Page 8 - Logistics News Sept/Oct 2017
P. 8

Feature


               In the future, warehouse robots



                            will learn on their own




                                                                               By Cade Metz, courtesy New York Times

               The robot was perched over a bin fi lled with random objects, from a box of instant oatmeal to a
               small toy shark. This two-armed automaton did not recognize any of this stuff, but that did not
                matter. It reached into the pile and started picking things up, one after another after another.









































            The robot at Berkeley can pick up irregular objects it has never seen before, like a toy shark.
            Credit Jason LeCras for The New York Times


            “IT FIGURES out the best way to grab each          distribution centres — where sorting through stuff
            object, right from the middle of the clutter,” said   is the primary task — armies of humans still do
            Jeff  Mahler, one of the researchers developing the   most of the work.
            robot inside a lab at the University of California,   The Berkeley robot was all the more
            Berkeley.                                          remarkable because it could grab stuff  it had
               For the typical human, that is an easy task. For   never seen before. Mr. Mahler and the rest of the
            a robot, it is a remarkable talent — something that   Berkeley team trained the machine by showing
            could drive signifi cant changes inside some of the   it hundreds of purely digital objects, and after
            world’s biggest businesses and further shift the   that training, it could pick up items that weren’t
            market for human labour.                           represented in its digital data set.
               Today, robots play important roles inside           “We’re learning from simulated models
            retail giants like Amazon and manufacturing        and then applying that to real work,” said Ken
            companies like Foxconn. But these machines are     Goldberg, the Berkeley professor who oversees
            programmed for very specifi c tasks, like moving a   the university’s automation lab.
            particular type of container across a warehouse or    The robot was far from perfect, and it could
            placing a particular chip on a circuit board. They   be several years before it is seen outside research
            can’t sort through a big pile of stuff , or accomplish   labs. Though it was equipped with a suction cup
            more complex tasks. Inside Amazon’s massive        or a parallel gripper — a kind of two-fi ngered hand


        24                                                                 September/October 2017  |  Logistics News
   3   4   5   6   7   8   9