Agility A2 + ExoArm-7 Exoskeleton Arm VLA Suite | Wearable Bimanual Teleoperation Platform for Embodied AI Research, Robot Learning, Rapid POC Validation

The Agility A2 + ExoArm-7 VLA Suite is a wearable dual-arm teleoperation and robot learning platform designed for embodied AI research, human demonstration data collection, and rapid POC validation. It combines the Agility A2 dual-arm robot with the ExoArm-7 exoskeleton motion capture system, enabling high-fidelity human motion acquisition and end-to-end VLA workflow execution.

Ideal for embodied AI research, robotics education, robot learning, and rapid VLA validation.

Powered by the open-source LeRobot framework, the system provides a complete pipeline for human demonstration capture, synchronized multimodal data collection, dataset generation, model training, and real-time robot inference.

Users can operate the robot via the ExoArm-7 wearable exoskeleton system, capture synchronized human motion data, RGB-D vision data, and robot state information, and process it through the control workstation/Ubuntu system running the OpenarmX-LeRobotVLA full pipeline for:

• Vision-based action learning (VLA)
• Imitation learning & reinforcement learning
• Human demonstration-based robot policy training
• Model inference and autonomous task execution

This system is ideal for universities, AI & robotics research labs, teaching labs, industrial R&D teams, and advanced robotics startups.
It supports ROS2 and NVIDIA Isaac frameworks for seamless integration, rapid prototyping, and end-to-end deployment of teleoperation and VLA workflows in controlled research and precision-demanding experimental environments.

Key Features:
• Dual 7-DOF Arms – Human-like bimanual manipulation with 5 kg rated / 12 kg peak payload.
• ExoArm-7 Wearable Control – 14-DOF high-precision motion capture with <20 ms latency for natural human demonstration recording.
• Human Demonstration Capture – High-fidelity motion mapping for imitation learning and VLA dataset generation.
• Multi-Camera 3D Vision – Intel D435 + dual D405 wrist cameras for synchronized RGB-D and point cloud data acquisition.
• High-Performance Controller – AMD Ryzen 9 Pro, 16GB RAM + 1TB SSD (optional higher configs for advanced VLA models).
• High-Speed USB & CANFD – Multi-sensor synchronization and industrial-grade robotic communication.
• Advanced Motion Planning – ROS2-control + MoveIt trajectory optimization + Gravity Compensation for stable, repeatable motion.
• VLA Pipeline Ready – ACT inference by default; supports SmoLVLA, Pi0, Pi0.5, XVLA with high-performance controllers.
• Simulation & Development Ecosystem –ROS2, NVIDIA Isaac Sim, MuJoCo, Python, C++, MoveIt, URDF, VLA, IK（inverse kinematics).

Ideal For:
• Research & Education
Universities, research labs, and robotics courses
Human demonstration collection, dual-arm manipulation studies, and embodied AI teaching
Hands-on VLA workflow exploration with real human motion data

• Rapid VLA Validation
Validate VLA workflows using real human demonstration datasets
Teleoperation data collection, AI training pipelines, imitation learning, and autonomous robotic execution
Tasks: grasping, assembly, insertion, transport, simulation-to-real experiments

Your Purchase Includes:
1x Agility A2 Dual-Arm Robot
1x ExoArm-7 Wearable Exoskeleton System
1x 48V / 20A Power Adapter
1x CANFD Communication Box
1x Head Camera (Intel D435)
2x Wrist Cameras (D405 with vision gripper assembly)
1x High-Speed USB Expansion Hub
1x Control Workstation (AMD Ryzen 9 Pro, 16GB RAM + 1TB SSD)
1x OpenarmX-LeRobotVLA Preconfigured Software Environment