lightweight models on embedded devices, they often involve a lengthy development process that includes data collection, annotation, training, validation, and deployment.</p><p>Cloud-based object detection simplifies this workflow. The ESP32-CAM captures an image and uploads it to a cloud service, which analyzes the image and returns information about the detected objects, their count, and confidence scores. This approach significantly reduces development time and lowers the hardware requirements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*PN4-NiAPJldSFY95nZ62jw.jpeg" /></figure><h3>How the System Works</h3><p>The workflow is straightforward:</p><ol><li>A push button connected to the ESP32-CAM acts as the trigger.</li><li>When the button is pressed, the camera captures an image.</li><li>The image is transmitted over Wi-Fi to the cloud-based object detection API.</li><li>The cloud service analyzes the image using pre-trained AI models.</li><li>Detection results are returned to the ESP32-CAM.</li><li>The identified objects, confidence scores, and object counts are displayed in the Serial Monitor.</li></ol><p>This architecture allows developers to add AI-powered vision capabilities to embedded projects without worrying about model development or inference optimization.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*LiJfPsEC0DxtQJXkOCPxRA.jpeg" /></figure><h3>Hardware Required</h3><p>The project uses only a few components:</p><ul><li>ESP32-CAM module</li><li>Push button</li><li>Breadboard</li><li>Jumper wires</li><li>USB-to-Serial adapter (for ESP32-CAM boards without onboard USB programming support)</li></ul><p>The simplicity of the hardware makes this project suitable for beginners who want to explore AI-powered <a href="https://circuitdigest.com/internet-of-things-iot-projects">IoT projects</a>.</p><h3>Hardware Connections</h3><p>The push button is connected to GPIO13 of the ESP32-CAM. Pressing the button initiates image capture and triggers the object detection process.</p><p>The ESP32-CAM handles both image acquisition and Wi-Fi communication, making the overall circuit compact and easy to assemble.</p><h3>Setting Up the Cloud API</h3><p>Before uploading the code, you’ll need access to the object detection API.</p><p>The setup process involves:</p><ul><li>Creating an account on the cloud platform</li><li>Generating an API key</li><li>Selecting the object classes you want to detect</li><li>Configuring a confidence threshold</li><li>Testing the API with sample images</li></ul><p>A confidence threshold helps control detection accuracy. Lower values detect more objects but may introduce false positives, while higher values provide more reliable results but can miss certain objects.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*P6jeauMUbskP2vLgvoKtrQ.jpeg" /></figure><h3>Programming the ESP32-CAM</h3><p>The firmware is written using the Arduino framework.</p><p>The main responsibilities of the code include:</p><h3>Connecting to Wi-Fi</h3><p>The ESP32-CAM first connects to a wireless network so it can communicate with the cloud service.</p><h3>Initializing the Camera</h3><p>The OV2640 camera sensor is configured for JPEG image capture. Camera settings such as brightness, contrast, resolution, and image quality can be adjusted to improve recognition performance.</p><h3>Capturing Images</h3><p>When the button is pressed, the camera captures a frame and stores it in memory.</p><h3>Sending Data to the Cloud</h3><p>The captured image is transmitted securely using HTTPS requests. The API key authenticates the request and allows the cloud service to process the image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*74GpHFvQy2tNuw9aWg2ITg.jpeg" /></figure><h3>Receiving Detection Results</h3><p>After processing, the cloud service returns:</p><ul><li>Object names</li><li>Number of detected objects</li><li>Confidence scores</li></ul><p>These results are displayed through the Arduino Serial Monitor for easy debugging and analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*B899v_86gEPmBkXz3LE_EQ.jpeg" /></figure><h3>Improving Detection Accuracy</h3><p>The quality of the captured image directly affects detection performance.</p><p>To achieve better results:</p><h3>Ensure Good Lighting</h3><p>Poor lighting often leads to blurry or noisy images, reducing detection accuracy.</p><h3>Use Appropriate Camera Settings</h3><p>Adjusting resolution and image quality can significantly improve recognition performance.</p><h3>Tune the Confidence Threshold</h3><p>Experiment with different confidence values until you achieve a balance between accuracy and sensitivity.</p><h3>Position the Camera Correctly</h3><p>Keeping the target object centred and clearly visible improves recognition reliability.</p><h3>Common Challenges and Troubleshooting</h3><h3>Camera Initialization Errors</h3><p>These issues are often caused by incorrect wiring, insufficient power, or camera module connection problems.</p><h3>Wi-Fi Connectivity Problems</h3><p>Ensure that the network credentials are correct and that the ESP32-CAM receives a strong Wi-Fi signal.</p><h3>Blurry Images</h3><p>Adjust focus, improve lighting, and verify camera settings.</p><h3>No Objects Detected</h3><p>Check whether the selected object classes match the objects present in the image and verify the confidence threshold settings.</p><h3>Cloud Processing vs Edge AI</h3><p>One interesting aspect of this project is the decision to perform object detection in the cloud rather than directly on the ESP32-CAM.</p><p>Cloud processing offers several advantages:</p><ul><li>No model training required</li><li>Faster project development</li><li>Access to more powerful AI models</li><li>Better detection accuracy</li></ul><p>However, it also introduces some limitations:</p><ul><li>Requires internet connectivity</li><li>Dependent on API availability</li><li>Potential usage limits</li><li>Slight latency due to image transmission</li></ul><p>For many IoT applications, this trade-off is acceptable because it dramatically simplifies development while delivering reliable results. Community discussions around ESP32 vision projects frequently highlight these hardware limitations and the benefits of offloading AI workloads to external systems.</p><h3>Real-World Applications</h3><p>This project can serve as the foundation for many practical systems, including:</p><ul><li>Smart surveillance cameras</li><li>Inventory monitoring systems</li><li>Industrial object counting</li><li>Classroom AI demonstrations</li><li>Home automation triggers</li><li>Retail analytics solutions</li></ul><p>By combining the ESP32-CAM with cloud-based AI services, developers can quickly prototype intelligent vision systems without specialized machine learning expertise.</p><h3>Final Thoughts</h3><p>Object detection is often perceived as a complex field requiring advanced machine learning knowledge and powerful hardware. This project demonstrates that modern cloud AI services can remove much of that complexity.</p><p>With just an ESP32-CAM, a push button, and a Wi-Fi connection, it’s possible to create a capable object detection system that identifies multiple objects in real time. The cloud-based approach eliminates the need for dataset preparation, model training, and optimization, allowing developers to focus on building innovative applications instead of managing AI infrastructure.</p><p>If you’re interested in exploring embedded vision, <a href="https://circuitdigest.com/ai-projects-and-tutorials"><strong>AI projects</strong></a>, IoT, or ESP32-CAM development, this project provides an excellent starting point while keeping the hardware and software requirements refreshingly simple.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FuEJxAupugaE%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DuEJxAupugaE&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FuEJxAupugaE%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/fd67b1021c30b5c42a30a281361e9780/href">https://medium.com/media/fd67b1021c30b5c42a30a281361e9780/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d5ed0b0b240f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Smart Attendance System Using ESP32-CAM and WhatsApp Alerts]]></title>
            <link>https://circuitdigest.medium.com/smart-attendance-system-using-esp32-cam-and-whatsapp-alerts-2473fea8a446?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/2473fea8a446</guid>
            <category><![CDATA[iot-projects]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[esp32-tutorial]]></category>
            <category><![CDATA[diy]]></category>
            <category><![CDATA[diy-projects]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Thu, 25 Jun 2026 13:26:29 GMT</pubDate>
            <atom:updated>2026-06-25T13:26:29.642Z</atom:updated>
            <content:encoded><![CDATA[<h4>Build an IoT-powered attendance tracker with photo verification and real-time notifications</h4><p>Attendance tracking is still surprisingly manual in many schools, coaching centers, hostels, and small offices. Teachers spend valuable time calling names, maintaining registers, and correcting attendance errors. Besides being time-consuming, traditional attendance methods provide little proof that the right person was actually present.</p><p>What if a low-cost IoT device could record attendance, capture a photo as evidence, and instantly send a notification through WhatsApp?</p><p>In this <a href="https://circuitdigest.com/microcontroller-projects/esp32-cam-attendance-system-using-circuitdigest-cloud"><strong>ESP32 Cam attendance system</strong></a> project, we build a smart attendance system using the ESP32-CAM, an OLED display, and a rotary encoder. The system allows users to mark their entry or exit, captures an image, records the timestamp, and sends the attendance information through the cloud in real time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*8cEuJmoAVZcA1g5ChCw3Nw.jpeg" /></figure><h3>Why Build a Smart Attendance System?</h3><p>Most attendance systems fall into one of three categories:</p><ul><li>Manual registers</li><li>RFID card-based systems</li><li>Biometric attendance devices</li></ul><p>While these solutions work, they often have limitations. Manual systems are prone to errors, RFID cards can be shared among users, and biometric systems can become expensive for small deployments.</p><p>The ESP32-CAM offers a practical alternative. By combining image capture, internet connectivity, and cloud-based messaging, it creates a reliable attendance solution at a fraction of the cost of commercial systems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*erRpeR19sG4OqqZqsskITw.jpeg" /></figure><h3>What Makes This Project Different?</h3><p>The most interesting aspect of this project is that every attendance record includes visual verification.</p><p>Whenever a user marks attendance, the device:</p><ul><li>Records the person’s name</li><li>Captures a photograph</li><li>Adds an accurate timestamp</li><li>Identifies whether the event is an entry or exit</li><li>Sends all information through WhatsApp</li></ul><p>This creates a transparent attendance log that is much harder to manipulate compared to traditional methods.</p><h3>Hardware Used</h3><p>The project is built around a few inexpensive components:</p><ul><li>ESP32-CAM module</li><li>OLED display</li><li>Rotary encoder</li><li>Perfboard or prototype board</li><li>Enclosure for mounting</li></ul><p>The ESP32-CAM acts as both the microcontroller and camera system, eliminating the need for additional image-processing hardware. The OLED display provides a simple user interface, while the rotary encoder allows users to navigate menus and make selections.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*JFsgiuhaF1k4mqsfWsylNA.jpeg" /></figure><h3>How the System Works</h3><p>The workflow is intentionally simple so that anyone can use it.</p><p>When the device powers up, it connects to a Wi-Fi network and synchronizes its internal clock using an online time server. Accurate timestamps are critical because every attendance event must be recorded precisely.</p><p>The OLED display then shows a list of registered users.</p><p>Using the rotary encoder, a person scrolls through the list and selects their name. After selecting a name, they choose whether they are entering or leaving the premises.</p><p>Once confirmed, the system starts a short countdown.</p><p>This countdown gives the user enough time to position themselves in front of the camera.</p><p>When the countdown reaches zero, the ESP32-CAM captures an image and prepares the attendance record.</p><p>The system then packages:</p><ul><li>User name</li><li>Attendance status (IN or OUT)</li><li>Timestamp</li><li>Captured photograph</li></ul><p>and sends the information to the cloud service through an API. Moments later, the designated WhatsApp number receives a message containing all attendance details along with the captured image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*X7-YuYzqRnVdDG1y5xrNrQ.jpeg" /></figure><h3>User Interface Design</h3><p>One of the strengths of this project is its simplicity.</p><p>There are no touchscreens or complicated menus.</p><p>The OLED display provides clear prompts, while the rotary encoder serves as the only input device.</p><p>This approach keeps the hardware cost low while making the system easy to operate for students, employees, or visitors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*AdxdJ6o9xtXYCs-a3na2iA.jpeg" /></figure><h3>Software Architecture</h3><p>The firmware running on the ESP32-CAM is divided into several logical sections.</p><h3>Wi-Fi and Cloud Connectivity</h3><p>The device first establishes an internet connection using stored Wi-Fi credentials. This connection enables communication with cloud services and time synchronization.</p><h3>Camera Management</h3><p>The camera is configured to capture compressed JPEG images, reducing memory consumption and improving upload performance.</p><h3>User Selection Logic</h3><p>The rotary encoder is continuously monitored to detect scrolling and button presses. This allows users to navigate through names and attendance options.</p><h3>Display Handling</h3><p>The OLED updates dynamically based on user actions, creating an intuitive menu-driven experience.</p><h3>Data Transmission</h3><p>Once attendance is confirmed, the device uploads the captured image and associated attendance information to the cloud platform, which then delivers the message through WhatsApp.</p><h3>Real-World Applications</h3><p>Although designed as an attendance system, the same concept can be adapted for many other applications.</p><h3>Educational Institutions</h3><p>Schools and coaching centers can automatically record student attendance while notifying parents in real time.</p><h3>Corporate Offices</h3><p>Organizations can track employee entry and exit records without maintaining paper logs.</p><h3>Hostels and Dormitories</h3><p>Wardens can monitor student movement and maintain accurate records.</p><h3>Visitor Management</h3><p>Businesses can create a lightweight visitor logging system with photo-based verification.</p><h3>Small Factories and Workshops</h3><p>Managers can track worker attendance without investing in expensive biometric systems.</p><h3>Challenges and Considerations</h3><p>Like any <a href="https://circuitdigest.com/internet-of-things-iot-projects">IoT projects</a>, there are a few practical considerations.</p><p>The system depends on a stable Wi-Fi connection because timestamps and notifications require internet access.</p><p>Camera placement is also important. Poor lighting conditions may reduce image quality and make visual verification less effective.</p><p>For larger deployments, a cloud database can be added to store attendance records and generate analytics dashboards.</p><h3>Future Improvements</h3><p>This project already demonstrates a powerful combination of IoT and cloud communication, but there is significant room for expansion.</p><p>Some interesting upgrades include:</p><ul><li>Face recognition for automatic identification</li><li>Cloud database integration</li><li>Mobile application support</li><li>GPS-based location tracking</li><li>RFID verification</li><li>Fingerprint authentication</li><li>Attendance analytics dashboard</li></ul><p>These additions can transform the project from a simple attendance logger into a complete workforce or student management platform.</p><h3>Final Thoughts</h3><p>The ESP32-CAM continues to prove that sophisticated IoT solutions no longer require expensive hardware. By combining image capture, wireless connectivity, cloud services, and messaging platforms, it’s possible to create a reliable attendance system that provides both convenience and accountability.</p><p>Whether you’re an electronics enthusiast, an IoT developer, or a student looking for your next embedded systems project, this build demonstrates how modern hardware can solve everyday administrative problems practically and affordably.</p><p>Explore these hands-on <a href="https://circuitdigest.com/esp32-projects"><strong>ESP32 projects</strong></a><strong> </strong>with complete circuit diagrams, source code, and step-by-step tutorials to build innovative IoT, robotics, AI, and automation applications: ESP32 Projects Collection.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FSxHRiSCLeL0%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DSxHRiSCLeL0&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FSxHRiSCLeL0%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/cc3ec40f3a54e7216092cc6ce0d351df/href">https://medium.com/media/cc3ec40f3a54e7216092cc6ce0d351df/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2473fea8a446" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build a Voice-Controlled ESP32 Drone]]></title>
            <link>https://circuitdigest.medium.com/build-a-voice-controlled-esp32-drone-21b2981ff8d5?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/21b2981ff8d5</guid>
            <category><![CDATA[python]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[voice-controlled-device]]></category>
            <category><![CDATA[drone-project]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Thu, 21 May 2026 17:26:56 GMT</pubDate>
            <atom:updated>2026-05-21T17:26:56.025Z</atom:updated>
            <content:encoded><![CDATA[<h3>Control a LiteWing drone using Python, offline speech recognition, and simple voice commands</h3><p>Drones are already exciting to build and fly, but controlling one entirely through voice commands adds a completely different level of interaction. Instead of relying on joysticks and buttons, this project uses speech recognition to let an ESP32-powered drone respond to spoken instructions like “take off,” “forward,” and “land.”</p><p>In this project, I built a voice-controlled drone using the LiteWing ESP32 drone platform, Python, and the Vosk offline speech recognition engine. The system listens to voice commands through a microphone, converts speech into text in real time, and instantly sends commands to the drone over Wi-Fi. The best part is that the entire system works offline without needing cloud APIs or internet connectivity.</p><p>The project is ideal for students, robotics enthusiasts, and embedded developers who want to explore drone programming, human-machine interaction, and offline AI-based control systems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*U_O-4qJVqwygfFOFHjH-Vg.jpeg" /></figure><h3>Why Build a Voice-Controlled Drone?</h3><p>Traditional drone controllers are effective, but they also create a learning curve for beginners. Voice interaction makes drone control feel more natural and intuitive. Instead of memorising controls, users can simply speak commands and interact with the drone conversationally.</p><p>This ESP32 <a href="https://circuitdigest.com/microcontroller-projects/esp32-voice-controlled-drone-using-litewing"><strong>Voice Controlled Drone</strong></a> project also demonstrates how lightweight AI tools can be combined with embedded systems to create responsive real-world applications. It is a practical example of integrating speech recognition, wireless communication, and drone control into a single system.</p><p>Community discussions around ESP32 drones show growing interest in programmable and open-source drone ecosystems because they allow developers to customise everything from flight control to automation features.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*-dbK3GqXDccU2TACTUJ9hg.jpeg" /></figure><h3>Hardware Used</h3><p>The hardware setup is intentionally simple. The LiteWing drone handles all flight operations while a laptop processes voice recognition and sends commands over Wi-Fi.</p><p>The main components include:</p><ul><li>LiteWing ESP32 drone</li><li>LiteWing positioning module</li><li>Laptop or desktop computer</li><li>Microphone for voice input</li><li>LiPo battery for the drone</li></ul><p>The positioning module improves flight stability and helps the drone maintain smoother movement during operation. Correct installation is important because the module orientation affects sensor readings and flight behaviour.</p><h3>Software Stack</h3><p>The software side of the project is where the real magic happens.</p><p>Python acts as the main control layer, coordinating speech recognition and drone communication. The microphone captures audio in real time using PyAudio, while the Vosk speech recognition framework converts speech into text locally on the computer. Once the text is recognised, the LiteWing Python library sends movement commands directly to the drone.</p><p>The entire workflow is fast enough to allow real-time interaction without noticeable lag.</p><p>The major software tools used are:</p><ul><li>Python 3</li><li>Vosk speech recognition</li><li>PyAudio</li><li>LiteWing Python library</li></ul><p>One of the strongest advantages of this setup is that Vosk works completely offline. Since the laptop connects directly to the drone’s Wi-Fi network, internet access is not required.</p><h3>How the System Works</h3><p>The control flow is straightforward.</p><p>The microphone continuously listens for voice input. When the user speaks, the Vosk model processes the audio and converts it into text. The program then scans the text for predefined keywords and maps them to drone actions.</p><p>For example:</p><ul><li>“Take off” arms the drone and start flight</li><li>“Forward” moves the drone ahead</li><li>“Turn left” rotates the drone</li><li>“Land” safely lands the drone</li></ul><p>The commands are transmitted wirelessly to the LiteWing drone using Wi-Fi communication.</p><p>This keyword-based approach keeps the system lightweight, reliable, and easy to customise for additional commands.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*_TXSzT3J_fmfEIbgnEXEhA.jpeg" /></figure><h3>Setting Up the Voice Recognition Engine</h3><p>The core of the project is the Vosk offline speech recognition engine.</p><p>Unlike cloud-based voice assistants, Vosk performs speech processing locally on the computer. This reduces latency and ensures the drone can still operate without internet access.</p><p>The microphone streams audio at 16kHz, which is then passed into the Vosk recogniser. Once a valid phrase is detected, the recognised text is forwarded to the command-processing system.</p><p>The responsiveness of the system makes the interaction feel surprisingly natural during flight tests.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*a-aLYdXclvM6FxexCbcvuQ.jpeg" /></figure><h3>Command Processing Logic</h3><p>The command-processing section acts as the brain of the project.</p><p>The software continuously checks recognised text for specific keywords. Once a matching command is detected, the corresponding drone action is triggered immediately.</p><p>The system supports commands for:</p><ul><li>Takeoff and landing</li><li>Forward and backward movement</li><li>Left and right motion</li><li>Rotation</li><li>Altitude adjustment</li><li>LED color changes</li><li>Emergency stop</li></ul><p>This architecture also makes it easy to add custom voice commands in the future.</p><p>For example, developers could extend the project with:</p><ul><li>Autonomous flight routines</li><li>Gesture recognition</li><li>AI-assisted navigation</li><li>Computer vision tracking</li><li>Voice authentication</li></ul><p>Many developers in the ESP32 drone community are already experimenting with custom firmware, ROS integration, and autonomous navigation systems, showing how flexible these platforms can become.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*lGKrnkNwpdvB0Ii9nr9u0w.jpeg" /></figure><h3>Flight Experience</h3><p>Testing the drone with voice commands feels futuristic at first.</p><p>Speaking simple instructions like “up,” “left,” or “land” creates a far more immersive interaction than using a traditional controller. The system responds quickly enough that controlling the drone through speech becomes surprisingly intuitive after a few minutes.</p><p>LED colour commands also add a fun, interactive element during demonstrations.</p><p>Because speech recognition happens locally, response times remain consistent even without internet connectivity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*5AmETevykSWqQVX0Naajiw.png" /></figure><h3>Challenges Faced During Development</h3><p>Like most drone projects, this build required careful tuning and testing.</p><p>Background noise can sometimes affect speech recognition accuracy, especially in crowded environments. Adding microphone calibration and confidence filtering helped reduce false triggers.</p><p>Flight stability was another important factor. The positioning module significantly improved hover performance and smoother movement control.</p><p>Battery management and safe indoor testing were also essential during development.</p><h3>Future Improvements</h3><p>This project opens the door to several exciting upgrades.</p><p>Future versions could include:</p><ul><li>AI-based obstacle avoidance</li><li>Computer vision tracking</li><li>Voice authentication for security</li><li>Autonomous waypoint navigation</li><li>Smartphone voice integration</li><li>Edge AI processing directly on the ESP32</li></ul><p>It would also be interesting to combine gesture control and voice control into a hybrid interaction system for even more intuitive flight experiences.</p><h3>Final Thoughts</h3><p>Building a voice-controlled<a href="https://circuitdigest.com/microcontroller-projects/DIY-wifi-controlled-drone"> <strong>ESP32 drone</strong></a> is an excellent way to combine embedded systems, wireless communication, and AI-powered interaction into one practical project.</p><p>What makes this build especially interesting is how accessible the technology has become. With Python, offline speech recognition, and an open-source ESP32 drone platform, it is now possible to create advanced interactive systems without expensive hardware or cloud infrastructure.</p><p>Projects like this highlight how embedded AI and human-machine interaction are becoming more approachable for makers, students, and developers alike.</p><p>If you enjoy experimenting with drones, robotics, or embedded AI systems, this project is a great starting point for exploring smarter and more interactive flying machines.</p><p>Build innovative<a href="https://circuitdigest.com/drone-projects"> DIY drone projects</a> with ESP32, Arduino, and Raspberry Pi using step-by-step tutorials for robotics, IoT, and autonomous flight experiments.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F6Y0AwSqIanU%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D6Y0AwSqIanU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F6Y0AwSqIanU%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/6cb3a89ff2cdd54bee1f0575673e7423/href">https://medium.com/media/6cb3a89ff2cdd54bee1f0575673e7423/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=21b2981ff8d5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build an Offline Voice Control System]]></title>
            <link>https://circuitdigest.medium.com/build-an-offline-voice-control-system-0a50f8fbcafe?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/0a50f8fbcafe</guid>
            <category><![CDATA[su-03t]]></category>
            <category><![CDATA[offline-voice-recognition]]></category>
            <category><![CDATA[modules]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Thu, 21 May 2026 10:56:27 GMT</pubDate>
            <atom:updated>2026-05-21T10:56:27.675Z</atom:updated>
            <content:encoded><![CDATA[<h3>A practical, privacy-friendly alternative to the VC-02 voice recognition module</h3><p>Voice-controlled electronics are everywhere today, from smart speakers to industrial automation systems. But most voice assistants rely heavily on cloud processing, which introduces latency, internet dependency, and privacy concerns. For embedded projects and DIY automation, this can become a major limitation.</p><p>This project demonstrates how to build a completely offline voice recognition system using the low-cost SU-03T module — a practical alternative to the popular VC-02. The entire system works without Wi-Fi or cloud APIs, making it ideal for local automation, robotics, assistive devices, and educational projects. The module listens for predefined commands and instantly triggers actions through GPIO pins.</p><p>The best part is that the setup is inexpensive, beginner-friendly, and surprisingly reliable for simple command-based applications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*09O18_B1HgLLXJbvvg5udg.jpeg" /></figure><h3>Why Use Offline Voice Recognition?</h3><p>Cloud-based voice systems send audio data to remote servers for processing. While powerful, they also come with several drawbacks:</p><ul><li>Internet connectivity is mandatory</li><li>Response time depends on network speed</li><li>Privacy becomes a concern</li><li>Reliability drops during outages</li></ul><p>Offline voice recognition solves these issues by processing commands directly on the device itself. The SU-03T module stores trained commands locally and responds immediately when a match is detected. This makes it suitable for low-latency and privacy-focused applications.</p><h3>Why the SU-03T Instead of the VC-02?</h3><p>The VC-02 by Ai-Thinker is widely used because of its official SDK support and documentation. However, the SU-03T offers a much cheaper alternative while still supporting offline command recognition, GPIO triggering, and wake-word functionality.</p><p>For hobbyists, prototypes, and budget projects, the SU-03T provides excellent value without sacrificing core functionality. Many developers have also confirmed that English commands work reliably despite older documentation suggesting otherwise.</p><h3>Hardware Required</h3><p>The complete setup uses only a few basic components:</p><ul><li>SU-03T <a href="https://circuitdigest.com/tutorial/offline-voice-recognition-module-alternatives-to-vc-02"><strong>Offline Voice Recognition Module</strong></a></li><li>Electret microphone</li><li>Small speaker</li><li>USB-to-TTL converter</li><li>Two LEDs</li><li>100Ω resistors</li><li>Breadboard</li><li>Jumper wires</li></ul><p>The microphone captures voice input while the speaker provides audio feedback. LEDs connected to GPIO pins are used to demonstrate output control.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*N5K61fEa62jHNs7JbP_jPw.jpeg" /></figure><h3>How the System Works</h3><p>The workflow is straightforward:</p><ol><li>Voice commands are configured using the Ai-Thinker SDK portal</li><li>Firmware containing those commands is generated</li><li>The firmware is flashed into the SU-03T module</li><li>The module continuously listens for speech</li><li>When a stored command is recognised, the GPIO pins trigger predefined actions</li></ol><p>For example, saying “Turn on LED” activates one GPIO output, while “Turn off LED” deactivates it.</p><p>Since everything runs locally, the response is nearly instantaneous and continues to work even without internet access.</p><h3>Wiring the Hardware</h3><p>The hardware connections are simple enough for beginners.</p><p>The USB-to-TTL converter powers the module and is also used for firmware flashing. The microphone connects to the module’s audio input pins, while the speaker connects to the onboard audio output.</p><p>GPIO pins from the SU-03T connect to LEDs through current-limiting resistors. These GPIO outputs change state when voice commands are recognised.</p><p>One important detail is that the module operates at 3.3V and is not 5V tolerant. Accidentally powering it with 5V may damage the board.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*jfStjFnfnBhwYoJnJZLuqQ.jpeg" /></figure><h3>Configuring the Voice Commands</h3><p>The SU-03T uses the Ai-Thinker Voice SDK portal for configuration and firmware generation.</p><p>The setup process involves:</p><ul><li>Creating a product profile</li><li>Selecting offline mode</li><li>Defining wake words</li><li>Adding custom commands</li><li>Assigning GPIO actions</li><li>Generating firmware</li></ul><p>You can define wake words such as:</p><ul><li>“Hello”</li><li>“Hi”</li><li>“Assistant”</li></ul><p>After the wake word is detected, the module listens for command phrases like:</p><ul><li>“Turn on light”</li><li>“Lights off”</li><li>“Fan on”</li></ul><p>Each command can also include a spoken reply from the module itself.</p><p>The platform supports wake-free commands as well, allowing specific actions to work without saying the wake word every time.</p><h3>Flashing the Firmware</h3><p>Once configuration is complete, the SDK generates firmware files for the module.</p><p>These files are uploaded using the UniOneUpdateTool utility over a USB-to-TTL interface. The process is simple but slightly inconvenient because some tools are still partially untranslated from Chinese.</p><p>Several hobbyists in the electronics community have noted that the flashing process can be confusing initially, but once configured correctly, the module performs reliably.</p><h3>Real-World Applications</h3><p>Even though this project demonstrates LED control, the same concept can be extended much further.</p><h3>Smart Home Automation</h3><p>Control lights, fans, relays, and appliances entirely offline without cloud dependency.</p><h3>Assistive Technology</h3><p>Voice commands can help elderly or physically challenged users interact with electronics more comfortably.</p><h3>Industrial Systems</h3><p>Factories and workshops often require reliable local control systems where internet access may not always be available.</p><h3>Automotive Projects</h3><p>Offline voice systems can be integrated into vehicles for hands-free control without relying on mobile connectivity.</p><h3>Educational Projects</h3><p>This setup is excellent for learning embedded AI concepts without needing expensive hardware or cloud subscriptions.</p><h3>Limitations You Should Know</h3><p>Like most low-cost embedded voice systems, the SU-03T does have limitations.</p><p>The number of fully custom commands is limited, and pronunciation accuracy matters. Background noise can also reduce recognition reliability.</p><p>Some users have reported firmware flashing issues or accidental module corruption during improper uploads. Careful handling and correct firmware selection are important.</p><p>Still, for simple offline command recognition tasks, the module performs surprisingly well for its price range.</p><h3>Future Improvements</h3><p>This project can easily evolve into a much larger automation platform.</p><p>Possible upgrades include:</p><ul><li>Controlling multiple relays or appliances</li><li>Integrating motors and servos</li><li>Adding display feedback</li><li>Combining with ESP32 or Raspberry Pi systems</li><li>Building a complete offline smart home hub</li></ul><p>Developers looking for more advanced offline AI capabilities may also explore ESP32-based TinyML voice recognition systems in the future.</p><h3>Final Thoughts</h3><p>Offline voice recognition is becoming increasingly important as makers and developers seek faster, more private, and internet-independent solutions.</p><p>The SU-03T module proves that you do not need expensive hardware or cloud AI services to build practical voice-controlled systems. With a few inexpensive components and some firmware configuration, you can create a responsive offline voice interface suitable for automation, accessibility, and embedded projects.</p><p>For beginners entering the world of voice-enabled electronics, this project is an excellent starting point. It combines embedded systems, human-machine interaction, and edge AI concepts into a simple and affordable build.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F8ugrDMHoMDg%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D8ugrDMHoMDg&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F8ugrDMHoMDg%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/a57bc094ebbcebe5053272a13a7fd0be/href">https://medium.com/media/a57bc094ebbcebe5053272a13a7fd0be/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0a50f8fbcafe" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build a Smart Grid Monitor with Arduino UNO R4 WiFi]]></title>
            <link>https://circuitdigest.medium.com/build-a-smart-grid-monitor-with-arduino-uno-r4-wifi-8dd8ba06c93e?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/8dd8ba06c93e</guid>
            <category><![CDATA[arduino]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[monitoring-system]]></category>
            <category><![CDATA[smart-grid]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Tue, 19 May 2026 06:51:22 GMT</pubDate>
            <atom:updated>2026-05-19T06:51:22.170Z</atom:updated>
            <content:encoded><![CDATA[<p>Electricity powers nearly everything we use, yet most of us only see how much energy we consume when the monthly bill arrives. What if you could monitor voltage, current, power, and energy usage in real time, both locally and from anywhere in the world?</p><p>That is exactly what this project does.</p><p>In this tutorial, you will build a practical <a href="https://circuitdigest.com/microcontroller-projects/smart-grid-iot-using-arduino-uno-r4"><strong>smart grid in IoT</strong></a> monitoring system using the Arduino UNO R4 WiFi and the PZEM-004T. The system measures key electrical parameters, displays them on an OLED screen, and uploads them to ThingSpeak for remote logging and visualisation.</p><p>This project is useful for:</p><ul><li>Monitoring household electricity consumption</li><li>Tracking the energy usage of appliances</li><li>Detecting abnormal voltage or current conditions</li><li>Learning how smart grid systems work</li><li>Building energy dashboards for homes, labs, or small industries</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*cuxc4_W7gqC3HDZXzYpJyA.jpeg" /></figure><h3>What Is a Smart Grid?</h3><p>A smart grid is an electrical system enhanced with sensors, communication, and software that continuously monitors how power is generated, distributed, and consumed.</p><p>Traditional grids only deliver electricity.</p><p>Smart grids also provide data.</p><p>By measuring parameters such as voltage, current, frequency, and power factor, smart grids help:</p><ul><li>Improve energy efficiency</li><li>Detect faults early</li><li>Reduce electricity costs</li><li>Enable predictive maintenance</li><li>Support renewable energy systems</li></ul><p>This project demonstrates a simple smart grid node that collects electrical data and sends it to the cloud.</p><h3>What This Project Measures</h3><p>The PZEM-004T module provides six important electrical parameters:</p><ol><li><strong>Voltage (V)</strong> — Supply voltage.</li><li><strong>Current (A)</strong> — Load current.</li><li><strong>Power (W)</strong> — Instantaneous real power.</li><li><strong>Energy (kWh)</strong> — Total energy consumed.</li><li><strong>Frequency (Hz)</strong> — Grid frequency.</li><li><strong>Power Factor</strong> — Efficiency of power usage.</li></ol><p>Together, these values provide a complete picture of how a load is behaving.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*vqvDz9nkIIrEgRtxvpy8kg.jpeg" /></figure><h3>How the System Works</h3><p>The project follows a simple edge-to-cloud architecture.</p><h3>1. Measurement Layer</h3><p>The PZEM-004T measures electrical parameters from the AC mains.</p><h3>2. Processing Layer</h3><p>The Arduino UNO R4 WiFi reads the data via Modbus RTU over UART.</p><h3>3. Local Display</h3><p>An OLED screen shows live readings.</p><h3>4. Cloud Layer</h3><p>The Arduino uploads data to ThingSpeak over Wi-Fi.</p><h3>5. Data Visualisation</h3><p>ThingSpeak stores and plots the data for remote monitoring.</p><h3>Why Arduino UNO R4 WiFi?</h3><p>The Arduino UNO R4 WiFi is an excellent choice for this project because it combines a powerful 32-bit microcontroller with built-in wireless connectivity. Explore 500+ <a href="https://circuitdigest.com/arduino-projects">Arduino projects</a> with complete circuit diagrams, source code, and step-by-step tutorials on CircuitDigest Arduino Projects, covering IoT, robotics, home automation, sensors, and beginner-friendly electronics builds.</p><p>Key advantages include:</p><ul><li>48 MHz Renesas RA4M1 microcontroller</li><li>Built-in Wi-Fi and Bluetooth</li><li>5V logic compatibility</li><li>More memory than the classic UNO</li><li>Hardware serial ports for reliable Modbus communication</li></ul><p>This means no external Wi-Fi module is required.</p><h3>Why Use the PZEM-004T?</h3><p>The PZEM-004T simplifies AC power measurement significantly.</p><p>Instead of designing analog circuits and performing calibration, you can directly read processed values through a serial interface.</p><p>Features include:</p><ul><li>Voltage measurement from 80–260V AC</li><li>Current measurement up to 100A using a CT clamp</li><li>Built-in calculations for power, energy, frequency, and power factor</li><li>Modbus RTU communication</li><li>Good accuracy for DIY and educational projects</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*uxcNL0Lo_Gcg0hLI9dhilQ.jpeg" /></figure><h3>Components Required</h3><p>You will need the following components:</p><ul><li>Arduino UNO R4 WiFi</li><li>PZEM-004T energy meter</li><li>Current transformer (included with PZEM)</li><li>128×64 I2C OLED display</li><li>Push button (optional)</li><li>Breadboard or PCB</li><li>Jumper wires</li><li>USB cable</li><li>AC load for testing</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*vA-OBmfQnoGnFDlQ6CUWZg.jpeg" /></figure><h3>Wiring Overview</h3><h3>PZEM-004T to Arduino UNO R4 WiFi</h3><ul><li>TX → RX1 (Pin 0)</li><li>RX → TX1 (Pin 1)</li><li>VCC → 5V</li><li>GND → GND</li></ul><h3>OLED Display to Arduino</h3><ul><li>SDA → A4</li><li>SCL → A5</li><li>VCC → 5V</li><li>GND → GND</li></ul><h3>Push Button</h3><ul><li>One side → Pin 7</li><li>Other side → GND</li></ul><p>The button can be used to switch display screens.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5b8Ys-xqd6g7jpCwXmRy8A.jpeg" /></figure><h3>Safety Warning</h3><p>This project works with mains AC voltage, which can be lethal.</p><p>Always:</p><ul><li>Disconnect power before wiring.</li><li>Use insulated tools.</li><li>Keep exposed conductors covered.</li><li>Test with proper caution.</li><li>If you are inexperienced, seek assistance from a qualified electrician.</li></ul><p>Safety should always come first.</p><h3>Setting Up ThingSpeak</h3><p>ThingSpeak is a free cloud platform for storing and visualizing sensor data.</p><p>To set it up:</p><ol><li>Create an account on ThingSpeak.</li><li>Create a new channel.</li><li>Add six fields:</li></ol><ul><li>Voltage</li><li>Current</li><li>Power</li><li>Energy</li><li>Frequency</li><li>Power Factor</li></ul><p>4. Copy your Channel ID.</p><p>5. Copy your Write API Key.</p><p>You will use these values in the Arduino code.</p><h3>Required Libraries</h3><p>Install the following libraries using the Arduino Library Manager:</p><ul><li>PZEM004Tv30 or compatible PZEM library</li><li>Adafruit SSD1306</li><li>Adafruit GFX</li><li>ThingSpeak</li><li>WiFiS3</li></ul><h3>Program Workflow</h3><p>The firmware performs the following loop:</p><ol><li>Connect to Wi-Fi.</li><li>Read all parameters from the PZEM-004T.</li><li>Display the values on the OLED.</li><li>Send data to ThingSpeak.</li><li>Wait for the next update interval.</li><li>Repeat.</li></ol><p>This continuous loop creates a real-time energy monitoring system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*qJKKYDXJcH-oru24vTuPlg.jpeg" /></figure><h3>Reading Data from the PZEM</h3><p>The PZEM library handles all Modbus communication internally.</p><p>The Arduino simply requests data:</p><ul><li>Voltage</li><li>Current</li><li>Power</li><li>Energy</li><li>Frequency</li><li>Power factor</li></ul><p>This keeps the code clean and easy to understand.</p><h3>Displaying Data on OLED</h3><p>The OLED provides instant visual feedback.</p><p>Typical display content includes:</p><ul><li>Voltage: 230.4 V</li><li>Current: 1.23 A</li><li>Power: 283 W</li><li>Energy: 1.47 kWh</li><li>Frequency: 50.0 Hz</li><li>PF: 0.98</li></ul><p>Even without internet access, you can still monitor the readings locally.</p><h3>Uploading Data to ThingSpeak</h3><p>Each parameter is assigned to a field:</p><ul><li>Field 1 → Voltage</li><li>Field 2 → Current</li><li>Field 3 → Power</li><li>Field 4 → Energy</li><li>Field 5 → Frequency</li><li>Field 6 → Power Factor</li></ul><p>ThingSpeak automatically generates charts from incoming data.</p><h3>Real-World Applications</h3><p>This project can be adapted for many practical uses.</p><h3>Home Energy Monitoring</h3><p>Track how much electricity appliances consume.</p><h3>Industrial Monitoring</h3><p>Identify overloaded machines and abnormal operating conditions.</p><h3>Renewable Energy Systems</h3><p>Compare solar generation with load consumption.</p><h3>Educational Labs</h3><p>Teach electrical measurements and IoT fundamentals.</p><h3>Preventive Maintenance</h3><p>Spot unusual trends before equipment fails.</p><h3>Testing the System</h3><p>When you connect a load:</p><ol><li>The PZEM measures electrical values.</li><li>The Arduino reads and processes them.</li><li>The OLED updates instantly.</li><li>Data appears in ThingSpeak graphs.</li></ol><p>You can observe:</p><ul><li>Startup current spikes</li><li>Voltage fluctuations</li><li>Changes in power factor</li><li>Long-term energy accumulation</li></ul><p>These insights are difficult to obtain from a conventional energy meter.</p><h3>Optional Improvements</h3><p>Once the basic system is working, you can extend it with:</p><ul><li>Email or SMS alerts</li><li>SD card logging</li><li>Web dashboard</li><li>Relay control</li><li>Multi-channel monitoring</li><li>MQTT integration</li><li>Mobile app notifications</li></ul><p>The modular design makes future upgrades straightforward.</p><h3>Common Troubleshooting Tips</h3><h3>No Data from PZEM</h3><ul><li>Check TX/RX wiring.</li><li>Verify the baud rate.</li><li>Confirm common ground.</li></ul><h3>OLED Not Working</h3><ul><li>Check the I2C address.</li><li>Confirm SDA and SCL connections.</li></ul><h3>Wi-Fi Connection Fails</h3><ul><li>Verify SSID and password.</li><li>Ensure 2.4 GHz Wi-Fi is available.</li></ul><h3>ThingSpeak Not Updating</h3><ul><li>Confirm Channel ID and API key.</li><li>Respect ThingSpeak’s update interval.</li></ul><h3>What I Learned from Building This</h3><p>Projects like this turn electrical concepts into something tangible.</p><p>Voltage fluctuations are no longer abstract numbers.</p><p>Power factor becomes meaningful.</p><p>Current spikes reveal exactly how motors and appliances behave.</p><p>By visualising data continuously, you gain a much deeper understanding of electrical systems.</p><p>That is why energy monitoring projects are so rewarding to build.</p><h3>Conclusion</h3><p>This project demonstrates how to build a practical smart grid monitoring node using the Arduino UNO R4 WiFi and PZEM-004T.</p><p>With just a few components, you can:</p><ul><li>Measure six electrical parameters</li><li>Display them locally</li><li>Upload them to the cloud</li><li>Analyse trends over time</li></ul><p>The result is a powerful and affordable IoT solution for learning, experimentation, and real-world energy monitoring.</p><p>Whether you are a student, maker, or engineer, this <a href="https://circuitdigest.com/internet-of-things-iot-projects"><strong>IoT project</strong></a> is an excellent introduction to both electrical measurement and IoT-based smart grid technology.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FSy0ap-J-04A%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DSy0ap-J-04A&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FSy0ap-J-04A%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/646ccae69ac8343d10aa78790bc84b64/href">https://medium.com/media/646ccae69ac8343d10aa78790bc84b64/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8dd8ba06c93e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Smart Waste Segregation with Arduino and TinyML]]></title>
            <link>https://circuitdigest.medium.com/smart-waste-segregation-with-arduino-and-tinyml-0f0ab0bb730c?source=rss-d4da06c4ac74------2</link>
            <guid isPermaLink="false">https://medium.com/p/0f0ab0bb730c</guid>
            <category><![CDATA[waste-segregation]]></category>
            <category><![CDATA[arduino-projects]]></category>
            <category><![CDATA[ai-projects]]></category>
            <category><![CDATA[arduino-uno-q]]></category>
            <dc:creator><![CDATA[Circuit Digest]]></dc:creator>
            <pubDate>Tue, 19 May 2026 06:27:18 GMT</pubDate>
            <atom:updated>2026-05-19T06:27:18.557Z</atom:updated>
            <content:encoded><![CDATA[<h3>Build an AI-powered system that detects and sorts paper, plastic, cardboard, and batteries automatically</h3><p>Waste segregation is one of the simplest habits that can make a big environmental impact. Yet in homes, offices, and public spaces, recyclable and hazardous materials are still often thrown into a single bin.</p><p>To address this problem, I built a smart waste segregation system using the Arduino UNO and machine learning. The system uses a USB camera to identify different types of waste in real time and then automatically directs each item to the appropriate bin using a servo motor. If a battery is detected, the system sounds a buzzer to warn the user instead of sorting it.</p><p>This project combines computer vision, embedded AI, and basic electronics to create a practical prototype for automated recycling.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*_GAlkhkca2xWsFzeAy9-fA.jpeg" /></figure><h3>Why Automated Waste Segregation Matters</h3><p>Improper waste disposal creates several environmental problems:</p><ul><li>Recyclable materials become contaminated.</li><li>Hazardous items, such as batteries, leak toxic chemicals.</li><li>Landfills fill up faster.</li><li>Recycling becomes more expensive and less efficient.</li></ul><p>An intelligent sorting system reduces human effort and improves recycling accuracy. While this prototype is small, the same concept can be scaled for schools, apartment complexes, and smart city infrastructure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*FtuJ0-nGAdFoqQVvjMCTVg.jpeg" /></figure><h3>How the System Works</h3><p>The system follows a simple workflow:</p><ol><li>A USB camera captures an image of the waste item.</li><li>A machine learning model classifies the object as:</li></ol><ul><li>Paper</li><li>Plastic</li><li>Cardboard</li><li>Battery</li></ul><p>3. A Python script validates the detection across multiple frames.</p><p>4. The script sends a command to the Arduino.</p><p>5. The Arduino rotates a servo motor to the correct position.</p><p>6. If a battery is detected, a buzzer sounds instead of moving the servo.</p><p>This creates a fully automated sorting mechanism that works in real time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*RgD-8lbWBzg3GD2-VQcSrg.jpeg" /></figure><h3>Components Required</h3><h3>Hardware</h3><ul><li>Arduino UNO R4 WiFi or Arduino UNO Q</li><li>USB camera</li><li>Micro servo motor (SG90 or similar)</li><li>Active buzzer</li><li>USB hub</li><li>Jumper wires</li><li>Cardboard or acrylic for the bin structure</li><li>Laptop or PC</li></ul><h3>Software</h3><ul><li>Edge Impulse</li><li>Arduino IDE</li><li>Arduino App Lab</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*5HgnJTyyb6kAA75rsjGelw.jpeg" /></figure><h3>Why Use Arduino UNO Q?</h3><p>The Arduino UNO Q supports Arduino App Lab and RouterBridge, which allow Python applications to communicate directly with the board. This makes it possible to run a machine learning model on the computer while using the Arduino for physical control.</p><p>This hybrid architecture is ideal for AI-powered hardware projects because it separates high-level vision processing from low-level actuator control.</p><h3>Hardware Connections</h3><p>The wiring is straightforward:</p><ul><li>Servo signal pin → Digital Pin 9</li><li>Buzzer positive pin → Digital Pin 8</li><li>Servo VCC → 5V</li><li>Servo GND → GND</li><li>Buzzer GND → GND</li><li>USB camera → Laptop via USB</li><li>Arduino → Laptop via USB</li></ul><p>The laptop powers both the Arduino and the camera.</p><h3>Building the Waste Sorting Mechanism</h3><p>I created a simple cardboard structure with two collection sections:</p><ul><li>Left side for biodegradable waste (paper and cardboard)</li><li>Right side for non-biodegradable waste (plastic)</li></ul><p>A servo motor controls a flap that directs each item into the correct section.</p><p>Batteries are treated separately as hazardous waste, so the system only triggers an audible alert.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*CEGbmKWz-CwMSH0LF7K6tw.jpeg" /></figure><h3>Training the Machine Learning Model</h3><p>The object detection model was trained using Edge Impulse.</p><h3>Dataset Classes</h3><ul><li>Paper</li><li>Plastic</li><li>Cardboard</li><li>Battery</li></ul><h3>Steps</h3><ol><li>Collect images of each waste type from different angles and lighting conditions.</li><li>Label each image.</li><li>Train an object detection model.</li><li>Evaluate model accuracy.</li><li>Export the model for use in Python.</li></ol><p>For best results, include at least 100 images per class.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*mgyGYZVicaGb1xXimoaQ_w.jpeg" /></figure><h3>Real-Time Detection Logic</h3><p>The Python script processes video frames and applies several safeguards:</p><h3>Confidence Threshold</h3><p>Detections are accepted only when confidence exceeds a predefined value.</p><h3>Stability Counter</h3><p>The same class must be detected in multiple consecutive frames.</p><h3>Cooldown Timer</h3><p>Prevents repeated triggering when the same object remains in view.</p><p>This combination makes the system much more reliable in real-world conditions.</p><h3>Sorting Logic</h3><p>When a valid detection is confirmed, the following actions are performed:</p><h3>Paper and Cardboard</h3><p>The servo rotates to 0° and directs the object to the biodegradable bin.</p><h3>Plastic</h3><p>The servo rotates to 180° and directs the object to the non-biodegradable bin.</p><h3>Battery</h3><p>The buzzer turns on for two seconds to warn the user.</p><p>After sorting, the servo returns to its neutral 90° position.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*455XiVadM3LP7igFZASSRw.jpeg" /></figure><h3>Arduino Firmware Overview</h3><p>The Arduino sketch performs three main tasks:</p><ol><li>Initialises the servo and buzzer.</li><li>Registers callable functions:</li></ol><ul><li>set_servo(angle)</li><li>buzz(duration)</li></ul><p>3. Continuously listens for commands from Python.</p><p>This keeps the microcontroller code compact and easy to maintain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*ryxhUOxbF0OygvghsL8sLA.jpeg" /></figure><h3>Python Application Overview</h3><p>The Python application handles:</p><ul><li>Camera input</li><li>Running the object detection model</li><li>Filtering detections</li><li>Sending commands to the Arduino</li><li>Updating a web dashboard</li></ul><p>This architecture makes it easy to refine the AI logic without modifying the Arduino code.</p><h3>System Startup Sequence</h3><p>When the system powers on:</p><ol><li>The servo moves to 90°.</li><li>The buzzer emits a short double beep.</li><li>The camera begins streaming.</li><li>The object detection model loads.</li><li>The system starts monitoring for waste items.</li></ol><h3>Performance and Accuracy Tips</h3><p>To improve detection reliability:</p><ul><li>Use good lighting.</li><li>Avoid cluttered backgrounds.</li><li>Capture diverse training images.</li><li>Set appropriate confidence thresholds.</li><li>Use multiple stable frames before triggering actions.</li></ul><p>With a well-trained model, the system can identify waste classes accurately and consistently.</p><h3>Applications</h3><p>This project can be adapted for many real-world scenarios:</p><ul><li>Smart home recycling systems</li><li>School science exhibitions</li><li>College final-year projects</li><li>Office waste management</li><li>Public recycling kiosks</li><li>Smart city prototypes</li></ul><h3>Future Improvements</h3><p>Possible upgrades include:</p><ul><li>Additional waste categories such as metal and glass</li><li>Conveyor belt automation</li><li>Load cells to measure bin weight</li><li>IoT-based fill-level monitoring</li><li>Cloud analytics dashboard</li><li>Solar-powered operation</li></ul><h3>Challenges Faced</h3><p>Building this <a href="https://circuitdigest.com/microcontroller-projects/automatic-waste-segregation-system-using-arduino-uno-q"><strong>Automatic Waste Segregation System</strong></a> project involved several practical challenges:</p><ul><li>Collecting a balanced image dataset</li><li>Avoiding false detections</li><li>Calibrating servo positions</li><li>Managing timing between detections</li><li>Handling hazardous items separately</li></ul><p>These challenges made the project an excellent learning experience in <a href="https://circuitdigest.com/ai-projects-and-tutorials"><strong>AI Projects </strong></a>and embedded systems.</p><h3>What I Learned</h3><p>This project helped me understand:</p><ul><li>TinyML and Edge AI</li><li>Computer vision fundamentals</li><li>Servo motor control</li><li>Hardware–software integration</li><li>Real-world automation design</li></ul><p>More importantly, it demonstrated how simple engineering solutions can contribute to environmental sustainability.</p><h3>Conclusion</h3><p>This smart waste segregation system shows how machine learning and microcontrollers can work together to solve everyday problems.</p><p>Using a camera, an Arduino, and a trained AI model, the system automatically identifies and sorts waste into the appropriate category while safely alerting users when hazardous materials are detected.</p><p>It is a practical, low-cost project that blends embedded systems, artificial intelligence, and environmental responsibility into one meaningful prototype.</p><p>If you are interested in <a href="https://circuitdigest.com/arduino-projects"><strong>Arduino Projects</strong></a>, TinyML, or sustainable technology, this project is a rewarding build that offers both technical depth and real-world relevance.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FBy-1X25D02Y%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DBy-1X25D02Y&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FBy-1X25D02Y%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/4229576623a61f0e98236449dabaac24/href">https://medium.com/media/4229576623a61f0e98236449dabaac24/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0f0ab0bb730c" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>