SOMA Robotics launches mobile dual arm robot for collecting training data

Robotics firm SOMA Robotics has announced a new dual armed hardware system called soma one, designed to help researchers teach artificial intelligence how to handle physical tasks. The mobile machine comes fitted with two mechanical arms, depth sensing cameras, and a moveable base. Humans can operate the robot remotely using virtual reality gear, allowing the system to log precise physical movements and video that can be directly fed into computer learning models.

For smart robots to assist with everyday household chores or factory work, they need massive amounts of real world practice data. Gathering this physical data has historically been slow, expensive, and difficult to standardize across different research facilities. By offering a standardized hardware package with built in recording software, SOMA Robotics aims to reduce the setup time for labs testing physical automation, which could eventually speed up how quickly capable helper robots enter workplaces and homes.

Full technical specifications and pricing details remain unannounced until the planned September launch, according to the company. While reducing setup barriers is a helpful step for academic labs, it remains to be seen whether a single standardized hardware setup can capture enough varied real world scenarios to train truly adaptable AI agents. Additionally, real world environments present unpredictable obstacles that rigid laboratory data collection might fail to anticipate.