Boston Dynamics gives Atlas humanoid new four-finger hand for complex industrial tasks

韩国先驱报正文已收录本站
Boston Dynamics’ Atlas humanoid robot uses its four-finger hand to manipulate a power drill in a demonstration. (Boston Dynamics)
Boston Dynamics’ Atlas humanoid robot uses its four-finger hand to manipulate a power drill in a demonstration. (Boston Dynamics)

Boston Dynamics, the robotics arm of Hyundai Motor Group, unveiled a robot hand with four fingers on Thursday. The hand was developed to help its humanoid robot Atlas perform continuous and complex tasks.

Humanoid robots are expected to work in spaces and use tools designed for humans, making versatility and flexibility important. Robotic hands are a key component in the capability.

The new hand has 13 degrees of freedom across four fingers, with four in the thumb and three in each of the other fingers. Pressure-based tactile sensors on the fingertips and palm detect subtle changes in contact force, allowing the hand to adjust finger movements and grip strength based on an object’s shape, material and contact conditions.

In a video released by Boston Dynamics, Atlas inserted a drill bit into a drill, pressed the trigger and drilled a hole in wood. It also tightened a small nut by rotating it between its fingers, spun a thin drumstick and simultaneously manipulated two golf balls in its hand.

The four-finger design was developed based on industrial applications, considering design complexity, power consumption, cost efficiency, durability and maintenance. The hand is roughly human-sized, with modular components and a simplified structure.

“We tested various designs, including the number of thumbs and whether a pinky was necessary, through experiments, simulations and 3D-printed mockups,” said Alberto Rodriguez, head of Atlas behavior manipulation technology at Boston Dynamics.

Boston Dynamics also used a Sim-to-Real approach, training the robot in virtual environments before applying the results to a physical machine. The simulations incorporated hardware characteristics and friction while varying motor torque, object shapes and external disturbances for reinforcement learning.

The company recently opened the Robotics Metaplant Application Center, which replicates Hyundai Motor Group’s manufacturing environment. The center will collect Atlas’ manufacturing data and validate its learning through data collection, training, field application, feedback and retraining.


jyoungseon8@heraldcorp.com