Demand for e-commerce in the United States is soaring, and the labor shortage is not going away. One way that logistics companies can mitigate these problems is with better mobile robot automation. Many warehouses are now starting to think about full workflow automation.
Achieving this requires integration of a combination of specialist mobile robots that each do different things. But those who seek to go down this route quickly encounter a problem: few vendors offer a full range of robots that can do it all. Even if they do claim to do it all, no single vendor offers the best product for every specialized warehouse application. The ideal situation from the customer’s point of view is to choose 'best-of-breed' products from different manufacturers and integrate them. Finding a software platform that is sophisticated enough to do this effectively is a major challenge.
One reason that now is the time to consider integrating robots from multiple vendors is simple: mobile robots have matured. The market currently abounds with cheap, capable, and reliable mobile robots. Integrating products from multiple vendors is the logical next step.
Another reason logistics managers can now consider integrating robots from multiple vendors is that there have recently been major advances in the necessary planning software technology. Multi-robot planning is a hard problem to solve, but it is a problem that has been well studied by artificial intelligence (AI) scientists. A lot of relevant progress stems from video game development. In video games, you have multiple agents, and the challenge was to try to plan their movements. That problem was cracked 10-20 years ago, and this technology has now been adapted for real-world robots.
Working with robots requires considering the nitty gritty details that aren’t a factor in games. Issues such as machines breaking down and how to manage acceleration. A lot of work has been done on that in the last five years, but it was mainly applied to homogenous robot fleets (where all the robots were of the same type and from the same vendor). The other big breakthrough, which is even more recent, has been successfully applying multi-robot planning technology to a heterogenous fleet of robots from different vendors. This is exponentially harder, but it is a problem that has now been solved.
Are customers ready for it?
The wave of adoption of heterogenous multi-robot fleets is only just beginning. Many mobile robot customers haven’t yet automated on a large enough scale to encounter the tough problems inherent in multi-robot planning. But the largest mobile robot users have tried to move things forward and undertake workflow automation combining “best-of-breed” robots from multiple vendors, and they’ve encountered big problems while trying to do it.
Logistics companies are great at managing humans but trying to integrate diverse mobile robots into warehouse operations takes a special set of robotics expertise and tends to require a radical reimagining of the workflow. Often, third-party advice is required, and the mobile robot vendors may not be the most appropriate people to turn to: they are experts at making very good robots, but they have less expertise at integrating their robots with those of other vendors. Specialist mobile robot software companies and integrators are better suited to this task.
Consider an automated production line. No one would expect every machine and robot on the line to come from the same industrial automation vendor. A specialist integrator comes in and combines varied products into a bespoke solution tailored to the needs of the end user. The highly automated warehouses of the near future will be no different. Mobile robot vendors will deliver specialist segmented solutions and then integrators will combine them with other solutions to create automated workflows.
As the big mobile robot users crack the problem of full workflow automation, the efficiencies they realize will push the entire market in this direction. The adoption of heterogenous fleets of mobile robots will then happen rapidly. This will result in a plethora of unique mobile robot solutions.
How heterogenous fleets are managed
Some current solutions work by connecting existing robot systems. This is done by functioning as a higher layer between the robot’s control system and the warehouse management system. But it is difficult to attain high levels of optimization in this way.
The best solutions actually integrate the robots themselves: talking directly to the machines. You can think of this as essentially building a unified robot control system which optimizes for the heterogeneous robot fleet, together with a warehouse control system that optimizes the business workflow. Different robots use different guidance systems and talk different “languages.” One of the capabilities needed to enable seamless mixed fleet orchestration, allowing robots to see and talk to each other and to collaborate effectively, is a “universal translator,” an open platform that can coordinate with robots from any vendor.
The best mobile robot logistics software solutions are already offering this capability and the benefits are strong.
Forward-thinking warehouse managers shouldn’t wait for their competitors to act first: now is the time to consider integrating large and diverse fleets of mobile robots into fully automated workflows.