The university of California, Berkeley, has developed a highly flexible industrial robot that can be found in warehouses, factories, hospitals and family Settings.
The robot, developed by Ken goldberg and one of his graduate students, has a number of skills, such as sorting your drawers, with amazing speed and skill.It picks up objects based on a software called dex-net, which can pick up objects in a virtual environment and train deep neural networks by trial and error.
The latest version of the system includes a high-resolution 3-d sensor and two robotic arms, each controlled by a different neural network.One arm is equipped with a traditional robotic claw, and the other has an inhalation system.The robot scans an object through software and then looks at two neural networks to determine whether a particular object is captured or sucked in real time.If you're not sure how to grab an object, the robot will push the object to better observe it.
The researchers have also developed a way to measure robot grabbing performance: a measure known as "average hourly crawl".This indicator will help the lab better share their results.
In five years, goldberg says, he expects the robot to be "more efficient than the average human being" per hour.