Alternative approaches are plenty, but ...
An open-API approach is no alternative to interface standardization
At first glance, an open-API approach might sound tempting as a potential alternative to standardization. Each mobile robot vendor would publish its proprietary API, and fleet manager and middleware vendors would use these open APIs to interface with the vehicles. However, this approach has two significant disadvantages:
- Firstly, it doesn’t really lead to an open market. Fleet managers will still only selectively support robot APIs, and customers will still have to be very careful to find the right combination of fleet manager and supported hardware vendors.
- Secondly, standardization is not only about the API, but also about ensuring that all vehicles share minimum capabilities. Fleet managers can rely on these capabilities to bring operational excellence into the customer’s intralogistics processes. And this level of excellence is guaranteed to prevail if new types of robots are added to the fleet. On the other hand, with an open-API approach, a lot of optimizations become impossible simply because the vehicle API does not support them.
The most prominent example is traffic management, where deadlocks between robots from different vendors can only be avoided if a central traffic optimizer knows the destinations and planned routes of all vehicles and can coordinate access to locations with constrained access space. Another topic is ensuring that dependencies between transport jobs are obeyed, even if they are performed by a hybrid fleet, e.g., ensuring the right order of just-in-sequence production material supplies. Finally, a holistic transport job prioritization is key to avoiding delays. This becomes impossible if the decision to yield or to proceed is left to individual mobile robots which are unaware of the overall intralogistics picture.
Decentralized control approaches don’t perform well on a large scale
A decentralized control approach does not perform well on a large scale. Look at any inner-city street or highway in the world – all of them suffer from regular traffic jams and mobility constraints despite human intelligence being in the mix. Perhaps, in fact, precisely because of the insufficiency of individual decision-making.
Now, regardless of whether the autonomous car will reign supreme and roam our roads in the near or distant future: If, in theory, all cars were to be centrally orchestrated, traffic jams and clashes would cease to exist as unit movements would be synchronized. Individual decision-making is being taken out of the equation. If you apply this logic to a busy shop floor in which our Intralogistics Management Platform steers more than 100 AGVs simultaneously, AMRs are the ones disturbing the synchronicity: While individually responsive, their movements are limited by their algorithms' lack of access to the full floor picture.
