Presented by:

December 04, 2018      
Matt Wicks

Good help is hard to find. But e-commerce demand keeps growing, with associated distribution demands accelerating at a rate of 25% annually. Addressing this is a challenge, as 80% of distribution centers operate manually — but this likely won’t last. Industry growth outpaces the labor pool by a factor of 6 to 1.

These conditions would indicate the liberal deployment of robotics as companies strive to close the gaps in their labor pool. However, that is not the case. Why?

The dynamic, unstructured environment of distribution centers has several key differences from the manufacturing facilities where robotics have historically thrived. E-commerce means high product variability, intense growth, and, above all, change. This poses a multi-faceted challenge to robots, putting their work in distribution centers at odds with the predictable, repetitive tasks of manufacturing environments.

From variable package types and sizes to new items and seasonal surges, robots must be able to adapt and improve over time — with minimal manual intervention. Fortunately, today’s logistics robotics market is evolving faster than ever, with a confluence of innovative technologies enabling a higher level of capability and performance than ever before.

­Innovation comes to distribution centers

Humans have many inherent characteristics that make them a great fit for various tasks in distribution centers. We humans have the most advanced perception, cognition, and motion-planning systems in the world. They make general tasks like washing dishes, brushing teeth, and, yes, picking orders and unloading freight in distribution centers seem so easy.

Robots, on the other hand, have historically lacked the ability to generalize beyond a specific use. Vision systems can be designed to identify a specific object, and motion-planning algorithms can enable a certain function. But change the object, task or tool the robot has to work with, and historically, the system would fail.

It all comes back to the robot’s “brains” — the combination of sensors, data science, and machine learning that enables the robot to make the best use of available tools when executing a certain task.

Advances in perception and processing power are enabling robots to better perceive their environment and more precisely manipulate gripping tools. Data science and machine learning assume the mantle of answering question of how – how can the robot position its tools to grasp and move objects, how much force to apply and more.

Honeywell Intelligrated palletizing robot

Source: Honeywell Intelligrated

Handling environments with the amount of variation of a distribution center requires a lot of data. While a robot can be programmed to reliably handle cases or items of the same size and mass, introducing another variant would have previously required another development cycle to adjust the algorithm.

However, a data-driven machine learning approach pulls data from a wide variety of package types, sizes, items, and tasks to train the model. In practice, this allows the robotic systems to recognize and handle new items or packages, even if it has not processed them previously. With e-commerce throwing variable carton sizes, polybags, apparel, electronics and more at robots in distribution centers, these attributes are critical for success.

Integration — putting technology to good use

With such high growth and intense competition, e-commerce logistics operations are finely tuned machines designed to maximize service levels and control costs. As such, introducing new technology and processes requires a comprehensive view of the entire operation to make sure interdependent systems deliver top-line benefit.

Everyone from established industrial companies to hot new startups can bring robots to distribution centers for a specific task. However, that alone is no guarantee of success. Finding a solutions provider that pulls together the distribution fulfillment experience and robotics expertise is a rare but critical piece to the e-commerce puzzle.

Success requires putting new, smart robotics processes in the context of established workflows, and seamlessly integrating them with other automated – and manual – components.

Building on the right foundation

There’s more technology than just robots hitting distribution centers. They’re becoming increasingly connected, with an exploding amount of data available to better inform management decisions, track and optimize processes, and, more recently, inform machine learning.

Making the most of robotics in the connected paradigm starts with the processing power to handle the volume of data available in real time and ensure consistent performance, rather than leave systems susceptible to paralysis by analysis.

Beyond that comes a growing layer of added value, including connected maintenance services to ensure system health, and the crown jewel of contemporary artificial intelligence, machine-to-machine and site-to-site machine learning. This capability allows robots to handle never-before-seen circumstances successfully and to continuously improve — no additional programming or intervention required.

Driving business benefit in distribution centers

Retail has always featured tight margins, and with competition only a click away, e-commerce brings new pressure to keep costs down and service levels high to attract — and retain — online shoppers. This leaves little room for ill-fated robotics investments.

Keeping the characteristics of innovation, integration and connectivity in mind are critical. Working with the right partner can ensure a system that covers all these bases. Trade organizations like the Robotic Industries Association can aid in vetting partners, offering certification programs, events, and other resources.

Matt Wicks, vice president, product development, Honeywell Intelligrated

About the author:

Matt Wicks is vice president, product development, at Honeywell Intelligrated.