Mr. Schildknecht, you took a closer look at the combination of 5G and safety - in what context did this take place?

Thomas Schildknecht: 5GANG stands for '5G applied in industry'. As part of the BMBF-funded project entitled '5G applied in industry', we at the Research
Institute for Rationalization and the Machine Tool Laboratory at RWTH Aachen University have been working intensively on the combination of 5G with safety since 2017. The investigations were carried out on the 5G campus network there and focused on the specific question of how the transmission of Profisafe via Profinet using 5G behaves in terms of jitter, latency and reliability.

5G and safety: what kind of applications would this combination make sense for?

A preferred application for 5G campus networks is undoubtedly the operation of automated guided vehicle systems with their large number of participants. In addition to the actual data transmission, these systems require an emergency stop function as a prerequisite for safe operation in terms of people and material.

"We are certain: 5G campus networks are basically real-time capable!"

The task here is to implement this function both locally on the vehicle and between the vehicles or from a central location; furthermore, this functionality should be independent of the safety protocol used.

Is 5G suitable for safety per se or what still needs to be done?

Today's safety protocols are designed to communicate between a safety controller and a remote IO, for example, via a network cable. The conversion to a wireless channel is basically solved by the black channel principle. This also met the requirements for safety via 5G at the start of the project. We then wanted to specifically investigate how the data of a PLC can be tunneled via a 5G campus network base station and how the 3.4 to 3.8 GHz band behaves in a factory environment - especially in comparison to WLAN and Bluetooth in the unlicensed bands.

The result?

WLAN showed a significantly lower latency in the undisturbed case when looking at the PN connection 'end-to-end'. However, the legally prescribed coexistence measures for WLAN when using the unlicensed 2.4 GHz band lead to an undefinable jitter of the PN transmission! This can then be up to several 100 ms, which in practice leads to a system standstill with manual acknowledgement in a safety application!

In the licensed 3.4 to 3.8 GHZ band, on the other hand, such mechanisms are not used, resulting in significantly lower jitter. As far as the PN update time is concerned, we worked with a 1 ms setting on the fieldbus cable side in the test and were able to achieve latency values over 5G down to the single-digit millisecond range.
We conclude from this that a 5G campus network is basically suitable for real-time capable data transmissions!

What does your further roadmap look like?

The next step is to clarify the behavior when many hundreds of such vehicles are operated on a 5G campus network instead of a single transport system. As cellular networks are fundamentally service-oriented, this poses the challenge of maintaining equal cyclical communication with all participants. The challenge here is not data volume or latency, but rather the cyclical communication of very small but very large data packets with a defined maximum jitter. Standardization for this requirement is expected with 3GPP Release 16 for uRLLC. But even with the currently valid Release 15, Profisafe communication with many vehicles via 5G is already possible.