Home▸Robotics & Kinematics▸Behavior Cloning & Distribution Shift

Behavior Cloning & Distribution Shift in Robot Manipulation

Interactive behavior-cloning manipulation policy: a k-nearest-neighbor regression policy learns from recorded expert demonstrations, and you control how far test rollouts start outside the training distribution to watch covariate shift and compounding error live.

Robotics & Kinematics3DAdvanced60 FPS📱 Mobile-adapted
Behavior Cloning & Distribution Shift ↗ Open standalone

Instead of hand-coding a control law, this simulator trains a manipulation policy the way real learning-based robots are trained: by cloning an expert's demonstrations. Recorded (state, action) pairs from an idealized expert form a training dataset, and a k-nearest-neighbor regression policy reproduces the expert's behavior anywhere near that data. Push the rollout's start point outside the demonstrated region and watch covariate shift take hold live — the policy's commanded direction gets noisier the farther it is from any demonstration, the gripper drifts into even less-covered territory, and the error compounds exactly as it does in real imitation-learning failures, the core problem that motivated on-policy corrections like DAgger and RL fine-tuning.

⚙ Under the hood

A k-nearest-neighbor behavior-cloning policy learns a manipulation task from recorded expert demonstrations; push test rollouts outside the training region to watch covariate shift and compounding error take hold live.

roboticsbehavior cloningimitation learningmanipulationreinforcement learningsim-to-real

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install

What did you find?

Add reproduction steps (optional)