Visual Servoing & Precision Assembly
Eye-in-hand image-based visual servoing (IBVS): a robot arm with a wrist camera drives a peg-in-hole insertion by continuously reducing the error between current and desired image feature points, compensating for part-position miscalibration in real time.
This simulator visualizes image-based visual servoing (IBVS) on an eye-in-hand robot arm performing a peg-in-hole insertion. Rather than driving to a pre-programmed coordinate, the controller continuously compares the current image positions of the hole's corner features against their desired positions and turns that pixel-space error into a velocity command — closing the loop on vision itself. Shift the part's true position with the calibration-error slider and watch the same feedback loop find and insert the peg anyway, then read the feature-error curve as it collapses toward zero with each control iteration.
Eye-in-hand image-based visual servoing (IBVS) on a peg-in-hole assembly: a wrist camera measures the image-plane error between current and desired hole-corner feature points each control tick and converts it into an end-effector velocity command, converging without a pre-programmed part coordinate. A slider injects part-position miscalibration; a live graph plots feature error collapsing toward zero.
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install