Deep Learning Robotics’ patent pending technology revolutionizes user-robot interaction and robot-to-robot collaboration. The Deep Learning Robotics device captures 3D images of a user performing the intended task on a specific object. With the ability to differentiate between the human hand and the object and using advanced Artificial Intelligence machine learning methods such as neural networks, the sequence is automatically translated into robot control language. This intuitive task-teaching process is independent of any specific robot hardware.
The simple and cost-effective Deep Learning Robotics implementation breaks down the current barriers to widespread adoption of robots in low-volume, highly versatile production lines.
The graph below shows a sequence of points in time and space that were detected automatically by tracking the hand of an expert user while she was performing a dip soldering task. After optimization, the coordinates and elapsed times between points were transferred to a robot, which then performed the same task without user intervention