DC robotics: the future

by Dave Luton
Dave Luton is a consultant in the Greater Toronto Area.
Dave Luton is a consultant in the
Greater Toronto Area.

The employment of warehouse robotics will ultimately depend on the evolution of the warehouse itself.

From a supply chain perspective the biggest change to outbound supply chains is the continuing diversification of sales channels. Thus warehouses now have to have the capability to support the following sales channels/order selection types:

A    Traditional wholesale direct to warehouse: pallet/
layer pick
B    Retail direct to store: case/less than case pick
C    E-commerce direct to end consumer: less than case

Accompanying this diversification of warehouse order picking needs are related pressures caused by the demand for free shipping for ecommerce and the need to reduce order turnaround time.

To support these multichannel needs, warehouses are evolving into order fulfillment centres. There is a potential role for robots in all of these differing distribution channel requirements.

Robotic applications to support traditional wholesale warehouses are a mature technology and have been around for 50 years. They include robotic technologies like AS/RS (automated storage and retrieval systems) for order picking and AGVs (automated guided vehicles) for horizontal transportation.

Initially designed for pallet-in/pallet-out order selection requirements, newer AS/RS systems can handle layer section needs.
Historical fixed-path AGV horizontal transportation systems have been upgraded through the use of new laser range-finding sensing technology to locate origin and destination and to avoid obstacles. These now have true mobility and can function autonomously.

Even for these traditional systems, further robotic upgrades are made possible through the use of new technology. Consider robotic devices like drones for such tasks as inventory counting and command and control trouble shooting if problems occur. The old inventory count sheets can be replaced by a video record of the count.

For layer pick, case pick and less-than-case pick applications different robotic technologies are employed.

There are four types of common robotic technologies and each is commonly used in different warehouse applications.

The common applications include:
Stationary Articulated Robots—These are the familiar robots from assembly line applications (e.g. painting) or in manufacturing—palletizing applications.
Many warehouse palletizing needs are different from traditional manufacturing (except for floor-loaded receiving) because of the need to build multi-SKU pallets. Historically this has proven more difficult because traditional robotic applications used sensing technologies like photo eyes that used electrical imputs and outputs.
Newer sensing technologies combined with processing algoritiums to optimize trailer loading (e.g. more stable pallets combined with optimal truck axle loading and route planning) offer the promise of resolving this issue in future.

Gantry Robots—These are used in case selection/layer picking applications in which SKUs are picked  from overhead using vacuum lifting heads.

Robotic Arms—Long-term these offer the potential to  pick from shelves.

Mobile Robots—These are used for horizontal transportation applications similar to older generation AGVs. They use machine vision to provide better warehouse applications. Amazon’s Kiva technology is a good example where the improved sensors combine with processing software to move the stock on shelves to the picker/packer.

Continuing safety enhancements is another key evolution in warehouse robotic applications which permit robots to work alongside humans. Earlier versions required segregation via cages with lockdown controls. The newer versions involve improved sensing (sonar, machine vision or lasers) combined with torque sensing motors. Safe robotic design such as no sharp edges is also important, along with software that slows the speed down when humans are in close proximity.

In this environment each species (dare I say that) will do what each does best. Robots are good at repetitive, unpleasant and ergonomically challenging functions that result in injuries, therefore no more human saying “Oh, my back.” Humans will perform the actions that require agility and variation.

This was forecast over a decade ago in a column I wrote entitled “Meet “Hal”, your new warehouse manager”.

Even so-called state-of-the-art systems can be improved. An obvious opportunity is the Kiva example mentioned earlier, where while the horizontal transportation has been automated, the shelf-picking and packing look as though they might become a possible robotic application.