Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Robust Online Epistemic Replanning of Multi-Robot Missions

Lauren Bramblett, Branko Miloradovic, Patrick Sherman, Alessandro V. Papadopoulos, Nicola Bezzo

TL;DR

This paper tackles robust coordination of multi-robot missions under intermittent communication by coupling a centralized GA-based mTSP planner with a decentralized epistemic planning layer. Robots exchange information via intentional rendezvous and propagate belief/empathy states to reason about others' knowledge when disconnected, using DEL and Monte Carlo Tree Search to optimize policies online. Key contributions include a novel interaction-reward mechanism to incentivize informative rendezvous, a belief-propagation framework for distributed replanning, and an MCTS-based epistemic planner that adaptively reallocates tasks in the face of faults or disturbances. The approach demonstrates superior performance over a baseline heuristic in simulations and real-world aerial experiments, enabling resilient task completion in environments with unreliable or missing communications.

Abstract

As Multi-Robot Systems (MRS) become more affordable and computing capabilities grow, they provide significant advantages for complex applications such as environmental monitoring, underwater inspections, or space exploration. However, accounting for potential communication loss or the unavailability of communication infrastructures in these application domains remains an open problem. Much of the applicable MRS research assumes that the system can sustain communication through proximity regulations and formation control or by devising a framework for separating and adhering to a predetermined plan for extended periods of disconnection. The latter technique enables an MRS to be more efficient, but breakdowns and environmental uncertainties can have a domino effect throughout the system, particularly when the mission goal is intricate or time-sensitive. To deal with this problem, our proposed framework has two main phases: i) a centralized planner to allocate mission tasks by rewarding intermittent rendezvous between robots to mitigate the effects of the unforeseen events during mission execution, and ii) a decentralized replanning scheme leveraging epistemic planning to formalize belief propagation and a Monte Carlo tree search for policy optimization given distributed rational belief updates. The proposed framework outperforms a baseline heuristic and is validated using simulations and experiments with aerial vehicles.

Robust Online Epistemic Replanning of Multi-Robot Missions

TL;DR

This paper tackles robust coordination of multi-robot missions under intermittent communication by coupling a centralized GA-based mTSP planner with a decentralized epistemic planning layer. Robots exchange information via intentional rendezvous and propagate belief/empathy states to reason about others' knowledge when disconnected, using DEL and Monte Carlo Tree Search to optimize policies online. Key contributions include a novel interaction-reward mechanism to incentivize informative rendezvous, a belief-propagation framework for distributed replanning, and an MCTS-based epistemic planner that adaptively reallocates tasks in the face of faults or disturbances. The approach demonstrates superior performance over a baseline heuristic in simulations and real-world aerial experiments, enabling resilient task completion in environments with unreliable or missing communications.

Abstract

As Multi-Robot Systems (MRS) become more affordable and computing capabilities grow, they provide significant advantages for complex applications such as environmental monitoring, underwater inspections, or space exploration. However, accounting for potential communication loss or the unavailability of communication infrastructures in these application domains remains an open problem. Much of the applicable MRS research assumes that the system can sustain communication through proximity regulations and formation control or by devising a framework for separating and adhering to a predetermined plan for extended periods of disconnection. The latter technique enables an MRS to be more efficient, but breakdowns and environmental uncertainties can have a domino effect throughout the system, particularly when the mission goal is intricate or time-sensitive. To deal with this problem, our proposed framework has two main phases: i) a centralized planner to allocate mission tasks by rewarding intermittent rendezvous between robots to mitigate the effects of the unforeseen events during mission execution, and ii) a decentralized replanning scheme leveraging epistemic planning to formalize belief propagation and a Monte Carlo tree search for policy optimization given distributed rational belief updates. The proposed framework outperforms a baseline heuristic and is validated using simulations and experiments with aerial vehicles.
Paper Structure (18 sections, 16 equations, 8 figures, 2 algorithms)

This paper contains 18 sections, 16 equations, 8 figures, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Notation, Communication, & Control
  5. Epistemic Logic
  6. Problem Formulation
  7. Approach
  8. The Centralized Planning Algorithm
  9. Interaction Reward Mechanism
  10. Belief & Empathy Propagation
  11. Epistemic Replanning
  12. Epistemic Updates
  13. Balanced Workload Partitioning
  14. Monte Carlo Tree Search
  15. Simulations
  16. ...and 3 more sections

Figures (8)

  • Figure 1: Pictorial representation of the problem presented in the paper. The green robot fails, and the blue robot observes that its belief is false. The blue robot routes to share this information with the red robot, reallocating tasks in the environment before searching for the green robot.
  • Figure 2: Diagram of the proposed approach. The contributions of this paper are within the green box.
  • Figure 3: Graphical representation of chromosome encoding.
  • Figure 4: Examples of tasks generated as a result of belief updates
  • Figure 5: Ideal mTSP allocation for 4 robots is shown in (a) and (b) is the solution with our proposed method.
  • ...and 3 more figures