Real-time robot software framework on Raspberry Pi using Xenomai and ROS2

This presentation is about a developed real-time framework based on Xenomai to run 20-sim-generated models on a Raspberry Pi 4. The framework can interact with ROS2 to log the outgoing data and to instruct the framework with setpoints. The software and hardware choices are made based on a design space exploration and analysis of the various options that can fit the set requirements. 

The viability of the framework is measured using multiple 20-sim-generated tasks implemented on the Raspberry Pi. Various tests are run to analyze the performance of the created framework, including core affinity and isolation, stress testing, and comparisons to actual 20-sim simulation data. Test results indicate that the framework’s performance meets the set requirements and can effectively provide the user with a stable RT framework that can be used in educational courses or research setups.

To join the presentation via Microsoft Teams click here