To date, the majority of real-time communication in industrial automation has been based on wired Ethernet technology, which, however, limits the flexibility of the machines and robots used in production due to the physical connection of the participants. And flexibilization is picking up speed: Assembly lines are being replaced by mobile and flexibly deployable platforms, mobile robots perform their work in changing locations and then move autonomously to the next job. Such scenarios require real-time communication without cables: wireless communication. However, the wireless technologies available to date are not real-time capable, as the characteristics of the wireless medium make it difficult to achieve the required determinism.

OPC UA with TSN and Pub/Sub: the possibilities

What are the main challenges that still exist today in the context of real-time radio communication? What are possible solutions? What are the limitations? And to what extent would OPC UA with TSN and Pub/Sub be compatible with real-time radio communication? OPC UA is a generic description and exchange platform that allows components from all manufacturers to communicate with each other at different levels. Various low-layer technologies can be used as the communication medium. Current OPC UA specifications place particular emphasis on use cases in the non-real-time area. Efficient communication architectures are provided for this purpose, which are based in particular on the server/client model and provide standard mechanisms for transport and protection. However, these can only be used to a limited extent or are inefficient when it comes to real-time-capable applications.

For this reason, the publisher/subscriber model was proposed for real-time applications, which can be applied more naturally to the communication patterns and is also much easier to implement. The latter plays a particularly important role for resource-limited components such as sensors. Other key features of the new OPC UA extension relate to the ability to implement data exchange in a local network with reduced effort. Furthermore, the correct choice must also be made for the low-layer communication technology in order to guarantee real-time capability. TSN is a manufacturer-independent solution that can be used flexibly for various applications and is largely based on inexpensive hardware components.

The selectivity of the radio channel, i.e. the fact that the reception quality varies unpredictably in terms of location, time and frequency at all times, is a major challenge for deterministic applications. This effect is amplified by the mobility of the planned applications themselves, but also by the mobility of other objects and people in production. Furthermore, the propagation of the signals in space requires comprehensive and deterministic management of the radio participants so that interference is avoided, which would prevent real-time capability.

The challenge of radio communication

As a result, the available spectrum in the unlicensed and licensed areas is severely restricted and limited. However, this also limits the available capacity in terms of cycle time, data throughput, reliability and the number of network participants. Products available today based on existing and largely standardized wireless technologies such as WLAN, Bluetooth, ZigBee and 4G were not developed for real-time requirements and therefore cannot be used in the complex environments of production with a large number of network users. However, new approaches - based on WLAN technology (EchoRing) or in the cellular area (5G URLLC) - are already being developed and are already available as products or for initial evaluation for the first applications.

Crucial point: real time

In order to be able to support OPC UA with TSN and Pub/Sub in radio technology in the future, three essential requirements must be met:

1. very high reliability/availability of the wireless connection,
2. support for very short cycle times and
3. time synchronization at a micro-second with regard to a central master clock.

Due to the inherent challenges in the field of radio communication, an efficient (in terms of the necessary spectrum) and cost-effective (in terms of hardware costs and complexity) implementation is of particular importance.

The architecture options

As the quality of the radio signal is not easily predictable, the issue of redundancy plays a major role. The approach to be chosen is typically a compromise between the spectrum used, complex hardware, supported form factor, achievable cycle time and scalability. Accordingly, an overall efficient implementation is the key to success. The most promising approaches from today's perspective are

Frequency redundancy

In this case, the signal is sent on the direct path on at least two different frequencies at the same time for security purposes. However, this design, which is good for a short cycle time, requires more complex transmitter and receiver structures (hardware and software) and more spectral resources. In addition, the variability of the radio channel in frequency is a function of the physical properties of the transmission environment, which cannot be influenced. This means that the diversity effect cannot be controlled and may only be available to a limited extent.

Multi-antenna system

The potential increase in reliability with a short cycle time by transmitting and receiving signals via multiple antennas at the transmitter and receiver has a significant impact on the respective form factor and the complexity of the transmitter and receiver structures. Both ultimately increase the costs and power consumption of the system solution.

Cooperative systems

In contrast to the two alternatives, the signal is secured here with the aid of other transmission paths via other radio network partners and therefore has a decisive advantage in terms of achievable diversity gains. For example, if direct paths are no longer available due to mobility in the factory environment. At the same time, this approach works with very simple and inexpensive hardware and results in very efficient utilization of the available spectrum. On the other hand, cooperation has an impact on the cycle time that can be supported.

Scaling with short cycle times - the sticking points

As with cable-based systems, the choice of components and interfaces used in the communication chain as well as the respective software have a significant influence on the overall cycle time. Depending on this design and the requirements of the application, a maximum time budget is available for the radio link between transmitter and receiver - including processing time.

Compliance with this time budget depends largely on the available bandwidth, i.e. the spectrum, and its use. Typically, the available spectrum must be exclusively available for the following topics and divided up accordingly and therefore represents a relevant cost factor that has a different impact in unlicensed (e.g. EchoRing) and licensed spectrum (e.g. 5G URLLC):

• Reliability: Irrespective of the system architectures described, additional repetitions of the data packets are typically provided, which only need to be kept available for this case if the system has a guaranteed service level.
• Capacity: A key aspect is the data rate required by the application in each case, which results from the packet lengths and their periodicity. The higher the periodicity and the shorter the packet length, the higher the overhead and therefore the lower the net data rate that can be supported in the defined time interval.
• Susceptibility to interference: Due to the dynamically changing radio signal strengths (strongly influenced by moving production factors such as production personnel, tools, production materials and logistics) and the uncontrollable environments in terms of security, 'emergency channels' should be provided that can be switched to in real time if necessary.
• Coverage : Radio network coverage also plays a major role, and not just for the scalability of the supported applications. In principle, a frequency can only be used once in the same location and time. The same frequency can only be reused within a certain spatial distance (analogous to radio network planning in the cellular environment).

It is very difficult to make a general statement about latency times, as the wireless channel is subject to other external factors, such as the thickness of walls or the positioning of antennas. In the case of EchoRing, it can be assumed that the number of participants in a radio cell with a cycle time of 16 ms, taking into account the expected latencies in the other components, is in the high single-digit to medium double-digit range (per 20 MHz radio channel). Lower cycle times are possible, but this significantly reduces the maximum number of participants per network. At the same time, these radio cells typically have to be set up with a narrow mesh - 25 m x 25 m - in order to meet the high requirements and the limited spectrum.

Time synchronization at 1 µs

In addition to the challenges mentioned, time synchronization is the biggest hurdle from today's perspective: Time synchronization to a central master clock must be implemented on each component, taking into account the respective interfaces and hardware and software aspects. In order to meet the jitter requirements, this leads to a high level of complexity and high costs, not least because of the necessary selection of supporting hardware and special software.

With regard to the radio components, the problem arises in the same way, so that the interface used, the processing in the radio component and the access to the radio medium must be optimized. Such a realization, for example on a software-defined radio with FPGAs, can already be implemented today, but is only suitable for marketing to a limited extent due to the high hardware and software costs. As far as the currently available standard radio components are concerned, these have not been developed and optimized for this application. Accordingly, it is desirable that the necessary prerequisites are available in the next generations of WLAN 802.11 AX or 5G URLLC.

Current solutions

As things stand today, the requirements for time synchronization in the 1 µs range are not yet available as a product together with a real-time radio system. However, there are already tried and tested systems based on existing standard radio components that can be used for a wide range of applications.

Existing wireless technologies do not yet meet the requirements of deterministic data traffic.

© R3 Communications

The EchoRing product from R3 Communications, for example, is based on a WLAN chip from Texas Instruments and can be used in the 2.4 and 5 GHz radio spectrum. It uses cooperative mechanisms to reduce susceptibility to interference. It also has features for securing the wireless connection and roaming for spatial coverage. A key advantage of the system is its transparent architecture, which enables it to support different protocols on higher layers. This functionality has already been validated with existing industry protocols. Such systems already support leading industry protocols with cycle times of 16 ms, or even less depending on the design, while offering a very high level of reliability.

Real-time radio with OPC UA and TSN

A key prerequisite for supporting OPC UA with TSN and Pub/Sub is the transparent implementation of a real-time wireless communication system. In contrast to Bluetooth, for example, where the protocol also includes higher layers, the desired solution must be limited to layers 1 and 2.

The TSN stack and wireless: The integration of wireless technologies in TSN networks follows the wired approach.

© R3 Communications

Another important point is the support of different communication patterns. While previous wireless technologies are particularly suitable for implementing the server/client model - one-to-one - next-generation wireless technologies either already support the pub/sub communication pattern today (EchoRing) or will do so when they are introduced (5G URLLC). To this end, the broadcast properties of the wireless medium are combined with efficient multicast addressing to map one-to-many and many-to-many relationships.

The changes proposed by the OPC Foundation with regard to increasing efficiency - allowing UDP instead of TCP as the transport protocol, or the option of even feeding user data directly to layer 2 - reduce the data rates that the communication system has to provide. This is extremely helpful, particularly in view of the limited spectrum in the wireless sector.

The last important point is the internal linking of wired and wireless technologies: Due to the network management functionality of TSN, especially 'Time Aware Traffic Shaping', data packets can be forwarded to the radio medium and received by the radio medium according to their priority. The transparent architecture of the wireless technology is crucial for this in order to meet the requirements of the application in a cycle time-optimized manner.

What can we expect?

The development of real-time radio technology is currently being pursued with high priority in various areas. Existing solutions (such as EchoRing) already enable access to real-time radio communication on inexpensive standard hardware. However, due to the current lack of time synchronization in the 1 µs range, time-isochronous applications are not yet supported. It can be assumed that the first demonstrators and prototypes will show the missing time synchronization functionality at trade fairs in the near future. However, the marketing of cost-effective standard hardware is not realistic until after 2020.

Author:
Florian Bonanati is the founder and Managing Director of R3 Communications GmbH.