ABB Robotics and PSYONIC Collaborate to Advance Robotic Dexterity Using Human Data

ABB Robotics has announced a strategic collaboration with PSYONIC, a bionics company based in California, focused on advancing the field of robotic dexterity. The partnership aims to improve how robots handle objects by utilizing a novel approach: incorporating real-world manipulation data derived from the use of human prosthetics. At the core of this initiative is the integration of PSYONIC's Ability Hand with ABB's GoFa™ collaborative robot. This setup will allow researchers to collect and analyze touch and motion data generated by prosthetic users performing everyday tasks. This rich dataset will then be used to train robots, enabling them to execute delicate and variable tasks that have traditionally been difficult to automate due to their complexity and the need for nuanced interaction. Marc Segura, President of ABB Robotics, highlighted the significance of this collaboration, stating that replicating human dexterity and the intuitive understanding of object handling is one of the most challenging aspects in industrial robotics. He emphasized that this is a fundamental requirement for developing truly autonomous and versatile robotic systems. The project is seen as a crucial step in the development of the next generation of physical AI, where robots will learn to perceive and interact with the world in a manner similar to humans. The ultimate goal is to close the long-standing gap between human and robot dexterity. By doing so, the collaboration expects to unlock new possibilities for automation across a wide range of industries. This could lead to robots capable of performing more intricate assembly, handling fragile materials, or even assisting in environments where human-like precision and adaptability are paramount. The use of human-generated data from prosthetics offers a unique and highly relevant source of information for teaching robots fine motor skills and adaptive gripping strategies.