Amazon’s Improved Algorithm Aims to Enhance Human-robot Collaboration

The latest collaboration between robots and humans is occurring in Amazon Robotics fulfillment centers. Recent research from the Institute for Operations Research and the Management Sciences highlights Amazon’s redesigned algorithm that controls the flow of goods via robot carriers to human workers who have been tasked with picking and packing orders. 

 

A mobile robot transports goods to workers. Image used courtesy of Allgor et al.

 

Robotic Picking Algorithm 

Instead of workers having to walk through the warehouse, locate each item, and pick it up (perhaps having to reach a high shelf or stoop to a low shelf), the workers can remain at their workstations and have the goods delivered to them. Ultimately, this system reduces the number of steps workers must take and limits lifting injuries compared to a traditional warehouse.

So far, this improved algorithm has paid off. Early results show that average worker travel has been reduced by 62%. The fulfillment time has also increased, as the robots can cycle through different workstations with goods as they maintain the list of all items required. From there, artificial intelligence (AI) optimizes the path of these robots to deliver the right goods to the right workstation at the right time, all while avoiding collisions with workers, objects, and other robots.

 

A worker picks the items from shelves delivered by robot. Image used courtesy of Allgor et al.

 

Robot vs Human?

Any discussion of new robotics capabilities can cause some worry for some workers. The age-old question “will a robot take my job?” is raised once again. There is always debate as to whether automation is good or evil in some moral sense. Perhaps a better question to ask is whether automation improves life or detracts from it. 

Tasks that are dangerous, repetitive, require high consistency or accuracy, or are otherwise boring or risky can and should be automated. Minecarts used to be stopped by hand, sometimes by children jamming wood blocks in the wheels—now they are stopped by a computer-controlled system; very few collisions and no children losing fingers. 

In the case of the latest robots, it’s clear that they improve life. Reaching, lifting, stooping, and carrying loads lead to musculoskeletal injuries, both in the aging worker and in the young worker who thinks they can do it all. While there are safe lifting trainings, it’s too tempting to carry a box a little farther rather than move the cart or carry too many boxes instead of making multiple trips. Robots will never do that—they will faithfully deliver packages to the worker at the right time and height when needed.
 

Advancements in robot-human collaboration opens up new possibilities for both robots and humans. Image used courtesy of Mike Mareen - stock.adobe.com

 

The Future of Automated Shipping

Perhaps the next step will be to automate the picking, packing, and shipping operations fully. This would cut humans out of the loop entirely, with robots picking the items, placing them in packaging materials, sealing the boxes, addressing them, and sending them out for delivery. In this case, some jobs may be lost at the fully-automated facility. 

Never fear, as there are always new opportunities. The high cost of entry to automation means there will always be a need for handling, picking, and packaging, whether as a supplement to automation systems at a large facility or a smaller startup company. Furthermore, with automation comes the opportunity for upskilling. Many of these systems require specialists to install and maintain the system, and there is a strong need to keep these systems running long after warranties and service agreements expire. 

Finally, and perhaps the most important point, any advancement in robot-human collaborative efforts will open new possibilities for both robots and humans. The cliche in education is “you’re training for the job that doesn’t exist yet,” and this is a cliche that definitely applies to the field of automation. Amazon has made some breakthroughs in the AI and cobot space, which will lead to new opportunities that have yet to be discovered.