Here’s what we learned at our webinar ‘Optimize Fulfillment with Collaborative Robotic Picking’

Rohma Abbas September 26th, 2018

There’s been a lot of buzz around collaborative robotic picking in warehouses. So, we recently joined forces on a webinar with DHL and Modern Materials Handling to try to define collaborative robotics and answer the most pressing questions. The webinar featured a panel of experts: DHL Supply Chain’s Adrian Kumar, VP of Solutions Design, our co-CEO Jerome Dubois and MMH’s Executive Editor Bob Trebilcock.

If you couldn’t make it, here are a few big takeaways about collaborative mobile warehouse robotics and new robotic picking strategies on the market today. Plus, we answer a few questions from attendees.

Why is collaborative robotic picking hot right now?

Warehouse automation has been around for a while, and traditional solutions are certainly efficient in boosting productivity. But a few key factors have converged into the perfect storm to give rise to the collaborative mobile robotics automation market:

  • Retail has changed, and customers’ demands are at an all-time high. Unit picking is labor-intensive, and the rise of e-commerce orders demand even more labor.
  • But labor is at a premium, as warehouse jobs continue to go unfilled and wages for these jobs rise. Demand for these jobs continue to rise.
  • Meanwhile, traditional automation has become too restrictive both financially and infrastructurally, hampering operational flexibility.
  • And finally, collaborative robotic technology is now proven, with widespread adoption across many disciplines.

Takeaway 1: There are five robotic pick techniques that you should know about. Embrace big data

In the world of robotic warehouse automation, there are five main robotic pick strategies out on the market today. There’s no one-size-fits-all approach. DHL has piloted all these strategies:

  • “Lead me” approach: The robot is integrated into the site’s Warehouse Management System. It navigates the warehouse autonomously. It leads pickers by displaying the item and quantity of the pick at each location, optimizing pick routes. In this approach, the robot-to-person ratio is typically 1.5 to 2.
  • “Follow me” approach: The robotic cart follows behind the picker, so operators don’t have to push the cart. The cart is not integrated with the Warehouse Management System.
  • “Swarm me” approach: The robot, integrated with the Warehouse Management System, waits for nearby pickers to interact with it. This technique is good for zone picking but operators need more robots per picker to make this technology work.
  • “Shuttle” approach: These technologies move products across different zones, mimicking pick-and-pass conveyor systems.
  • “Retrieval” approach: These solutions act like goods-to-person systems. The robot to person ratio is typically 10 to 1. Operators get great efficiency through completely eliminating travel, but at the cost of greater capital and less flexibility.

Takeaway 2: Some warehouses are a better fit for collaborative mobile robotics than others.

Particularly for the “lead me” approach, there are a few conditions that make some warehouses a great fit for this technology, and more likely to see ROI in under a year. These conditions in include:

  • Warehouses with manual cart operations that are labor intensive.
  • Warehouses with a variety of SKUs and a large area to walk in.
  • Warehouses with the ability to run robots for more than one shift.
  • Warehouses with pressure to get orders out the door in under two days.

This chart goes into more detail about the ideal site profile for collaborative robotics:

Takeaway 3: Flexibility is the biggest reason these solutions are gaining steam right now.

Bolting conveyor and shuttle systems is a big investment, particularly at a time when companies are moving away from large, regional warehouses and opening up “urban warehouses” to speed up delivery. Warehouse automation now needs to be more flexible to accommodate the changing landscape of warehousing. Collaborative mobile robots scale and adapt with future business needs, while traditional warehouse automation often requires its own infrastructure, including storage and movement systems, and costs tens and millions of dollars to implement.

Robots can be deployed in less than a month, at a fraction of the cost of traditional automation. And, they can be moved around between various sites within a network. Additionally, robotic solutions give warehouses the flexibility to incrementally add more automation as needed. This enables businesses to adapt to change more easily, where a traditional automation upgrade might cause major operational interruptions.

Takeaway 4: The pricing structures are extremely flexible, too.

The flexibility in pricing is similar to how forklift trucks are priced. You can buy a fleet or rent. Similarly, with robots, you can buy a fleet or rent as needed. This is very different from a capital expenditure purchase, which depreciates over time. At 6 River Systems, what we find is our customers tend to buy for their off-peak and rent during peak season.

Takeaway 5: The software on the robots is as impressive as the hardware.

At the end of the day, the robot itself rapidly becomes commoditized, so the software on the robot becomes a huge differentiator. These software components help operators manage work in the warehouse and optimize pick paths and pick techniques (discrete, batch, replenishment).

The beauty of robotic warehouse automation is that as the software on the robot gets better, the robot gets better. The same robot can become 10 percent more productive year over year as the software on it improves. This is exciting, because you don’t have to rip up or tear out your warehouse to get more out of your automation investment.

You asked, we answered: Top questions from the webinar

We received several insightful questions about robotics and particularly what we do here at 6 River System. Here are some top questions, with answers from our co-CEO and co-founder Jerome Dubois:

Q: “Does your technology work with perishable and inconsistent products? We work with local, small-farm produce, meat, dairy, etc..”

Jerome: In most of our implementations, the WMS (host system) tells us what to pick and where to pick it from. So perishability is not something we are generally concerned about. We do rely on barcoding to confirm that the item is being picked. While this step can be skipped, it will impact accuracy.

Q: “Question for Jerome, what I think what I’m hearing is that order picking in bins with low lines per order is the sweet spot for this follow-me technology. How does that compare to batch picking to a consolidation point to reduce walking?”

Jerome: “Follow-me” from the robot’s perspective, as the associate follows the Chuck through the entire process. As for pick strategies, batch picking vs. discrete order picking is always an open debate. We have customers who actually do both with our system, at the same time. During off peak, we pick discrete orders and batch singles. During peak season, we move to a batch pick to a sort wall. While this does increase pick rates by reducing walking times in-aisle, it adds additional labor and touches (sort) and lengthens the cycle time of the order.

Q: “Can Jerome/Adrian give us some idea of the efficiencies they are seeing? eg. Pick efficiency, FTE reduction etc? Can they quantify to give us an idea of the RoI?”

Jerome: It depends on your baseline. Some of our customers are already quite advanced and are replacing their cart with voice picking technology or are batch picking to unit sorters, so we are demonstrating a 30-50% there. Others are going from manual paper-based picking so we comfortably demonstrate a 3x improvement.

Q: “Would we see a robot that can do a pick independently as opposed to collaboratively with humans, anytime soon from 6 River Systems? Is 6 River doing any research in this area or is this very futuristic?”

Jerome: While there are a number of brilliant technologists working on this problem, the challenge is that whatever pick being attempted in-aisle needs to have a near 100% success rate for the technology to have wide-scale adoption. Today, these technologies do relatively well in sort functions where they are singulating multi-SKU totes to a separate destination or even sort-wall. But that problem is entirely different when trying to pick from a multi-SKU bin in an aisle. We are likely 10 years away from solving that problem consistently enough.