Working on various palletizing machines has both frustrated and inspired me.

There appears to be a pattern emerging: Reliable palletizers are slower, and the faster ones are less reliable.
While this makes sense in many ways, I did not expect it. I had the prejudice that plant manufacturers would achieve increased throughput with complex technology and higher costs. Such machines might exist, but their footprints or prices are not attractive, leading to their less frequent use.

What seems to be used when high throughput is needed are complicated machines with relatively simple technology. Elements like stops, light barriers, rollers, conveyor belts, timers, and counters are used. Many components are involved, most of them mechanical and under constant stress.

In contrast, the slower palletizers I'm familiar with are much more elegantly constructed. The most elegant ones use a robot to place the product directly onto the pallet after minimal pre-formatting. This design eliminates many components and functional groups. There's no need for a pallet lift or layer lift, no straightener, and it simplifies formatting slightly. However, there is a clear limit to the minimum size of the product and throughput when compared to non-robotic systems. A robot can only perform one movement at a time, and parallelization is difficult.

There are also systems that lie between these two extremes, but I haven't found much of interest there. Some of these palletizers use robots instead of pushers and stops for formatting. This simplifies the design but takes up quite a bit of space and faces the same scalability issues as other robotic systems.

Increasing the speed of the robots isn't really possible, as the product's own weight and acceleration tolerance severely limit it.

Now to my concept: From these comparisons, I conclude that an elegant system should offer high freedom of movement, achieve high throughput, and require few components and little space.
Furthermore, from other observations, I recognized that a system that is accommodating can ironically avoid many problems.

Based on these guidelines, I envision a palletizer that can operate independently in parallel or in series.
The inlets for empty pallets, interlayers, and materials are on one side of the palletizer. Empty pallets and interlayers are introduced at the bottom, and materials at the top.
The output side is on the opposite side. Depending on the configuration of the palletizers, interlayers, full or partially finished pallets, as well as material and scrap, can be output or passed on.
Pallets and interlayers are transported through the machine on straight sections of roller or chain conveyors. To reach the working position, a lifting device is provided for each. Centring devices are intentionally not necessary.
The material is transported through the machine on conveyor belts. Here, too, no guide rails or other centring aids are planned. At the inlets and outlets, a certain standard must be maintained so that the goods can be correctly handed over to external transport systems.

The central task of the palletizer is taken over by an area of planar drives. This is attached to the top of the palletizer. Unusually, the active side of the drives faces downwards. The area covers the interlayer station, the pallet station, and the material conveyor belt. A fixed maximum height is set by the drives, which are attached to the machine frame.
Movers move on the drives, performing their work while hovering and being moved and held by the drives. The movers can be moved in all six axes, with the highest dynamics and freedom in the X and Y linear axes and the Z rotational axis.
The conveyor belt, which delivers the material for palletizing, is mounted at a fixed distance from the drives. This, minus necessary distances, dictates the maximum height of the material.
Material-specific tools are mounted on the movers.

To fulfil the necessary tasks, movers with different tools and payloads are needed:

The geometry of the tools largely determines how efficiently the system can operate. It is important to design them in such a way that the necessary functions can be performed, but there is enough space so that the movers do not obstruct each other and have to wait. The set of tools described here seems to be a suitable solution, especially since it considers the construction dimensions of the tools and the need to place the material flush on the pallet. However, without tests in simulations and reality, this cannot be determined.
A possible combination of these mover functions is reserved.

In my opinion, the necessary information to control and regulate these functions should be done with cameras. Classic sensors cannot capture the necessary details that will be needed for high throughput, and the high number of sensors increases maintenance effort and adds weak points to the system.
In contrast, cameras and their data processing consist of relatively few parts. However, a problem in this processing chain can be fatal for the overall system. If certain influences lead to an unforeseen state and restrict function, a complete failure of the plant is to be expected, which probably cannot be remedied by technical service.
In contrast, classical systems are expected to have reduced productivity and more correctable failures.
However, these failures also increase with the age and use of the plant.

A camera-based system, however, allows variables to be automatically accommodating, something a classical machine simply cannot do to such an extent. Being accommodating in this context means that there are fewer specifications and greater tolerances in general. Pallets, materials, or the entire machine can be shifted, stretched, speeds varied, or light and material properties can differ significantly, and still allow operation at full throughput. This makes it possible to omit entire functional groups from the machine, eliminating associated weak points and thus increasing reliability.

One drawback of operating the planar drives with the active side facing downwards is that the movers must first be secured before the machine can be switched off. This may even have to be done if a person wants to enter the work area.
There are several ways to secure the movers. The most cost-effective that comes to mind are resting surfaces that rest between the drive and mover during operation. For the service position, the movers position themselves at the nearest parking position, distributed over the entire area, then the resting surfaces swing downwards, and the movers are set down on them. 

With this setup, the following process would take place when palletizing:

  1. Empty pallets, interlayers, and materials are fed to the palletizer.
  2. The palletizer recognizes the input and moves it so that it is roughly in the target area.
  3. An interlayer is laid on the pallet by the interlayer movers, if configured in the format, and its position is checked by the cameras and corrected if necessary.
  4. The material is separated by a system, then transported through the machine at a controlled speed from the conveyor belt.
  5. A set of the three types of material movers loads the material while it is moving on the belt.
  6. The material is transported to the pallet and laid down according to the format specifications. The camera systems monitor the position and relay necessary correction movements to the material aligner.
  7. Steps 5 and 6 are performed in parallel and by multiple sets of movers, and repeated until the layer is completed according to the format specifications.
  8. Step 3 is repeated after each layer. The pallet is lowered by the lifting system to make room for the next layer.
  9. Once the pallet is complete according to the format specifications, it is output from the palletizer.
  10. Material that was not processed due to exclusion criteria or other reasons is output.
  11. The pallet can now be further processed, e.g., wrapped and labelled.

I am quite satisfied with this concept. Although many sophisticated technologies are used, the effort in developing the software more than compensates for the otherwise necessary maintenance effort.
The lifting devices in the concept bother me a bit. I would like to find a way to do without them. However, in all my ideas, an apparatus was necessary to control the changing height difference between the growing pallet and the moving parts of the machine.
I also deliberately excluded pallet magazines and wrappers, as well as labellers from the concept. These are steps that would not really benefit from integration. It could be possible, but it would make the concept much less elegant.