Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CGoT: A Novel Inference Mechanism for Embodied Multi-Agent Systems Using Composable Graphs of Thoughts

Yixiao Nie, Yang Zhang, Yingjie Jin, Zhepeng Wang, Xiu Li, Xiang Li

TL;DR

The paper tackles efficient coordination of embodied multi-agent systems where ego-vehicles carry service robots to perform tasks in dynamic settings. It introduces Composable Graphs of Thoughts (CGOT), a Graph of Thoughts–based inference framework that allows dynamic agent merging to form composites and reason collaboratively through a decentralized GoT-based loop. The approach is formalized with a graph-centric model and a three-layer Execution pipeline, and is demonstrated in a simulated office-park delivery scenario where CGOT achieves planning performance comparable to GoT without merging while substantially reducing token consumption. The work highlights the potential for scalable, energy-efficient autonomous collaboration in mixed vehicle–robot platforms and lays groundwork for real-world deployment.

Abstract

The integration of self-driving cars and service robots is becoming increasingly prevalent across a wide array of fields, playing a crucial and expanding role in both industrial applications and everyday life. In parallel, the rapid advancements in Large Language Models (LLMs) have garnered substantial attention and interest within the research community. This paper introduces a novel vehicle-robot system that leverages the strengths of both autonomous vehicles and service robots. In our proposed system, two autonomous ego-vehicles transports service robots to locations within an office park, where they perform a series of tasks. The study explores the feasibility and potential benefits of incorporating LLMs into this system, with the aim of enhancing operational efficiency and maximizing the potential of the cooperative mechanisms between the vehicles and the robots. This paper proposes a novel inference mechanism which is called CGOT toward this type of system where an agent can carry another agent. Experimental results are presented to validate the performance of the proposed method.

CGoT: A Novel Inference Mechanism for Embodied Multi-Agent Systems Using Composable Graphs of Thoughts

TL;DR

The paper tackles efficient coordination of embodied multi-agent systems where ego-vehicles carry service robots to perform tasks in dynamic settings. It introduces Composable Graphs of Thoughts (CGOT), a Graph of Thoughts–based inference framework that allows dynamic agent merging to form composites and reason collaboratively through a decentralized GoT-based loop. The approach is formalized with a graph-centric model and a three-layer Execution pipeline, and is demonstrated in a simulated office-park delivery scenario where CGOT achieves planning performance comparable to GoT without merging while substantially reducing token consumption. The work highlights the potential for scalable, energy-efficient autonomous collaboration in mixed vehicle–robot platforms and lays groundwork for real-world deployment.

Abstract

The integration of self-driving cars and service robots is becoming increasingly prevalent across a wide array of fields, playing a crucial and expanding role in both industrial applications and everyday life. In parallel, the rapid advancements in Large Language Models (LLMs) have garnered substantial attention and interest within the research community. This paper introduces a novel vehicle-robot system that leverages the strengths of both autonomous vehicles and service robots. In our proposed system, two autonomous ego-vehicles transports service robots to locations within an office park, where they perform a series of tasks. The study explores the feasibility and potential benefits of incorporating LLMs into this system, with the aim of enhancing operational efficiency and maximizing the potential of the cooperative mechanisms between the vehicles and the robots. This paper proposes a novel inference mechanism which is called CGOT toward this type of system where an agent can carry another agent. Experimental results are presented to validate the performance of the proposed method.
Paper Structure (12 sections, 3 equations, 2 figures)

This paper contains 12 sections, 3 equations, 2 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Ego-vehicles and Service Robots
  4. LLMs Application
  5. Method
  6. Problem Formulation
  7. System Structure
  8. Experiment
  9. Basic Settings
  10. Execution Efficiency
  11. Token Consumption
  12. Conclusion

Figures (2)

  • Figure 1: The structure of the CGoT framework. The system operates in a continuous loop between the inference, conclude, and execution phases until all tasks are completed. When an emergency event or human input occurs, the emergency handling module joins in the loop to ensure that task execution proceeds smoothly and efficiently.
  • Figure 2: The contrast of GoT and CGoT on token consumption during a test. CGoT saves more tokens than GoT as the proceeding of the tasks execution.