Google DeepMind uses a recently developed three-fingered robotic hand as an example of how cutting-edge robotics technology is used to boost AI. In this regard, the most recent advancement in robotics is a complex three-fingered robotic hand designed by UK-based Shadow Robot, well-known for its proficiency in handling robotics. This cutting-edge technology, which mimics the functions of the human hand, provides unmatched force output and precision, making it possible to manipulate objects and complete complex tasks with extraordinary efficiency.
The Shadow Dexterous Hand series, recognized as the world’s most advanced 5-fingered robotic hand, is the foundation for this new iteration. Boasting superior accuracy and versatility, it empowers researchers to explore new frontiers in robotics and artificial intelligence, driving innovation across various industries.
The newly introduced three-fingered hand demonstrates exceptional resilience to rigorous conditions, withstanding repeated abuse and external impacts. Its rapid response time, transitioning from fully open to closed in just 500 milliseconds, coupled with a fingertip pinch capability of up to 10 newtons, underscores its effectiveness in real-world applications.
Engineered with a blend of durable metals and plastics, including aluminium, brass, and polycarbonate, the hand balances strength and flexibility. This enables seamless navigation of diverse tasks while ensuring robust performance and longevity.
The integration of Shadow’s electric “Smart Motor” actuation system empowers the hand with precise force and position control capabilities, facilitating seamless interaction with the surrounding environment. With 20 motors and advanced sensors, including temperature, voltage, and current sensors, meticulous monitoring of load and joint positions is ensured, enhancing overall performance and safety.
The fundamental idea behind Shadow Robot’s approach is data-driven precision. Researchers use neural networks and reinforcement learning approaches to train the robotic hand in simulation. This allows goal-oriented actions to be executed smoothly in real-world circumstances, powered by a constant supply of accurate sensor data.
With a comprehensive suite of sensors, including Hall effect sensors and inertial measurement units (IMUs), the hand delivers unparalleled motion tracking and feedback, enabling agile and adaptive interaction with its surroundings. This data-driven approach maximizes the hand’s capabilities, ensuring optimal application performance.
