A groundbreaking development from Crailsheim demonstrates what happens when the productivity of traditional industrial robots meets the high integration capability of cobots. The tog.519 can be quickly put into operation, requires little programming effort and picks up objects with high precision thanks to AI – and maybe best of all, it does it all tirelessly.
A success story
“tog” – the name says it all at Schubert. The abbreviation – borrowed from the English word ‘together’ – stands for an entire range of automation solutions that make work easier for human operators. Schubert began its journey towards automated collaboration in 2019 with the tog.519: an international team of talented engineers designed the pick & place cobot in their own start-up. The tog family now also includes the tog.101.
Two worlds, one machine
The tog.519 is not a collaborative robot in the true sense of the word. It works in a protective cell to ensure that its high performance remains constant. Classic cobots reduce their speed as soon as a person approaches. The tog.519 nevertheless shares the cobot’s ease of integration and quick commissioning. When picking and placing products of all kinds and from all industries, it runs at up to 90 cycles per minute, matching the high performance of classic industrial robots.
Productive thanks to the SCARA principle
Schubert has gained a wealth of experience with robot-based solutions for pick & place applications since the 1980s. With the tog.519, Schubert has added a fifth axis to the SCARA kinematics, making the cobot even more agile. It not only picks up objects straight on, but also at an angle, and places them precisely into the designated trays.
Detection with artificial intelligence
The cobot is exceptionally capable, and not only because of its five axes. A combination of two 2D cameras provides it with images of the products to be picked. It evaluates the images using a neural network and assigns the products on the conveyor belt to packaging categories. Engineers previously ‘trained’ the network with countless data sets. This is why it can also recognise the positions of individual products – even when they are in an unsorted random pile.