Skip to content

README.md

Latest commit

 

History

History
91 lines (67 loc) · 2.38 KB

README.md

File metadata and controls

91 lines (67 loc) · 2.38 KB

AeroPiper: A Dual-Hand Manipulation System

Demo

Images

PiPER Arm TetherIA Aero Hand

Description

AeroPiper is a dual-hand manipulation system that combines two AgileX PiPER 6‑DOF robotic arms with two TetherIA Aero Open hands, targeting dexterous, human-like manipulation tasks. The project pairs high-fidelity MuJoCo simulation assets with reinforcement-learning baselines so you can prototype, train, and evaluate complex bimanual skills rapidly—then transfer them to real hardware.

Official resources

  • TetherIA Aero Hand Open Docs: https://docs.tetheria.ai
  • AgileX PiPER product page: https://global.agilex.ai/products/piper
  • MuJoCo documentation: https://mujoco.readthedocs.io
  • Robosuite: https://robosuite.ai/

Installation

Setup (conda, Python 3.10)

conda create -n aeropiper python=3.10 -y
conda activate aeropiper
pip install -e .

Install (pip-only)

From the repo root:

python -m pip install -r requirements.txt
python -m pip install -e .

Verify Installation

python demos/run.py

This file send 14 random action values to the 14 DOFs.

Notes:

  • Tkinter: teleop/gui.py uses Tk. If your environment doesn't have it, install via conda (conda install -n aeropiper tk -y) or on Ubuntu/Debian: sudo apt-get install -y python3-tk

Quick Start

To run with any random action

python demos/demo_random_action.py

To run and integrate Teleoperation: Follow the current GUI control

python teleop/gui.py

This GUI file control 6+6 DOFs for both arms, and two control values applies the all the 6 DOFs of each Gripper.

Camera Integration

python demos/demo_with_camera.py

This demo displays the simulated Intel RealSense D435 camera feeds from both wrists alongside the robot simulation. The cameras are already integrated into the AeroPiper model. Press 'q' to quit, 'r' to reset, or SPACE to pause/unpause.

VR Teleoperation

Setup

  1. Launch SteamVR and connect Meta Quest
  2. Run teleoperation:
python teleop/vr_control.py

Calibration

To calibrate and record poses:

python teleop/vr_module/vr_joint_calibration.py