ArtificialGate teaching adaptation. T geometry and goal-coverage task are based on the open-source Diffusion Policy Push-T environment. This is not the benchmark's original Pymunk runtime, a trained Diffusion Policy, or VLM inference.
Planar constraints, contact parameters, damping and actuator limits are educational modeling choices, not hardware calibration. No manufacturer endorsement or certification is implied.