February 28, 2026

A Docker-Enabled Real-Time Framework for Robotic Applications in Heterogeneous ROS 2 Environments

We are pleased to announce the publication of our latest research article:

“A Docker-Enabled Real-Time Framework for Robotic Applications in Heterogeneous ROS 2 Environments”

Authors: Ji Min Lim, Keon Woo Kim, Byoung Wook Choi, and Raimarius Delgado

This paper marks a special milestone for our lab — it is the first publication co-authored with undergraduate researchers, highlighting our commitment to early research engagement and hands-on system-level experimentation.

We developed a Docker-enabled real-time evaluation framework that enables controlled and repeatable experiments under heterogeneous ROS 2 deployments. The framework integrates:

• An RT-PREEMPT–patched Linux kernel

• Dockerized ROS 2 distributions

• Configurable cross-version communication pathways

Using this framework, we:

• Quantified Docker-induced effects on task periodicity, jitter, and response time

• Measured end-to-end communication latency across:

  • Host-only deployments

  • Container-only deployments

  • Hybrid host–container configurations

To validate practical applicability, we applied the framework to a mobile robot system integrating legacy control software with new modules, including a reinforcement learning–based decision layer, within a mixed ROS 2 environment.

This work provides:

• Reusable real-time evaluation tooling

• Empirical analysis of Docker and cross-version ROS 2 overheads

• Actionable deployment guidelines for deterministic robotic systems

The results support safer and more practical modernization pathways for robotics systems operating under heterogeneous and containerized constraints.

We congratulate all authors and extend our gratitude to Prof. Byoung Wook Choi of SeoulTech — especially our undergraduate researchers on this significant achievement.

More updates and research developments from our lab will follow soon.