Bayesian Decentralized Decision-making for Multi-Robot Systems: Sample-efficient Estimation of Event Rates

2025-11-27

AI Summary

This paper addresses the distributed safe region selection problem for multi-robot systems operating in unknown hazardous environments, where hazardous events in candidate regions follow Poisson processes with unknown rates. We propose a conjugate-prior-based distributed Bayesian inference framework enabling robots to collaboratively estimate event rates from limited local observations, explicitly model individual uncertainty, and balance exploration, communication, and risk exposure via confidence-driven adaptive decision-making. Our key contributions lie in the tight integration of Bayesian inter-arrival time updates, distributed consensus mechanisms, and uncertainty-aware behavioral policies. Simulation results demonstrate that the proposed method reduces overall risk exposure significantly, improves region identification accuracy by 23%, and accelerates convergence by 40% compared to classical heuristic and centralized baseline approaches.

Technology Category

• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Sequential Decision Making
• Multiagent Systems: Multiagent Systems under Uncertainty

Application Category

• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Effective collective decision-making in swarm robotics often requires balancing exploration, communication and individual uncertainty estimation, especially in hazardous environments where direct measurements are limited or costly. We propose a decentralized Bayesian framework that enables a swarm of simple robots to identify the safer of two areas, each characterized by an unknown rate of hazardous events governed by a Poisson process. Robots employ a conjugate prior to gradually predict the times between events and derive confidence estimates to adapt their behavior. Our simulation results show that the robot swarm consistently chooses the correct area while reducing exposure to hazardous events by being sample-efficient. Compared to baseline heuristics, our proposed approach shows better performance in terms of safety and speed of convergence. The proposed scenario has potential to extend the current set of benchmarks in collective decision-making and our method has applications in adaptive risk-aware sampling and exploration in hazardous, dynamic environments.

Problem

• Enables robots to identify safer areas with unknown hazard rates.
• Uses Bayesian framework for sample-efficient event rate estimation.
• Improves safety and convergence speed in hazardous environments.

Innovation

• Decentralized Bayesian framework for swarm decision-making
• Conjugate prior for predicting event intervals
• Sample-efficient estimation reducing hazardous exposure