Sagawa Tests Deep Learning AI Parcels Diagnostics Solution

Sagawa Express and artificial intelligence solutions developer Automagi collaborate to develop a Deep Learning AI Solution that measures and evaluates parcels.

Sagawa Express and Automagi announced on 31 May 2017 that they undertook the testing of a solution that uses deep learning AI (artificial intelligence) to not only measure parcels but to distinguish its shape, the presence or the lack thereof damage, and handling of parcels, a project approved by the Ministry of Economy, Trade and Industry and awarded to NTT Data Co., Ltd. 

Objective: a solution to help suppliers keep up with the growing demand of parcel deliveries

When it came to light in February 2017 that Japan's No.1 parcel deliveries company, Yamato Transport Co., Ltd.'s Union was asking management to reduce the number of parcels the company was taking in, shockwaves hit Japan. In no other country has domestic parcels delivery demands have grown to a point where service providers physically experienced saturation and overflow. Yamato has been trying desperately to keep up with surging demand by hiring more staff and outsourcing where necessary. But when reports of suicide from overworking and pressure and hours of unpaid overtime work would not stop, even management had to say, enough is enough. 
Against this backdrop, NTT data proposed that the labor-intensive operational processes in parcel delivery get some help from AI-driven automation. The company believes that by applying the latest deep learning AI engines, such tasks as acceptance into the warehouse, sorting, and even loading and unloading trucks can be automated as just a start with many more applications to be developed thereafter. 

Image Diagnostics Tested at Sagawa Sorting Center

Sagawa's role was to install the proposed image diagnostics device in its sorting center. The company hopes that in the near future, such solutions will enable it to reduce manpower at sorting centers as securing staff has become and will increasingly be more difficult as demand continues to grow rapidly.
At the front line of parcels logistics, the diversity of the shapes and sizes of parcels that are handled has traditionally made it a challenging, or daunting,  segment to automate. NTT Data's solution can automatically identify up to 1,000 different types of parcels for size, shape, how they are to be handled (fragile, "this side up," and so on), volume, and whether they are damaged or not. To date, such "diagnostics" of parcels is predominantly done by human beings. Once the parcels can be correctly identified,  such data can then be applied to create automated processes for loading and/or offloading, inspecting, and packing, says NTT Data, and thus, will lighten the burden on the drivers. 

Meanwhile at Yamato...

Yamato's Chronogate already has a robot arm that offloads parcels from cages (0:30 ~) and 3D scanners that scans parcels as they move on conveyer belts (the cross belt sorter 2:01~) that automatically sorts them.
Youtube Video: Daily Cargo published 1 October 2013


Post a Comment