Agentic AI for Robot Teams
A presentation from the Johns Hopkins Applied Physics Laboratory describes recent work on agentic AI for collaborative robotic teams. It frames the core challenges of autonomy, coordination and adaptability across heterogeneous systems, then introduces a scalable architecture for agentic behaviors in multi-robot environments. The talk closes with challenges encountered and practical lessons learned from ongoing research.
Key Takeaways
- The work advances agentic AI for collaborative robotic teams at the Johns Hopkins Applied Physics Laboratory.
- It addresses autonomy, coordination and adaptability across heterogeneous systems.
- A scalable architecture is introduced to support agentic behaviors in multi-robot environments.
- The approach is demonstrated in hardware with a heterogeneous team of robots.
- The talk shares key challenges and lessons learned from research and development.
The focus of the presentation
The talk centers on a specific applied research effort.
- ›The work comes from the Johns Hopkins Applied Physics Laboratory.
- ›It targets agentic AI for collaborative robotic teams.
- ›It frames the core challenges before presenting an architecture.
The presentation highlights recent efforts to advance agentic AI for robot teams, beginning by framing the problem of enabling autonomy, coordination and adaptability across heterogeneous systems.
Core challenges addressed
Several challenges define the problem space.
- ›Enabling autonomy across different robotic systems.
- ›Coordinating multiple robots that may not share the same hardware.
- ›Adapting behavior across heterogeneous systems.
The proposed architecture
A scalable design supports agentic behaviors.
- ›The architecture is described as scalable.
- ›It is designed to support agentic behaviors in multi-robot environments.
- ›It builds on LLM-based AI agents.
The session introduces an approach to applying LLM-based AI agents to robotic teams, providing an introduction to those agents as part of the key learnings.
Hardware demonstration
The approach was shown running on real robots.
- ›Demonstrations show the approach running in hardware.
- ›A heterogeneous team of robots was used.
- ›The demonstrations support the described architecture.
Lessons and future work
The talk concludes with reflection.
- ›It presents key challenges encountered during the work.
- ›It shares practical lessons learned from ongoing research and development.
- ›It outlines future work in this area.
The material is offered as a free whitepaper for download.
Frequently Asked Questions
Who produced this work?
The Johns Hopkins Applied Physics Laboratory, described in the presentation on agentic AI for robot teams.
What is the goal of the work?
To advance agentic AI for collaborative robotic teams by enabling autonomy, coordination and adaptability across heterogeneous systems.
What does the architecture support?
Agentic behaviors in multi-robot environments, and it is described as scalable.
Was the approach tested on real robots?
Yes. The talk includes demonstrations of the approach running in hardware with a heterogeneous team of robots.
What technology underpins the agents?
LLM-based AI agents, which the presentation introduces and applies to robotic teams.
The presentation offers an LLM-based agentic approach to coordinating heterogeneous robot teams, backed by hardware demonstrations and lessons learned.
Continue Learning
Comments
Sign in to join the conversation