Solution
Autonomous Robots require Less Programming
With the ability to „see“ and “understand” , the robot can plan and execute many tasks autonomously that previously required laborious manual programming.
Applications
Application
Depalletizing of disarranged work pieces
Problem:
The robot is not able to locate the work piece on the palette, e. g., because the objects were disarranged during transportation. Manual intervention is necessary to restart the process. The unknown, random location of the work pieces can neither be hard-coded in the robot program nor taught in.
Application
Error-prone feeding systems
Problem:
Expensive feeding, separation, and centering devices are used to pass a single work piece to a robot at a defined place with a defined orientation. Not only are mechanical devices prone to clogging, they are also highly inflexible, being tailored to a specific product with high effort.
Application
Sorting and consignment of different products
Problem:
The robot must distinguish between different types of objects and sort them according to class (e. g., into packages or in bins with the same type of work pieces). This task becomes difficult and complex if the robot has to differentiate between high number of different objects.
Application
Machine tending from bulk
Problem:
The work pieces are laying randomly on piles or in bins. Naturally, their position and orientation is unknown to the robot without sensing. Robots are not able to pick from the bins directly and expensive machine feeding systems need to be used. But more often, this repetitive and exhausting task is assigned to human workers.
Application
Recognition of deformed objects in quality control
Problem:
Visual inspection for shape defects in the realm of quality management is typically carried out by a human workforce. Manual visual inspection is a laborious process, thus throughput is limited, and errors remain undetected.