Distributed Linear Parameter Estimation: Asymptotically Efficient Adaptive Strategies

TL;DR

This paper introduces a distributed adaptive estimation method for multi-agent networks that achieves asymptotic efficiency comparable to centralized estimators, despite partial information and unpredictable communication.

Contribution

It develops a mixed time-scale stochastic procedure that ensures asymptotic efficiency in distributed linear parameter estimation under weak assumptions.

Findings

Estimates are asymptotically efficient with asymptotic covariance matching centralized estimators.

The proposed method handles partial local information and unreliable communication.

The approach is based on convergence of non-Markovian stochastic approximation procedures.

Abstract

The paper considers the problem of distributed adaptive linear parameter estimation in multi-agent inference networks. Local sensing model information is only partially available at the agents and inter-agent communication is assumed to be unpredictable. The paper develops a generic mixed time-scale stochastic procedure consisting of simultaneous distributed learning and estimation, in which the agents adaptively assess their relative observation quality over time and fuse the innovations accordingly. Under rather weak assumptions on the statistical model and the inter-agent communication, it is shown that, by properly tuning the consensus potential with respect to the innovation potential, the asymptotic information rate loss incurred in the learning process may be made negligible. As such, it is shown that the agent estimates are asymptotically efficient, in that their asymptotic covariance coincides with that of a centralized estimator (the inverse of the centralized Fisher information rate for Gaussian systems) with perfect global model information and having access to all observations at all times. The proof techniques are mainly based on convergence arguments for non-Markovian mixed time scale stochastic approximation procedures. Several approximation results developed in the process are of independent interest.