PhysicalAI Star Arm Plugin

Fashion Star Arm 102 plugin for PhysicalAI. This package provides Star Arm 102-LD/102-HD leaders and a Star Arm 102-FL follower, and registers all variants with Physical AI Studio.

Features

• Star Arm 102-LD leader driver (StarArm102LDLeader) for read-only teleoperation input
• Star Arm 102-HD leader driver (StarArm102HDLeader) with passive mode (default) and optional assist mode for hold/action commands
• Star Arm 102-FL follower driver (StarArm102FLFollower) for position control
• Studio catalog plugin entries for LD leader, HD leader, and FL follower
• Bundled stararm102 URDF package

Installation

uv add physicalai-stararm-plugin

Quick start

from physicalai.robot import connect
from physicalai_stararm_plugin import StarArm102FLFollower

robot = StarArm102FLFollower(port="/dev/ttyUSB1", baudrate=1_000_000)

with connect(robot) as arm:
    obs = arm.get_observation()
    arm.send_action(obs.joint_positions)

Run with the PhysicalAI CLI

uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/teleop-hd-to-fl.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/teleop-ld-to-fl.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/read-joints-hd.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/read-joints-ld.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/read-joints-fl.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/move-joints-fl.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/hold-hd.yaml
uv run physicalai run --config packages/physicalai-stararm-plugin/examples/runtime/policy-follow-hd.yaml

Zero-pose calibration

StarArm102HDLeader.set_zero_position() (also on the LD leader) stores the arm's current pose as every servo's origin. Physical AI Studio runs it from its calibration step when you add a Star Arm leader: move the leader into the same rest pose as the follower with the gripper closed, then set zero.

URDF

from physicalai_stararm_plugin import get_urdf_path

urdf_dir = get_urdf_path()
hd_urdf = urdf_dir / "stararm102" / "urdf" / "stararm102_hd_description.urdf"

Development

uv sync
uv run pytest packages/physicalai-stararm-plugin/tests/