ROS 2 Executor: How to make it efficient,
real-time and deterministic?
Overview
The power of ROS as a software framework is evident in its application area from feather-weight drones to vacuum cleaners to autonomous driving trucks. While a consolidation with flexible abstraction layers in the area of communication middleware has already been achieved, we are currently seeing a lot of activity when it comes to tweaking the execution of ROS nodes for the individual use cases. In this workshop you will get deep insights into the latest developments in the community dealing with efficiency, real-time and determinism. An open-source reference system will be introduced that enables participants to gain practical experience.
Intended audience are all ROS users and developers that are interested in how software written with ROS is executed on the operating system and hardware, especially the ones who want to use ROS 2 in systems where they have to deal with resource constraints and real-time requirements.
ARM
x86
Agenda 10 AM - 2 PM CST
10.00 Motivation (Michael) watch video | download slides
10.10 Current status of executor in ROS 2 Galactic (William) watch video| download slides
10.30 Introduction to the reference system (Evan) watch video| download slides
10.40 Tracing the reference system (Christophe) watch video| download slides
11.00 Tune the system for real-time (Andrei) watch video| download slides
11.20 Callback groups executor (Ralph) watch video| download slides
11.40 Event executor (Alberto) watch video| download slides
​
12.00 Break - 10 min
12.10 PiCAS executor (Hyunjong) watch video| download slides
12.30 RCLC executor (Jan) watch video| download slides
12.50 Lock-free ROS 2 executor: A ring-buffer to rule them all (Pablo) watch video| download slides
1.10 Executor with wait-set and polling subscription (Michael) watch video| download slides
1.30 Panel discussion (all) watch video
2.00 End of workshop
​
Speakers
Windriver
Jan Staschulat
Bosch
​Klepsydra
Open Robotics
University of California, Riverside
Polytechnique Montréal
Apex.AI
iRobot