Shadow Robot, a renowned robotics company with a glowing reputation for its groundbreaking robot designs, has recently partnered with Google’s DeepMind to create a new, robust robot model tailored for machine learning research. The company, known for its high-profile clients such as NASA, ESA, Google, OpenAI, MIT, and several universities, has over 20 years of experience in delivering quality robot designs.
Its dexterous robot hand demo at Amazon re:MARS with Jeff Bezos is a testament to its technological prowess. With this new innovation, Shadow Robot is poised to revolutionize the field of machine learning research.
The latest version of the world’s most dexterous humanoid robot hand looks quite different from previous iterations. The new Shadow Hand, which was created by a company in collaboration with the Google DeepMind robotics team, features only three fingers in a gripper-like formation. This design was built with speed, flexibility, and precision in mind, but also with the ability to endure a significant amount of misuse.
According to the company, one of the biggest challenges in AI and robotics is to create hardware that is dexterous enough for complex tasks but also robust enough to withstand the trial-and-error process required for robots to learn. The new robot hand is designed to be able to test things in the real world, sometimes executing motions at the limit of their abilities, without causing damage to the hardware.
The robot hand has precise torque control and each of its fingers can exert up to 10N of fingertip pinch force. And with rapid movement speeds, the fingers can go from fully open to closed in just 500 milliseconds. The robot itself is compact and lightweight, measuring in at 350mm in length, 165mm in width, and 160mm in height. And with a total weight of just 4.1kg, it’s easy to maneuver and control.
The weight of a single finger is 1.2kg, while the whole hand weighs 4.1kg. It also requires a 48-V/200-W power supply.
The robot’s fingers are designed to provide a unique and advanced sense of touch. Each finger functions as a self-contained unit, hosting an array of 3-DOF tactile sensors and a stereo camera setup that captures high-resolution, wide-dynamic-range feedback in real-time.
The silicone skin covering the fingertip enables the robot to interact with the world around it through touch, providing a more intuitive and responsive experience. This cutting-edge technology empowers the robot to be more efficient and effective in its tasks, making it a valuable asset in any setting.
The Shadow Hand’s finger modules are designed to withstand limit-pushing AI experiments. However, in the rare event of fatal damage, the modules can be easily removed and replaced, minimizing downtime. What’s more, the tactile sensors are also interchangeable, allowing for seamless repairs.
With such advanced technology, the Shadow Hand is a game-changer in the robotics industry. Don’t miss your chance to witness its debut at ICRA 2024 in Yokohama, Japan. Be among the first to experience the future of robotics.
Additionally, the tactile sensors can also be removed and replaced as needed. I wonder how much the Shadow Hand will cost, though. According to the information available, the company will showcase it for the first time at ICRA 2024 in Yokohama, Japan.
