hub 4

Inspiration

Robots often respond through screens, sounds, or simple lights. I wanted to explore whether a robot could communicate through movement. I imagined a small creature with four different limbs, representing humans, animals, nature, and machines, that could observe its surroundings and physically express what it understood.

What it does

Hub4 is a four-limbed robotic creature that reacts to the world in real time. My laptop camera detects people and objects placed in front of it. Hub4 interprets what it sees and selects a matching personality or “mood.”

For example, it can greet a person, unfold like a flower when it sees a plant, move like a spider when it recognizes an animal, or perform a scanning motion when it detects technology. When nothing is present, it settles into a gentle idle animation.

How I built it

I built Hub4 using four Feetech STS3215 smart servos connected through a central communication bus. Each servo controls one expressive limb.

The vision system uses my laptop camera, OpenCV, and a YOLOv8 object-detection model. Detected objects pass through a semantic-matching system that connects them to different moods and animations. I created a custom choreography engine that coordinates all four servos using smooth movements instead of abrupt position changes.

I also designed the structure for rapid 3D printing, keeping each arm small enough to print in approximately 30 minutes.

Challenges I ran into

One major challenge was making the movement feel expressive instead of mechanical. Moving every limb simultaneously looked unnatural, so I developed staggered movements, waves, pauses, and smaller background motions.

I also had to prevent the robot from constantly changing reactions when the camera produced uncertain detections. I solved this by requiring several matching detections before changing moods and adding a minimum delay between transitions.

Setting up Python, camera access, object-detection files, servo communication, and hardware calibration also required significant troubleshooting.

Accomplishments that I’m proud of

I am proud that I connected real-time computer vision directly to physical expression. Hub4 does not simply identify an object on a screen; it transforms that recognition into coordinated movement.

I also created a modular system where I can add new objects, moods, poses, and animations without rebuilding the entire program. A dry-run mode lets me safely test the camera and recognition system without activating the motors.

What I learned

I learned that expressive robotics involves much more than moving motors. Timing, movement intensity, pauses, and transitions can give a machine a recognizable personality.

I also gained experience with computer vision, semantic matching, servo control, animation design, rapid 3D printing, hardware integration, and safe testing.

What’s next for Hub4

Next, I plan to create custom sculpted attachments for each limb, improve object recognition, and add sound and proximity sensing. I also want to replace the laptop with an onboard computer so Hub4 can operate independently as an interactive robotic creature.

Built With

• 3dprinted
• autodesk-fusion-360
• solder