Linux runtime and AI Creator for TwinCAT 3

Beckhoff Automation has introduced two extensions to its "TwinCAT 3" automation software: "TwinCAT Runtime for Real-Time Linux" enables several TwinCAT runtimes to be executed in one IPC, and "TwinCAT Machine Learning Creator" enables the automated creation of AI models.

"TwinCAT Runtime for Real-Time Linux" makes it possible, for example, to combine different system parts in one high-performance computer to streamline programming and diagnostics. It also makes it possible to access a GPU (graphics card) from the real-time environment, providing a hardware accelerator for particularly computing-intensive applications such as vision – and much more besides.

TwinCAT runtime for real-time Linux is based on Beckhoff’s own Linux distribution, which expands the existing spectrum of operating systems beyond Windows and TwinCAT/BSD. This distribution was developed using the free Debian operating system and includes a Linux real-time kernel as standard for real-time execution of the runtime. The Debian package system also makes it easy to install additional software.

This all combines to provide a TwinCAT runtime for stable real-time control on a Linux platform. Using container technologies such as Docker, Podman, or LXC, several TwinCAT runtimes can be executed on a single IPC to achieve modularity in machine and other control applications. The individual execution of several TwinCAT runtimes on a single IPC facilitates hardware consolidation with corresponding cost advantages due to the improved utilization of the available computing capacity.

Linux users can take advantage of a whole new world of automation possibilities, while the familiar TwinCAT application programming remains unchanged. Benefits include the optimized execution of several real-time applications on a single CPU, increased user flexibility thanks to the modular control code, simplified addition or replacement of application modules, and the targeted execution of updates for individual applications. There is even a reduced overall investment in terms of both time and costs for engineering.

TwinCAT runtime for real-time Linux will initially be available for the new ARM-based CX82x0 and CX9240 Embedded PCs (with ARM Cortex A53 CPU). The Beckhoff Linux distribution will then be successively rolled out for all of the company’s other industrial and embedded PCs.

Create AI models automatically

TwinCAT Machine Learning Creator is aimed at automation and process experts and adds the automated creation of AI models to the TwinCAT 3 workflow. This allows users to handle the entire process, from data collection to the trained model, for themselves – without any AI expertise of their own. The finished model is optimally adapted to real-time requirements in the control environment in terms of latency and accuracy.

A prime application for the versatile TwinCAT Machine Learning Creator (TwinCAT 3 component TE3850) lies in AI-supported image processing for quality assurance. This, along with various other applications, taps into the ease and standardization of creating AI models for automation through Beckhoff’s no-code development platform. It not only leverages open standards, interfaces, and best practices, but also provides the trained models in the open standard format ONNX. These latency-optimized AI models for control applications are specially adapted to run on Beckhoff IPCs and with TwinCAT products, although they can also be used as ONNX models beyond the Beckhoff product world.

With TwinCAT Machine Learning Creator offering fully automated AI model creation, the potential of artificial intelligence is now available to all companies – including smaller ones – offering a competitive edge and much-needed solutions for the growing shortage of specialist skills. And for seasoned AI experts, this solution serves to streamline their workload considerably while also minimizing the potential for errors.

The ability to speed up project development processes offers yet another clear advantage, particularly as the development tool provides extensive and transparent methods for displaying the behavior of the models created and comparing them with each other. Users can also benefit from automated report generation, which supports auditing processes for AI model creation. Another crucial aspect is that the required application-specific data remains protected, since it does not leave the company.