Automatic Parts Detection: Solving Complex Handling Tasks with AI

Cambrian Robotics will showcase its 3D vision system for robots at this year’s Motek. Thanks to artificial intelligence, the robot can detect objects with the help of a camera and execute highly complex tasks.

This makes it possible for us to implement a broad range of automation tasks in the production environment. Our customers include larger SMEs, tier 1 suppliers and OEMs. Typical tasks involve, for example, bin picking and multi-bin picking, cable handling including insertion or plug-in, precise feeding of parts to machines, assembly and placement tasks, as well as extensive materials handling and kitting operations – e.g. in combination with storage systems such as Kardex. Production steps that could only be performed by means of manual labour in the past can be automated with our system.
We unite a number of technological features, thus making it possible for us to automate difficult production steps. These features include 1) insensitivity to light – our system does not require structured light, 2) the range of parts – we can reliably process transparent, shiny and black parts as well, 3) speed – part, AI and image recognition takes less than 180 milliseconds, which is the fastest system on the market and 4) accuracy – we work with CAD parts data, so gripping and placement are very accurate.
We’re already working with manufacturers such as ABB, KUKA, UR, Neura, Denso, Doosan and Yaskawa.