This is how I built my first robotic arm.

@pathofseb

31.4K followers · usually 15.0K views a post

Views

436.6K

Likes

29.7K

Comments

209

Shares

1.8K

Selling · Other Product

robotic arm

Why it worked

The video demonstrates a complex DIY project with clear explanations and visual feedback, appealing to audiences interested in robotics, engineering, and computer science. The use of computer vision to control the robotic arm in real-time is a compelling visual element.

Summary

The creator demonstrates how they built a robotic arm using servo motors and computer vision. The arm's fingers move in response to hand gestures captured by a webcam and processed with Python code.

Structure

1Introduction to the robotic arm and its components
2Demonstration of finger retraction and extension mechanism
3Overview of the building process: 3D modeling, printing, and assembly
4Explanation of the computer vision integration using Python and MediaPipe
5Demonstration of hand gesture tracking and robotic arm response
6Conclusion on the functionality of detecting finger states and hand rotation

Product placement

The video showcases a custom-built robotic arm, detailing its construction with servo motors (SG90 and MG996r), 3D printed parts, and integration with computer vision for hand gesture recognition.

Original caption

This is how i built my first robotic arm. #robotics #engineering #computerscience #computervision #ai #polymath #student #learning #coding

More from @pathofseb

Their other posts in the index, biggest breakout first.

On-screen text

Fingers: 0 FPS: 20 Fingers: 3 FPS: 19 Fingers: 3 FPS: 20 Fingers: 0 FPS: 21 Fingers: 0 FPS: 17 Fingers: 4 FPS: 15 Fingers: 5 FPS: 21 Fingers: 0 FPS: 21 Fingers: 2 FPS: 21 Fingers: 0 FPS: 21 Fingers: 5 FPS: 21 Fingers: 3 FPS: 12 FPS: 20

Transcript

This is how I built my first robotic arm. I built it using these servo motors. SG90, MG996r. Pull on a fishing line like this, which makes the fingers retract. And then having an elastic band connected to the fingers, making them stretch out again when the tension is removed. The builder process for this was fairly simple. I started by finding a 3D model of a line that I liked. And made a few tweaks to better suit my design. Once printed, assembling the parts was pretty straightforward. I connected each servo to the board and powered everything using an external power supply. With hardware ready, the next step was integrating computer vision. With just a few lines of Python code, using the MediaPipe library, I was able to project 21 landmarks onto my hand in real time. These points track every movement on my hand, including bending, extension, and rotation. Using a simple function, I can detect whether my fingers are open or closed. And even track the overall rotation of my hand.