Some lessons from robotic picking
If the last 10 years have been revolutionary for the Supply Chain industry, the next decade will be even more dynamic. According to a study by Reportlinker, the global AI solutions for logistics market will grow at a CAGR of 39.6% between 2020 and 2027[1]. Such investment growth will lead to constant changes in warehouse operations and processes, gradually leading to complete automation.
Furthermore, the approach of increasing the automation of existing warehouses will certainly be essential for many operators in order not to defer the modernization of the Supply Chain until the arrival of a new warehouse. With this so-called “brownfield” approach, the evolution of operations is easier to control, even if it remains complex.
To prepare for this, I offer below some best practices that can make warehouse automation more accessible. I will use concrete examples from the field of robotic picking, which is the next big step in intralogistics, and will hopefully provide some useful lessons.
Rule #1: Focus on a specific process or a set of SKUs first
This may seem obvious, but the first step in an automation project is to analyze which part of the process or set of SKUs needs further automation. Robotic solutions are improving every day, but they cannot yet handle all kinds of use case or all types of products. It would be naive to expect that a new technology can be adopted and improve the flow of all your SKUs overnight. It is therefore essential to define a precise scope for iteration and progressive improvement of operations. This goes for any type of innovation implemented in the warehouse.
For example, in robotic piece-picking, for each new customer, it is essential to analyze the SKUs to be processed to determine which ones can be managed by the robot and which ones will not. This is the first phase of all our projects at Nomagic, and it is done with a dedicated tool that we call the “Product Scout”. The Product Scout essentially takes photos of SKUs sold during a given period and analyzes their shape, position and characteristics to determine if they can be handled by robotic picking.
Focusing on a smaller number of SKUs at a time brings the advantage that we can quickly achieve target KPIs such as system throughput or autonomy and make the solution viable for continuous operations. After this initial calibration and once the first robot is on track, we can move on to other SKU categories and gradually deploy more robots.
Rule #2: Validate in production as soon as possible
There is a lot of innovation in the intralogistics industry and you are surely often surprised by what you see at various conferences and events. But – returning to our example – robotic picking demos at conferences have no value, especially if they come from suppliers who cannot demonstrate real production experience. An isolated test environment won't tell you much about the robot's actual performance in a warehouse. This is true in part because each use case is different, and the types of exceptions encountered can vary greatly. The positions of the picking bin, the type of items handled, the type of product placement or packaging – the number of combinations is endless.
For this reason, it is essential to first clearly identify your needs, then look for a supplier who already has some production experience with comparable use cases. Otherwise, if you both start from scratch, a lot of time and effort will be wasted before satisfactory results are achieved.
However, one problem with testing in production is the cost of doing so. You must identify the right supplier and define processes together to be able to set up a test without having to invest hundreds of thousands of euros. This is exactly what pushed us at Nomagic to develop a dedicated system for our robots to interact directly with the operator screen of a goods-to-person station, thus removing the need for WMS integration during the pilot phase. Thanks to this approach, we can focus mainly during the pilot phase on ramping up the system on the set of selected SKUs (rule #1!) and work on WMS integration later.
Rule #3: Invest in remote operations
Finally, the most important characteristic of a fully automated system is its autonomy. But as warehouses increasingly integrate systems that use machine learning and tend to improve over time, autonomy should not be taken for granted but rather as the result of the process of integration and adaptation. As explained with the SKU selection process (see Rule #1), the picking system can only become autonomous gradually. And the main KPI used to measure this autonomy is the intervention rate or the frequency with which a person has to look after the system.
Given the gradual improvement of the system, initial configuration and ramp-up may require some time and intervention on your part or that of your supplier. The cure is not to find a vendor who can promise that this won't happen and that their machine learning system will be correct from day one. The best solution is to implement a remote operations management system that eliminates the need for on-site interventions. This means investing in a process by which people can intervene on the automated system (for us a robot) without having to go on site.
The way we handle this at Nomagic is to have our own remote operators, called "watchers", who remotely monitor the movements of any robot at our customers' premises. Our proprietary robot-watcher communication system alerts the watcher whenever the robot encounters a problem and allows it to take control of the robotic arm and remove the obstacle or place the object in a different way. For each remote intervention case, numerous photos and videos are stored and detailed positioning data is recorded so that engineers can analyze them after the fact and can make the system resilient autonomously in the case in question. A large majority of problems can be handled in this simple way, which significantly reduces the number of on-site interventions, results in better availability and gradually makes the 100% system automated.
Conclusion: Gradually towards lights out warehouses…
The end goal of warehouse automation, which is the “lights out” warehouse, is not something that will happen tomorrow. However, when it does, it will not be in the laboratory of a robotics startup or in the development center of an integrator. This will have to happen in a production warehouse with real products, real project engineers, for a selected part of the process (Rule #1), after careful validation on site (Rule #2) and with the support of remote operators ( Rule #3). The iterative nature of such solutions means that those who give themselves more time to adapt a solution to their needs will reap the most rewards. I hope that the approach I have described here will help make this process relatively smooth and can serve as a good reference for the introduction of full warehouse automation, especially for existing warehouses that can still be further automated .
[1]https://www.reportlinker.com/p06030748/Global-Artificial-Intelligence-AI-in-Supply-Chain-Industry.html
Read all the “Expert Opinion” articles on the SprintProject blog
