Explore / Case study / Askable Arm — Robot Action Selection in Simulation

Askable Arm — Robot Action Selection in Simulation

Self-hostWebmacOSTerminal

Choose a next movement or gripper action from a fixed skill library in ManiSkill.

by admin · listed 7 days ago

⚡ What System1 does here

Jev reads the goal, simulator-provided object positions, gripper state and recent actions. It selects a next step from about 30 hand-written primitives and a target object when needed. A Python executor performs the move before the next decision. Jev does not output torques or complete trajectories and does not process camera images.

About

An experiment that uses natural-language goals to direct a simulated Franka Panda arm in ManiSkill. Its existing skills cover tasks such as picking up a cube, placing it on a target and pressing buttons in sequence. The repository provides local voice/text interfaces and a batch evaluation script. Evidence and limits: A simulation demo, not a physical-robot deployment. Seed-0 examples do not establish cross-scene success rates. Privileged simulator coordinates do not test real-camera errors. Running it needs a TypeSafe API key. No published image or video was found in the repository. This review checked public documentation, code and media; it did not rerun simulations or test physical hardware. Sources: https://github.com/TarunTomar122/jev-askable-arm/blob/main/README.md https://github.com/TarunTomar122/jev-askable-arm/blob/main/jev_robotics/primitives.py

🧪 Community test reports

0 reports · 0 passed on the real model

No one has tested this yet. Be the first.

💬 Discussions

New post →

No discussions yet.