For robotic systems, real-time models of the environment are no longer created solely from the immediate surroundings and the robot's own sensors. Advanced data correlation and analytics massively improve situational awareness based on data from the network or the internet. Telerobotics in particular will benefit from the enhanced connectivity of systems. Robotics is like the rest of the IIoT - almost every industrial sector has to cope with the exponentially growing volumes of digital data. In medicine, for example, successful treatment today also means managing large volumes of data. This is because the average monitored patient generates 1.5 GB of raw physiological data per day.

The automotive industry is also undergoing the greatest upheaval since the invention of the car. Connected cars and autonomous driving are about to completely change the way we move. And the car of the future is nothing other than a robot on wheels.

Proven DDS standard

The RTI Connect DDS platform fulfills a wide variety of tasks in autonomous vehicles, for example it controls the interaction between vehicles or control systems.

© Real-Time Innovations

The Internet has reached a point where it is possible to handle complex distributed systems. Almost every industry is looking for similar system architectures that connect sensors to the cloud, enable interoperability of different technologies but also bridges between different industries.

This is a good opportunity to exchange experiences between industries. This is precisely why multinational organizations such as the Industrial Internet Consortium (IIC) are focusing on reference architectures for the IIoT that can be used between systems, but above all between different industries. It is not about developing more and new standards, but about applying proven technologies and standards in new contexts.

The Data Distribution Service (DDS) is one such internationally proven standard. ROS-2 (Robot Operating System), a set of libraries and tools to support software developers creating robotic applications, also relies on DDS as a middleware solution. NASA is currently testing robotic systems in the ISS that will make it possible to control Mars expeditions from orbit in the future.

Various application scenarios

DDS meets connectivity requirements at edge, fog and cloud level and simplifies the integration of different systems in the Industrial Internet of Things.

© Real-Time Innovations

Challenges in data communication include multi-second delays in the connection; ack/nack and normal Ethernet connections no longer work. In ESA's Telerobotics & Haptics Laboratory, an astronaut from the ISS shakes hands with the guest from 5000 km away with robotic assistance. Real-time control over an unstable, low-bandwidth connection from the ISS to the ground station is made possible by DDS connectivity.

The task in DLR's 'Minimally Invasive Robotic Surgery System' (MIRS) is completely different. Heart operations today are very risky, as the heart is 'immobilized' and the patient is connected to a heart-lung machine for the duration of the operation. Three robots enable precise surgical interventions on the beating heart, a task that is impossible even for the best doctors to perform manually. A static image of the beating heart is displayed on the monitor, the surgeon operates the haptic control elements and the robotic arms compensate for the movements of the heart. Precise control in extreme real time is critical here, the system operates at a data rate of 3 kHz between the robotic arm and the control system.

Although the requirements for these robotic systems from NASA, ESA and DLR are completely different, they are implemented using the same connectivity platform: RTI Connext DDS is used in all systems. RTI Connext DDS offers a market-leading implementation, powerful development tools and the experience gained from hundreds of projects in a wide range of industries to enable developers to quickly implement new requirements and new architectures.

Plug & Play' possible

DDS acts as a data bus for the integration of systems, devices and applications. Basically, it is a connectionless architecture that abstracts the complexity of direct point-to-point integration. Components that the data bus connects are publishers (producers) or subscribers (consumers) or both. The components do not need to know each other. Applications only need to know the data types and which data topics are to be processed. Applications simply write (publish) or read (subscribe) data. DDS recognizes and connects the components loosely coupled at runtime, thus enabling 'plug & play'. Existing applications do not have to be adapted when components are added or removed.

As in a relational database, the user defines the criteria for data exchange (data model, key) in the data-centric architecture and the robot fills and updates the entries for position, direction and acceleration, for example.

RTI Connext DDS works without servers or brokers, distributed data storage avoids bottlenecks and 'single points of failure', and the fastest direct access to the latest data is possible at any time. The key to the universal use of DDS in a wide variety of systems is the precise quality-of-service configurations offered by the middleware.

The data bus architecture enables consistent communication between devices, machines, between systems and across the entire chain from the sensor to the cloud without changing technology. DDS provides the relevant data - and only this data - securely and reliably at any time in system real time. No conventional messaging middleware or communication protocol offers this functionality; every application development would have to implement it anew (e.g. message parsing, filtering, message caching & state management, discovery, presence, marshaling).

The Open Source Robotics Foundation has also recognized these advantages and is relying on communication via DDS in the next generation (ROS-2).

Securing the systems

However, the current possibilities of data communication and networking also require new ways of securing systems. Unauthorized access to data and intervention in a system, for example during heart surgery, would have devastating consequences. This is why the Open Management Group (OMG) extended the DDS standard in 2016 with the DDS Secure standard.

In contrast to other approaches, DDS Secure does not require a 'one-fits-all', but enables fine-grained security structures. While conventional approaches such as TLS or DTLS seal off the entire system or encrypt the entire data stream, DDS Secure makes it possible to adapt the security functions to the actual requirements at data level for each topic. Not all information needs to be encrypted. It is often sufficient to ensure that data access is permitted or that the data has not been manipulated. Data-centric access control, secure multicast function, encryption and logging must be configured in detail and unnecessary system impairments - such as deteriorating latency times - are avoided.

As the DDS Secure extension has been implemented as a plug-in for RTI Connext DDS, it is also possible to subsequently secure existing DDS applications without changing the application code. Customer-specific encryption algorithms and certificate management can be integrated using a software development kit (SDK).

With the new application possibilities of robotics, the requirements for functional safety are increasing accordingly. Developers must ensure that their system always does what it is supposed to do and nothing else. The relevant safety standards, such as ISO 60601 for medical devices or ISO 26262 for autonomous vehicles, must be observed and their implementation must be verified. Here too, lessons can be learned from other industrial sectors, such as aerospace.

What's new in functional safety

RTI Connext DDS has been used for many years in a number of safety-relevant systems in the aerospace industry. RTI is the only provider worldwide to offer a DO-178C Level A certifiable DDS implementation with complete documentation (Certification Evidence). ISO 26262 and 60601 certifications are currently in progress. The use of this variant reduces the time and costs involved in approving the entire system. As the certifiable variant is compatible with the DDS standard and the protocol, interoperability with non-certified implementations is also ensured.

Author:
Reiner Duwe is Sales Manager Central Europe at RTI.