AI-Powered Robots Learn Complex Tasks Through Human Training On Factory Production Lines

As published in WIRED:

This innovative approach enables two-armed robots to learn complex tasks through human training and real-world practice on factory production lines. AgiBot’s system, which combines teleoperation and reinforcement learning, is currently being tested on the production line of Longcheer Technology, a prominent Chinese company that manufactures smartphones, VR headsets, and other electronic gadgets.

The integration of AI and human workers is transforming the capabilities of industrial machines, and this trend is expected to increase manufacturing productivity while potentially reducing the need for low-wage human workers. However, this shift may also lead to the creation of new job opportunities. Robots are already widely used in factories for tasks such as lifting boxes and moving bins, but they often struggle with complex assembly tasks that require dexterity, sensing, and adaptation.

AgiBot’s representative, Yuheng Feng, explains that the robot deployed at the Longcheer plant performs a specific task, taking components from a machine that performs testing and placing them onto a production line. This type of task is suitable for robots, as it does not require fine manipulation or handling of fragile or ← →