The space industry is increasingly taking the approach that what has been successfully proved on Earth can also work in space. This is how the suppliers of space robotics came to EtherCAT, adapting technology for motion control applications on earth is well-suited for corresponding applications in space.
The space industry is known for the fact that only the best is good enough for it: after all, applications in orbit have particularly high requirements with regard to reliability; and the environmental conditions during launch, operation and, if necessary, landing of space components are also extreme. Therefore, the space industry has mainly relied on technologies developed specifically for this case of application so far – and then again and again appropriate applications for these technologies were found for these on Earth. However, this approach naturally leads to high development costs and, not least because of quite modest number of suppliers, also to high costs for the components themselves and their operation.
For this reason, the space industry – not only the “New Space” companies with their unusual and pragmatic methods – is increasingly taking the opposite approach: what has been successfully proved on Earth can also work in space. The prerequisite, of course, is that the technology meets the additional requirements of the industry. And this is how the suppliers of space robotics came to EtherCAT: the most widely used communication technology for motion control applications on earth is ideally suited for corresponding applications in space.
On the recent Space Symposium in Colorado Springs, which is considered the most important meeting place for the space industry, leading manufacturers of space robots and related components have published a white paper together with the EtherCAT Technology Group (ETG) on “How Space Robotics benefits from the World Standard for Motion Communication”. Among the companies involved in the co-authorship of the white paper is mda, the Canadian aerospace company currently developing the Canadarm3 robotic arms for NASA’s Lunar Gateway. Canadarm in the Space Shuttle and the Canadarm2 robotic arm are also from mda. The latter plays a central role in the spacewalks on the International Space Station (ISS).
Motiv Space of California developed the robotic arm for NASA’s Mars Rover “Perseverance”, an arm more than 2 meters long with five joints that carries some of the Rover’s key scientific instruments for searching for signs of life on Mars. Tethers Unlimited is the manufacturer of the KRAKEN robotic arm, which provides the space industry with a compact, seven-degree-of-freedom manipulator that enables small spacecraft to perform assembly, manufacturing, and maintenance operations in space. The Institute of Robotics and Mechatronics at the German Aerospace Center DLR has been using EtherCAT in a variety of applications for many years – and has already had EtherCAT-based systems in use on the ISS. The EtherCAT robot CAESAR (Compliant Assistance and Exploration SpAce Robot) was developed for a variety of tasks in space, such as assembling structures, maintaining and repairing satellites or removing space debris.
Requirements of a Communication Standard for Space Robotics
The white paper first discusses the general and specific requirements of space robots for a communication technology. From the functional perspective there is no real difference between robot services on earth or in space. The performance requirements are similar: although the movement velocities in space are usually lower, the extreme lightweight construction means that the systems are less rigid and therefore must be controlled highly dynamically and with low overall system delay margin. This requires cycle times down to significantly below one millisecond and precise synchronization accuracy for equidistance of position scanning and synchronization of multi-degree-of-freedom (DOF) systems, since space robots are no longer seen as individual axes lined up in a row, but as a complete system moved dynamically and synchronously.
Space robotics applications require a network structure that can adapt to dynamic configuration changes such as the addition of payload tools, sensors, cameras etc. to a robotics system. The data bus must accommodate a changing network topology by automatically detecting, addressing, and communicating with added nodes with minimal network initialization delays and hardware configuration changes.
It is obvious that high availability is of much greater importance in space than on earth. System reliability requires different levels of redundancy depending on the mission: from cable redundancy via controller redundancy to independent primary and secondary networks. A communication system for space robotics must also be able to meet the extended environmental requirements for radiation robustness, temperature, shock, vibration and so on.
Furthermore, radiation tolerant hardware must be available. Technologies that provide IP cores for the radiation tolerant path to flight FPGAs significantly reduce development time and cost associated with qualifying an ASIC for the flight radiation environment. And by combining the node controller with application specific firmware in the same FPGA, developers can optimize the PCB form factors that support the volume constraints of the space system assembly.
Space systems and their interfaces must be tested very carefully because implementation errors cannot be corrected during operation or can only be corrected with great difficulty. Accordingly, the testing effort usually exceeds the actual implementation effort, also for the communication interfaces. Extensive and proven conformance test systems for the bus technology significantly reduce this effort.
It is often necessary to integrate subsystems using other bus technologies – such as grippers or equipment for test set-ups – into the space robot control system. It must also be possible to integrate the robot itself into other environments. Therefore, the availability of interfaces to other bus systems is an indispensable requirement.
Communication-integrated functional safety according to IEC 61508 is only slowly finding its way into space robotics. Up to now, humans have rarely been in the workspace of the moving robot in space, and so simple shutdown is usually sufficient for personal protection, also thanks to relatively slow movements combined with the low forces generated by robots in space. However, it is foreseeable that this will change. Therefore, a communication system for space robotics should fundamentally support integrated functional safety.
There are also strategic requirements for a communications standard for space robotics which are less detailed than the technical ones, but equally important. The first thing to mention here is openness: open access to technology avoids dependency on individual suppliers and is therefore a central strategic requirement for a space robotics bus system. Ideally, openness extends beyond the space industry to increase supplier diversity and ensure the momentum of ongoing technological progress. Only when the bus technology is backed by a large community broad support for space robotics component and system providers is available. Moreover, a large supporting community leads to long term availability.
Stability is another key. Technical progress, improvements and functional enhancements are important and necessary, but should preferably not be accompanied by new incompatible versions of the technology, which are particularly problematic for space applications because development cycles are relatively long.
Costs are also playing an increasingly important role in space applications, and it goes without saying that the robotic bus system must contribute to cost savings.
And even if a bus technology already must meet all technical and strategic requirements today, the existence of a technology roadmap would be helpful because it suggests that future requirements can also be covered with the same solution.
How EtherCAT meets these requirements
EtherCAT is known to be the fastest industrial Ethernet Technology, since it makes optimum use of the available bandwidth due to its functional principle. Typical cycle times start at 50µs, and 100 drives can be updated every 100 µs. The Distributed Clocks synchronization mechanism results in a jitter of less than 100ns, which is also achieved in networks without a precision clock in the master: the nodes share the clock time already present in the chips. The system scales very well, so that additional nodes have minimal effect on the latency, which is appreciated by every control software architect.
EtherCAT supports any topology without affecting performance and without the complexity that arises from cascading switches or hubs: line, tree, star topologies can be freely combined. There can be up to 65,535 nodes per segment and a controller can host several segments. EtherCAT main controllers can automatically detect network changes using a topology recognition feature which compares the actual network to the configuration expected by the master and can reconfigure accordingly. Hence, nodes can be connected and disconnected during operation. EtherCAT automatically assigns addresses to the nodes, and hence no manual addressing is required. This is highly supportive of the changing robotic manipulator configurations where the robotics must extend its internal data network to include external grappled loads and/or sensors.
EtherCAT achieves cable redundancy by the ring topology without the network nodes or their chips having to have special properties. If a neighboring node (or tool) is removed from the network, the port is automatically closed so the rest of the network can continue to operate. Very short detection times < 15 μs guarantee a smooth transition. This also prevents the limitation that a failure in one node can disable the whole segment. Controller redundancy with hot standby is also possible. EtherCAT can detect topology changes due to failures, disconnection, or addition of nodes with a discovery method by querying the nodes through the network whereas the nodes not only respond with their identification, but also with information regarding the connection status of each port. Furthermore, the network nodes can be equipped with several EtherCAT chips to achieve multiple redundancy – the combination of all these possibilities is used e.g. by NASA in their Modular Robot Vehicle.
Chips with EtherCAT node functionality are already available from 12 semiconductor manufacturers: including those for extended environmental requirements such as temperature, shock or impermeability. A Beckhoff EtherCAT ASIC type was put through extensive irradiation tests in preparation for ISS missions (LEO) and was found to be suitable for space use. In addition, there are several FPGA manufacturers for whose devices an EtherCAT IP core is available: also for their respective radiation-tolerant and radiation-hardened space grade devices. Thus, EtherCAT semiconductors are available for the full range of space mission requirements.
Testability and Verification
Well-functioning interoperability is the prerequisite for the success of an open communication technology. That is why the EtherCAT Technology Group has placed emphasis on testing and certification from the very beginning. The comprehensive EtherCAT Conformance Test Tool (CTT) tests devices with well over 1000 test cases for compliance with the standard, and accredited EtherCAT test centers in North America, Europe and Asia issue official test reports, on the basis of which the ETG issues certificates of conformance. The CTT can also be extended by the developer with additional test cases, which can be used e.g., in a consortium to test specific functional extensions.
Interfacing to other communication systems
The efficient bandwidth utilization of EtherCAT allows to tunnel other protocols over the network. These can be individual telegrams/frames or entire process images of fieldbus systems. With “Ethernet over EtherCAT” (EoE) any Ethernet protocols are tunneled via EtherCAT without affecting its real-time properties. The mapping of fieldbus gateways to EtherCAT is also standardized within the ETG, so that the process data and parameters are transferred consistently, and the controller does not have to differentiate functionally between native EtherCAT devices and devices connected to underlying bus systems. There are now gateways to 35 different fieldbus systems.
The TÜV approved protocol extension “Fail Safe over EtherCAT” (FSoE also known as “Safety over EtherCAT) has a proven residual error probability of < 10-9/h and meets the requirements of Safety Integrity Level (SIL) 3 with single channel communication, whereas SIL4 can also be achieved with additional measures. A well-developed ecosystem including a TÜV-certified test tool facilitates implementation, and Safety over EtherCAT also meets the more stringent requirements of the latest edition of IEC61784-3 without modification.
EtherCAT is an open standard that can be implemented and used by anyone. As an IEC standard EtherCAT, the EtherCAT drive profile and Safety over EtherCAT are internationally recognized. In countries where IEC standards are not automatically acknowledged (such as China and South Korea) EtherCAT is also a national standard. The specification is available in English, Chinese, Korean and Japanese. EtherCAT is supported and maintained by the EtherCAT Technology Group, with over 6700 member companies from 69 countries the world’s largest fieldbus organization. Several hundred of ETG’s members are active in the space and aerospace sector.
ETG has succeeded in advancing EtherCAT without versioning issues known from other communication technologies: with EtherCAT new features have been added without changing the existing ones. Older devices can be easily replaced by newer ones without having to consider network protocol versions. This provides a stability of the technology that is second to none and ensures long-term availability and investment security.
The costs are primarily determined by two factors: Vendor variety and implementation complexity. A large variety of vendors ensures low prices and fully featured products. EtherCAT has the widest vendor and product variety of all Industrial Ethernet solutions: over 3000 vendors have registered as official EtherCAT suppliers, offering the full range of products for any type of application. Simple implementation is particularly important in space programs, as it reduces the probability of errors. With EtherCAT, the complex part of the implementation is embedded in the chips and not in the stacks. The chips (including the IP cores) are deployed in many millions of nodes and are very mature. The EtherCAT protocol stacks are extremely lean – also on the controller side and have successfully implemented in thousands of products. The availability of tools from different manufacturers additionally contributes to easy implementation and thus to cost reduction.
EtherCAT is proven and mature a million times over, but it is far from the end of its possibilities: ETG is working on the next fully backward compatible extension: with bit rates of 1 Gbit/s and more, EtherCAT G provides even more bandwidth. A focus of this development is the seamless integration of 100 Mbit/s EtherCAT networks, so that current devices and developments will not become obsolete or redundant through EtherCAT G. EtherCAT G ensures that in 25 years EtherCAT will still be the technology of choice for fast, deterministic communication in control applications.
EtherCAT proven in Space Applications
EtherCAT has been used in space applications since 2015. It is permitted to report about the project “Kontur 2”, a joint project of the German Aerospace Center DLR and the Russian Federal Space Agency ROSCOSMOS, as well as about the “Haptics-2” flight experiment within the METERON project of the European Space Agency (ESA) in conjunction with NASA. In both projects an EtherCAT equipped joystick was deployed to ISS, and EtherCAT was selected for its determinism, its openness, and the radiation robustness of the ET1100 EtherCAT chip, which was tested extensively with different radiation sources and doses.
Currently the DLR Institute of Robotics and Mechatronics is setting up the CAESAR arm, which brings the benefits of terrestrial light weight robots and the philosophy used in the design of Cobots into orbit. The 7dof robot arm is equipped with torque sensors in each joint, intelligent joint control units and EtherCAT as the fast, deterministic communication system.
CAESAR was prominently displayed at the exhibition accompanying the Space Symposium exhibition: the ideal place to present the white paper on EtherCAT in space robotics.