Distributed Estimation with Information-Seeking Control in Agent Network

TL;DR

This paper presents a distributed Bayesian estimation and control framework for agent networks that enhances localization and tracking through information-seeking strategies, suitable for nonlinear, non-Gaussian scenarios.

Contribution

It introduces a novel cooperative estimation and control method combining belief propagation, consensus, and entropy maximization for decentralized agent networks.

Findings

Agents achieve high localization accuracy with minimal anchors.

Simulation demonstrates effective self-localization and target tracking.

Agents exhibit intelligent, cooperative behavior.

Abstract

We introduce a distributed, cooperative framework and method for Bayesian estimation and control in decentralized agent networks. Our framework combines joint estimation of time-varying global and local states with information-seeking control optimizing the behavior of the agents. It is suited to nonlinear and non-Gaussian problems and, in particular, to location-aware networks. For cooperative estimation, a combination of belief propagation message passing and consensus is used. For cooperative control, the negative posterior joint entropy of all states is maximized via a gradient ascent. The estimation layer provides the control layer with probabilistic information in the form of sample representations of probability distributions. Simulation results demonstrate intelligent behavior of the agents and excellent estimation performance for a simultaneous self-localization and target tracking problem. In a cooperative localization scenario with only one anchor, mobile agents can localize themselves after a short time with an accuracy that is higher than the accuracy of the performed distance measurements.