Shadow DEX-EE Hand

The DEX-EE robotic hand from Shadow Robot is a cutting-edge and versatile instrument specifically designed for the world’s leading AI research teams. This innovative hand was created in close collaboration with Google DeepMind to meet the demands of real-world machine learning projects. The DEX-EE robotic hand offers an ideal hardware platform for dexterous manipulation research, enabling dynamic and controlled movements in a robust and reliable design. This ensures that long-lasting experiments can be conducted without interruptions due to hardware failures.

Thanks to advanced high-speed sensor networks, the DEX-EE provides extensive sensor data, including position, force, and inertial measurements, as well as hundreds of channels of tactile sensing per finger. Shadow’s pioneering fingertip sensors offer hundreds of taxels with a massive dynamic range. The high-frequency torque and position control loops enable delicate and precise fingertip dexterity. Moreover, torque and inertial measurements make the entire hand sensitive to interactions with its environment.

The tactile fingertip sensors equipped with stereo cameras offer an unprecedented level of 3D interaction detail in a robust package. The multi-taxel sensors with three degrees of freedom on the middle and proximal phalanges provide additional information during grasping and manipulation.

The DEX-EE robotic hand is easy for users to maintain, requires only minimal training, and is designed to minimize failures and downtime by incorporating fail-safes and a gentle shutdown procedure. Its full ROS integration makes it an indispensable tool for research and development.

• Robustness: The DEX-EE robotic hand is designed to reliably meet the requirements of long-lasting reinforcement learning experiments. It features a long mean time to failure and a shortened repair time, leading to high availability. The robotic hand is resistant to repeated environmental impacts and intensive use from untrained handling.
• Excellent Collaboration: Shadow Robot works with customers long-term to provide support and engineering services and collaborate on development. The development of the DEX-EE has brought about significant technological advances and specialized knowledge that can be leveraged in a wide range of academic and applied research contexts.

Specifications:

• Dimensions:
  • Length: 350mm
  • Width and Height: 165mm x 160mm
• Speed:
  • Up to 180º/s per joint
• Force:
  • At least 8 N force at fingertip
• Weight:
  • Single finger: 1.2 kg
  • 3-fingered hand: 4.1 kg
• Joint Range of Motion:
  • J0 (Base Joint): -50º to 50º
  • J1 (Proximal Joint): -80º to 45º
  • J2 (Middle Joint): 0º to 80º
  • J3 (Distal Joint): -30º to 85º
• Actuation and Control:
  • Driven by five motors in an N+1 configuration
  • Joint position PID running at 1 kHz
• Sensor Channels per Finger:
  • 155 individual sensor channels
• Tactile Sensors:
  • Distal tactile sensors with a stereo camera pair viewing the inside surface of the silicone skin
  • Middle and proximal phalanges equipped with modular tactile sensing regions
• Communication Interfaces:
  • EtherCAT (100 Mbps Ethernet-based communications field-bus)
  • USB 3.0 for distal tactile sensor feed
  • Dual FAN bus within the finger (3 Mbps bus architecture)
• Power Supply:
  • Requires 48V 200W power supply

For comprehensive advice, please feel free to contact us at any time.