DOCSEARCHbyAlgolia
Submit Docs

OpenArm Documentation

website·Indexed 33 days ago

https://docs.openarm.dev

Documentation for the OpenArm robotic arm project, covering hardware assembly, specifications, and teleoperation. Includes guides for version 1.0 and 2.0, featuring the OpenArm 2.0 Arm and OpenArm KER (Kinematic Equivalent Replica). Provides technical details on CAN interface setup, VR teleoperation, MuJoCo simulation, and the OpenArm Dataset version 2.0 API for loading observations and actions.

Tokens
48.1K
Snippets
201
Records
389
Agent score
67%

What's inside OpenArm

  1. Overview of the OpenArm software ecosystem

    OpenArm provides a toolkit for bridging bimanual robots, offering a suite of tools for hardware interaction and intelligent control. The ecosystem includes modular robot descriptions (URDF/XACRO), high-frequency low-level control APIs, a CAN CLI for motor configuration, ROS2 middleware integration, and Dora-based Python nodes for dataflow-driven robot control and inference.

  2. Overview of the OpenArm CAN Library

    The OpenArm CAN Library is a communication bridge between high-level control applications and low-level motor protocols. It utilizes the Linux SocketCAN interface to provide an API for motor control and state monitoring. The library is structured into three layers: can/socket (high-level components like OpenArm, ArmComponent, and GripperComponent), damiao_motor (motor protocol implementation), and canbus (low-level SocketCAN interface).

  3. Overview of OpenArm VR Teleoperation capabilities

    The OpenArm VR teleoperation system enables real-time control of the OpenArm robot using streaming data from VR human body estimation. Key technical capabilities include:

    • 7DOF Control: Maps both position and rotation for intuitive manipulation.
    • Gesture-based Gripper Control: Uses VR finger distance to trigger real-time open/close actions.
    • Scalable Mapping: Supports full arm motion (upper and lower arm) for human-like movement.
    • Performance: Operates at approximately 20 FPS using TCP data streaming.

  4. Overview of OpenArm VR Teleoperation capabilities

    The OpenArm VR teleoperation system (Version 2.0) enables real-time control of the OpenArm robot using VR human body estimation streaming data. Key technical capabilities include:

    • 7DOF Control: Maps both position and rotation for intuitive manipulation.
    • Gesture-based Gripper Control: Uses VR finger distance to trigger real-time open/close actions.
    • Scalable Mapping: Supports full arm motion (upper and lower arm) for human-like movements.
    • Performance: Operates at approximately 20 FPS using TCP data streaming.

  5. Overview of OpenArm humanoid arm specifications

    OpenArm is an open-source 7DOF humanoid arm designed for physical AI research and contact-rich environments. It features a human-scale design (scaled for a person 160-165cm tall) and a safety-first architecture using QDD backdrivable motors and high compliance for safe human-robot interaction.

    Key technical specifications include:

    • Payload Capacity: 6.0kg peak / 4.1kg nominal.
    • Construction: Aluminum and stainless steel structural components for durability.
    • Accessibility: Fully open-source hardware and software, including CNC parts, 3D-printed casings, and electrical wiring data.
    • Cost: Approximately 6,500 USD for a complete bimanual system.

  6. Overview of Bilateral Control for force feedback

    Bilateral Control enables two-way force feedback between the leader and follower arms. Unlike Unilateral Control, this mode allows the operator to feel the physical interactions and contact forces experienced by the follower arm.

  7. Overview of OpenArm Isaac Lab Simulation

    OpenArm is a framework used to develop and evaluate reinforcement learning (RL), imitation learning, and foundation model–based approaches within the NVIDIA Isaac Sim and Isaac Lab environments. The project provides simulation assets, training workflows, and benchmarks designed for application to real-world robotic systems. The codebase is hosted on GitHub under an open-source license.

  8. Overview of OpenArm Isaac Lab Simulation

    OpenArm is a framework used to develop and evaluate reinforcement learning (RL), imitation learning, and foundation model-based approaches within the NVIDIA Isaac Sim and Isaac Lab environments. It provides simulation assets, training workflows, and benchmarks designed for application to real-world robotic systems.

  9. Overview of the OpenArm software ecosystem

    The OpenArm software ecosystem provides a toolkit for bimanual robot control, bridging hardware with intelligent control through several key components: modular XACRO files for URDF robot models, high-frequency low-level control APIs for motor interaction, ROS2 middleware compatibility, and advanced control algorithms for system identification.

  10. Overview of OpenArm Cell hardware

    OpenArm Cell (Version 2.0) is a standardized benchmark system designed for evaluating robotic foundation models. It provides a unified environment that includes the robot arm, lighting, cameras, and calibration procedures to ensure reproducibility in model performance evaluation.

  11. Use OpenArm KER for Low-Fatigue Teleoperation

    OpenArm KER (Kinematic Equivalent Replica) is a motorless leader arm that exactly matches the kinematics of the OpenArm 2.0. Because it has no actuators, it is lightweight and designed to reduce operator fatigue during long data-collection sessions where the operator must drive the follower arm for several hours.

  12. Understand OpenArm teleoperation thread architecture

    The project utilizes a multi-threaded structure managed by a custom class derived from PeriodicTimerThread. It spawns three specific threads to handle operations: the Leader thread (leader arm control logic), the Follower thread (follower arm control logic), and the Admin thread (coordination and communication). The Control class encapsulates the core logic and supports both Bilateral control (force-feedback) and Unilateral control (open-loop/single-arm).