Distributed publisher-subscriber architectures performance for robotics virtual reality applications: A case study on MQTT
Virtual Reality has been traditionally explored in many robotics systems, with applications such as off-line programming, trajectory planning, teleoperation, education, design, natural user interfaces and rehabilitation. Implementing these features in an all-in-one monolithic solution can be complex...
Saved in:
Published in: | 2017 Latin American Robotics Symposium (LARS) and 2017 Brazilian Symposium on Robotics (SBR) pp. 1 - 5 |
---|---|
Main Authors: | , , , |
Format: | Conference Proceeding |
Language: | English |
Published: |
IEEE
01-11-2017
|
Subjects: | |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Virtual Reality has been traditionally explored in many robotics systems, with applications such as off-line programming, trajectory planning, teleoperation, education, design, natural user interfaces and rehabilitation. Implementing these features in an all-in-one monolithic solution can be complex and slow. Also, new trends in automation, notably the Internet of Things, represent promissing technological improvements to the robotics field. This paper presents an alternative for developing such distributed and evolutive applications by defining a minimalistic set of commands for physical and virtual robots and integrating them with a publisher-subscriber network, commonly used with Internet of Things. This allows several configurations, depending on the available modules and on which elements subscribe to each topic. Thus, we discuss the performance of a small network comprising 3 agents: one input application acting as a real-time publisher and two subscribers associated with a physical and a virtual robot. All machine clocks are synchronized with the IEEE 1588-2008 protocol and the application packages are stored with timestamps on log files for evaluating metrics such as the package transmission rate and the latency. Results indicate that the overall performance on a local area network for this kind of application is beyond the physical robot's maximum update rate. Further tests might consider other time syncing mechanisms, wireless networks and different configurations with more network nodes. |
---|---|
DOI: | 10.1109/SBR-LARS-R.2017.8215340 |