[Paper Review] A ROS-based Software Framework for the NimbRo-OP Humanoid Open Platform

Philipp Allgeuer, Max Schwarz+4 more|arXiv (Cornell University)|Sep 28, 2018

Robotic Locomotion and Control7 references18 citations

TL;DR

This paper presents a ROS-based software framework for the NimbRo-OP humanoid robot, enabling hardware abstraction, visual perception, and behavior generation for humanoid soccer tasks. The framework was successfully deployed at RoboCup 2013, where the robot scored its first official goal, demonstrating real-time control and modularity in a competitive environment.

ABSTRACT

Over the past few years, a number of successful humanoid platforms have been developed, including the Nao and the DARwIn-OP, both of which are used by many research groups for the investigation of bipedal walking, full-body motions, and human-robot interaction. The NimbRo-OP is an open humanoid platform under development by team NimbRo of the University of Bonn. Significantly larger than the two aforementioned humanoids, this platform has the potential to interact with a more human-scale environment. This paper describes a software framework for the NimbRo-OP that is based on the Robot Operating System (ROS) middleware. The software provides functionality for hardware abstraction, visual perception, and behavior generation, and has been used to implement basic soccer skills. These were demonstrated at RoboCup 2013, as part of the winning team of the Humanoid League competition.

Motivation & Objective

• To develop a modular, open-source software framework for the NimbRo-OP humanoid robot based on ROS to support research in full-body motion and human-robot interaction.
• To replace the initial Robotis-based software stack with a native ROS implementation for improved extensibility and community integration.
• To enable robust execution of humanoid soccer behaviors, including walking, ball detection, and goal scoring, in real-world competition settings.
• To provide a flexible, extensible architecture that supports visualization, diagnostics, and runtime parameter tuning for development and debugging.
• To foster community-driven improvement by releasing the framework and hardware designs openly under an open-source model.

Proposed method

• The framework is built on the Robot Operating System (ROS) middleware, leveraging its modular architecture for hardware abstraction, device control, and inter-process communication.
• A custom ROS node architecture was implemented to manage joint control, sensor data processing (e.g., from the fisheye camera), and behavior execution.
• The system uses a hierarchical control stack with distinct layers: Robot Control, Motion Control, and Behavior Control, each exposing ROS services and topics.
• Visualization is enabled via RQT plugins, including RViz for 3D robot state rendering, a plotter for time-series data, and a parameter tuner for runtime configuration.
• The diagnostics and walk control widgets allow real-time monitoring of battery, servo temperature, and gait commands during execution.
• The framework supports dynamic parameter tuning and data logging via ROS bags, enabling post-hoc analysis and debugging.

Experimental results

Research questions

• RQ1 How can a ROS-based software framework be effectively designed and implemented for a large-scale humanoid robot like the NimbRo-OP?
• RQ2 To what extent can ROS enable modularity, reusability, and community collaboration in humanoid robotics software development?
• RQ3 Can a ROS-based framework support real-time, robust execution of complex behaviors such as walking and soccer gameplay in a competitive setting?
• RQ4 How effective are ROS visualization and debugging tools in supporting development and runtime monitoring of a humanoid robot?
• RQ5 What are the performance and reliability characteristics of a ROS-based control stack when deployed on a robot with 20 degrees of freedom and real-time constraints?

Key findings

• The NimbRo-OP ROS framework successfully enabled the robot to score its first official goal in a RoboCup 2013 Humanoid League match, validating real-time behavior execution.
• The framework demonstrated stable operation in dynamic, real-world environments, including public demonstrations at ICRA 2013 and the RoboCup German Open 2013.
• The use of RQT-based visualization tools, including the plotter and parameter tuner, significantly improved runtime debugging and parameter tuning efficiency.
• The modular ROS architecture allowed for seamless integration of perception, motion, and behavior modules, with clear separation of concerns.
• The framework's open-source release facilitated community engagement, leading to recognition such as the HARTING Open-Source Award at the RoboCup German Open 2013.
• The system achieved reliable communication between components via TCP/IP and UDP, adapting automatically to wireless or wired connections.