Thank you for the excellent work on the universal_manipulation_interface project. By significantly reducing hardware requirements for data collection, the framework makes real-world dataset acquisition much more accessible, which is highly valuable to the community.
As we explore deploying the collected data on real robotic systems for inference and execution, we have noticed that the currently supported robot arms and grippers are relatively expensive. This may limit broader adoption, especially in academic or hobbyist settings.
I would like to inquire whether there are any recommended low-cost alternatives for the robot arm and gripper that could be compatible with the system. For instance, would it be feasible to support configurations similar to the Lerobot robotic arm and its associated grippers?
Any recommendations or guidance regarding affordable hardware options would be greatly appreciated.
Thank you for the excellent work on the universal_manipulation_interface project. By significantly reducing hardware requirements for data collection, the framework makes real-world dataset acquisition much more accessible, which is highly valuable to the community.
As we explore deploying the collected data on real robotic systems for inference and execution, we have noticed that the currently supported robot arms and grippers are relatively expensive. This may limit broader adoption, especially in academic or hobbyist settings.
I would like to inquire whether there are any recommended low-cost alternatives for the robot arm and gripper that could be compatible with the system. For instance, would it be feasible to support configurations similar to the Lerobot robotic arm and its associated grippers?
Any recommendations or guidance regarding affordable hardware options would be greatly appreciated.