No Reboot Required: Real-Time Modular Robot Payload Integration
We present a novel approach for plug-and-play robotics that allows integrating modular payloads into a skillsbased task planner and execution engine of an autonomous robot. This integration occurs seamlessly without rebooting the host, restarting processes, or requiring user input. The core innovation is packaging payload operational and behavioral software-e.g. hardware drivers, support libraries, executable skills and domain-specific ontologies-within the payload itself. Upon physical connection to the host's data port, the payload components are automatically detected, transferred, and incorporated into the host's software stack in real-time. This approach eliminates downtime, manual integration steps, and the need for expert knowledge, significantly enhancing robot adaptability and operational flexibility. Furthermore, the approach leverages popular open-source software to ease and encourage its adoption by developers. The method was demonstrated on a mobile robot that integrated a modularized commercial device, automatically ingested new skills from it, and used those skills to complete a simple task that was beyond its previous capabilities.
Authors
- Robert Blake Anderson (ORCID: https://orcid.org/0000-0002-1847-5982)
- Steven Swanbeck (ORCID: https://orcid.org/0000-0002-7106-3530)
- Jonathan Salfity
- Corrie Van Sice
- Benjamin K. Chen (ORCID: https://orcid.org/0000-0002-5404-1997)
- Mitch Pryor (ORCID: https://orcid.org/0000-0001-5089-9964)
Institutions
- Walker (United States) (US)
- The University of Texas at Austin (US)
Publication Details
- Published
- 2025-06-30
- DOI
- https://doi.org/10.1109/ur65550.2025.11078050
- Primary Topic
- Modular Robots and Swarm Intelligence
- Type
- article
- Field-Weighted Citation Impact
- 0.00