Adaptive Link Selection Strategies for Distributed Estimation in Wireless Sensor Networks

TL;DR

This paper introduces adaptive link selection strategies for wireless sensor networks that enhance distributed estimation by reducing error and improving convergence through novel algorithms exploiting network topology.

Contribution

It presents two new link selection algorithms, including an exhaustive search and a sparsity-inspired method, to optimize network performance in poor-quality link environments.

Findings

Lower EMSE compared to existing methods

Faster convergence rates

Significant performance improvements in simulations

Abstract

In this work, we propose adaptive link selection strategies for distributed estimation in diffusion-type wireless networks. We develop an exhaustive search-based link selection algorithm and a sparsity-inspired link selection algorithm that can exploit the topology of networks with poor-quality links. In the exhaustive search-based algorithm, we choose the set of neighbors that results in the smallest excess mean square error (EMSE) for a specific node. In the sparsity-inspired link selection algorithm, a convex regularization is introduced to devise a sparsity-inspired link selection algorithm. The proposed algorithms have the ability to equip diffusion-type wireless networks and to significantly improve their performance. Simulation results illustrate that the proposed algorithms have lower EMSE values, a better convergence rate and significantly improve the network performance when compared with existing methods.