Autonomous pick­ing with­out mas­ter data

New soft­ware fea­ture: Mag­a­zino robots now pick indi­vid­ual items with­out prior knowl­edge of their dimen­sions.

Press releases

29. January 2020


Munich, Jan­u­ary 31, 2020 — The Ger­man robot­ics com­pany Mag­a­zino presents a new soft­ware fea­ture at the Logi­MAT 2020 show. The mobile robot TORU is now able to pick indi­vid­ual items with­out mas­ter data on dimen­sions. With the new capa­bil­ity, TORU can pick and stow indi­vid­ual items, like shoe boxes, even if only incom­plete or no mas­ter data (e.g. width, height and the length of an item) is avail­able. The robot now deter­mines this data itself and can also feed it back into the cus­tomer’s ware­house man­age­ment sys­tem (WMS). The new soft­ware fea­ture is part of the Advanced Coop­er­a­tive Robot Oper­at­ing Sys­tem (ACROS).

Fre­quently, in e‑commerce ware­houses there is incom­plete or incor­rect mas­ter data avail­able for cer­tain items, such as shoe boxes. Thus, the dimen­sions of a cer­tain items are not exactly doc­u­mented in the WMS. This is crit­i­cal not only for opti­mal space uti­liza­tion and the pack­ag­ing process at the ship­ping sta­tion, but also for autonomous pick­ing with mobile pick­ing robots. The less a robot knows about its envi­ron­ment in advance, the greater the chal­lenge to ful­fill its order quick and reli­able.

TORU in use alongside humans
The mobile robot TORU deter­mines object sizes with its own sen­sors and can there­fore pick items with­out com­plete mas­ter data.


Against this back­ground, Mag­a­zino has devel­oped this new fea­ture for its robot TORU: When pick­ing shoe boxes, TORU com­bines the data of an addi­tional sen­sor in the grip­per with the var­i­ous infor­ma­tion from its 3D cam­era. This means that the robot can now auto­mat­i­cally cap­ture all dimen­sions (width, height, length/depth) of an arti­cle while grasp­ing it. This not only enor­mously increases pick per­for­mance, robust­ness and process reli­a­bil­ity for unknown arti­cles – it also extends the robot’s pos­si­ble range of tasks.

The newly acquired object data is not only used by the respec­tive robot itself, it is also stored in a local cloud via ACROS. This means that the newly acquired mas­ter data can also be used by other robot col­leagues in the fleet and, more impor­tantly, can be sent back to the ware­house oper­a­tor’s WMS via the cloud. This allows the oper­a­tor to sup­ple­ment his data­bases and com­plete incom­plete or incor­rect mas­ter data.

These new capa­bil­i­ties are part of the Advanced Coop­er­a­tive Robot Oper­at­ing Sys­tem (ACROS). ACROS is one of the most com­pre­hen­sive soft­ware stacks for autonomous robots and was devel­oped by Mag­a­zino. ACROS func­tions as the oper­at­ing sys­tem of the Mag­a­zino robots, which com­bines the indi­vid­ual soft­ware mod­ules such as autonomous nav­i­ga­tion, order man­age­ment, image recog­ni­tion, com­mu­ni­ca­tion with the cloud and machine learn­ing. ACROS is hard­ware-inde­pen­dent, already in use on dif­fer­ent Mag­a­zino robots and can also be oper­ated on third-party hard­ware such as auto­mated guided vehi­cles (AGVs) or fork­lifts.

Mag­a­zino GmbH devel­ops and builds intel­li­gent, mobile robots that per­ceive their envi­ron­ment and make their own deci­sions. These autonomous robots work along­side peo­ple and make processes in e‑commerce, fash­ion and pro­duc­tion logis­tics more flex­i­ble and effi­cient than ever before. With over 110 employ­ees in Munich, Mag­a­zino is the largest Advanced Robot­ics team in Europe. Magazino’s investors include Kör­ber AG, Zalando and Fiege Logis­tik.

Down­load PDF, Word & pic­tures

Press kit 3MB