How William and His Team Built an AI LEGO Sorting Robot with Hiwonder JetArm
At Hiwonder, we firmly believe that robotics education should extend far beyond mere theory—students must have the hands-on opportunity to build, train, and deploy truly intelligent systems.
Recently, we were thrilled to see this philosophy realized in a student research project at an Indonesian school, where William and his team utilized the JetArm robotic arm to build an AI-powered LEGO sorting machine.
Not only was this project featured in their school publication, but it also brilliantly demonstrates how modern robotics education seamlessly integrates computer vision, machine learning, and real-world manipulation tasks.
From Ideation to Implementation: Building the Core Pipeline
The project centers on a classic automation challenge: enabling a robotic system to autonomously identify, locate, and sort various LEGO pieces.
To solve this, the students engineered an integrated, end-to-end pipeline that combines:
● Computer vision-based object detection
● Coordinate mapping between the camera and the robot's workspace
● Robotic arm motion control
● Automated pick-and-place logic
In doing so, the students transformed a simple toy scenario into a sophisticated, AI-driven robotic system.
The JetArm relies on YOLO-based computer vision to dynamically track LEGO bricks across its workspace, instantly translating visual data into the precise spatial coordinates required for real-time decision-making.
Since raw camera frames cannot directly govern physical joint angles, the team established a coordinate transformation process. This step effectively bridges the gap between visual perception and physical interaction by mapping pixel positions into real-world robot coordinates.
From there, the JetArm utilizes inverse kinematics control to calculate and execute precise pick-and-place sequences, closing the loop between sight and action.
Real-World Engineering Challenges
Building a fully autonomous system is rarely a straight line. Throughout development, William’s team faced several authentic engineering friction points:
Resolving these complex issues demanded rigorous, iterative testing—a core tenet of effective robotics engineering education.
Learning by Building
This project perfectly embodies a core conviction we hold at Hiwonder: robotics is best mastered through immersive, hands-on development and building real systems.
Through hands-on development, the students navigated essential pillars of real-world robotics engineering, specifically critical challenges like object detection reliability, coordinate calibration, and motion precision.
Above all, it gave them a front-row seat to how AI and robotics converge into embodied intelligent systems, bridging the gap between theory and practice.
Empowering a Global Community of Young Builders
We are proud to see Hiwonder robots being used by students and educators worldwide to explore robotics and AI in creative and meaningful ways.
Success stories like this highlight a thriving global community of young creators who are not just learning robotics but applying it to create real, working intelligent systems.
We can't wait to see what William's team—and other creators in the Hiwonder community—will build next!