Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Implicit Coordination using Active Epistemic Inference for Multi-Robot Systems

Lauren Bramblett, Jonathan Reasoner, Nicola Bezzo

TL;DR

This paper addresses implicit coordination in multi-robot systems when explicit communication is restricted or absent. It integrates Theory of Mind (ToM) up to third-order, dynamic epistemic logic (DEL), and active inference to infer others' beliefs and signal intentions while selecting actions. The approach maintains a posterior $q( rue{G})$ over possible goal configurations and minimizes the variational free energy $F$ (and the expected free energy $ ext{E}[F]$) to guide runtime task allocation. Across simulations and lab experiments, the method outperforms greedy and first-order baselines, reducing uncertainty (entropy) and enabling efficient, scalable coordination in heterogeneous robot teams.

Abstract

A Multi-robot system (MRS) provides significant advantages for intricate tasks such as environmental monitoring, underwater inspections, and space missions. However, addressing potential communication failures or the lack of communication infrastructure in these fields remains a challenge. A significant portion of MRS research presumes that the system can maintain communication with proximity constraints, but this approach does not solve situations where communication is either non-existent, unreliable, or poses a security risk. Some approaches tackle this issue using predictions about other robots while not communicating, but these methods generally only permit agents to utilize first-order reasoning, which involves reasoning based purely on their own observations. In contrast, to deal with this problem, our proposed framework utilizes Theory of Mind (ToM), employing higher-order reasoning by shifting a robot's perspective to reason about a belief of others observations. Our approach has two main phases: i) an efficient runtime plan adaptation using active inference to signal intentions and reason about a robot's own belief and the beliefs of others in the system, and ii) a hierarchical epistemic planning framework to iteratively reason about the current MRS mission state. The proposed framework outperforms greedy and first-order reasoning approaches and is validated using simulations and experiments with heterogeneous robotic systems.

Implicit Coordination using Active Epistemic Inference for Multi-Robot Systems

TL;DR

This paper addresses implicit coordination in multi-robot systems when explicit communication is restricted or absent. It integrates Theory of Mind (ToM) up to third-order, dynamic epistemic logic (DEL), and active inference to infer others' beliefs and signal intentions while selecting actions. The approach maintains a posterior over possible goal configurations and minimizes the variational free energy (and the expected free energy ) to guide runtime task allocation. Across simulations and lab experiments, the method outperforms greedy and first-order baselines, reducing uncertainty (entropy) and enabling efficient, scalable coordination in heterogeneous robot teams.

Abstract

A Multi-robot system (MRS) provides significant advantages for intricate tasks such as environmental monitoring, underwater inspections, and space missions. However, addressing potential communication failures or the lack of communication infrastructure in these fields remains a challenge. A significant portion of MRS research presumes that the system can maintain communication with proximity constraints, but this approach does not solve situations where communication is either non-existent, unreliable, or poses a security risk. Some approaches tackle this issue using predictions about other robots while not communicating, but these methods generally only permit agents to utilize first-order reasoning, which involves reasoning based purely on their own observations. In contrast, to deal with this problem, our proposed framework utilizes Theory of Mind (ToM), employing higher-order reasoning by shifting a robot's perspective to reason about a belief of others observations. Our approach has two main phases: i) an efficient runtime plan adaptation using active inference to signal intentions and reason about a robot's own belief and the beliefs of others in the system, and ii) a hierarchical epistemic planning framework to iteratively reason about the current MRS mission state. The proposed framework outperforms greedy and first-order reasoning approaches and is validated using simulations and experiments with heterogeneous robotic systems.
Paper Structure (15 sections, 1 theorem, 31 equations, 13 figures)

This paper contains 15 sections, 1 theorem, 31 equations, 13 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Formulation
  4. Approach
  5. Epistemic Structures for Multi-Robot Reasoning
  6. Active Inference for Decision Making
  7. Higher-Order Evidence-Based Reasoning
  8. Epistemic Allocation for Dimensionality Reduction
  9. Simulations
  10. Rendezvous
  11. Task Allocation
  12. Multi-Robot Multi-Task Allocation
  13. Experiments
  14. Conclusion
  15. Acknowledgements

Key Result

Lemma 1

In a multi-robot system, incorporating higher-order reasoning (second- and third-order) reduces the variance of the robots' belief distributions compared to first-order reasoning by leveraging joint likelihoods that account for dependencies among robots, assuming that the errors in the observations

Figures (13)

  • Figure 1: Pictorial representation of the problem presented in the paper. The robots are unable to explicitly communicate their beliefs and must convey their intentions through sensorimotor communication. In the left frame, the red and blue robot are unable to converge to a correct belief state using only first-order reasoning. In the right frame, the red and blue robot clearly display their intentions by using higher-order reasoning about their observations.
  • Figure 2: Diagram of the proposed approach
  • Figure 3: Overview of generative model and process used in our multi-robot application. We assume that the state is hidden and the robot is only able to observe using their own on-board sensing capability (e.g., depth sensors, cameras).
  • Figure 4: Pictorial depiction of observation mapping to evidence and depth of reasoning.
  • Figure 5: Sample comparison of where in first-order reasoning the red UAV does not consider the blue UGV's perception of its movements
  • ...and 8 more figures

Theorems & Definitions (2)

  • Lemma 1
  • proof