Transceiver Design for Clustered Wireless Sensor Networks --- Towards SNR Maximization

TL;DR

This paper proposes novel iterative algorithms for joint transceiver design in clustered MIMO wireless sensor networks to maximize SNR, considering noisy sensing and transmission, with analysis of efficiency and convergence.

Contribution

It introduces three new iterative algorithms for joint transceiver design in clustered sensor networks, optimizing SNR under realistic noise and fading conditions.

Findings

Algorithms effectively maximize SNR in simulated scenarios

Different algorithms offer trade-offs between efficiency and optimality

Convergence of algorithms is theoretically validated

Abstract

This paper investigates the transceiver design problem in a noisy-sensing noisy-transmission multi-input multi-output (MIMO) wireless sensor network. Consider a cluster-based network, where multiple sensors scattering across several clusters will first send their noisy observations to their respective cluster-heads (CH), who will then forward the data to one common fusion center (FC). The cluster-heads and the fusion center collectively form a coherent-sum multiple access channel (MAC) that is affected by fading and additive noise. Our goal is to jointly design the linear transceivers at the CHs and the FC to maximize the signal-to-noise ratio (SNR) of the recovered signal. We develop three iterative block coordinated ascent (BCA) algorithms: 2-block BCA based on semidefinite relaxation (SDR) and rank reduction via randomization or solving linear equations, 2-block BCA based on iterative second-order cone programming (SOCP), and multi-block BCA that lends itself to efficient closed-form solutions in specific but important scenarios. We show that all of these methods optimize SNR very well but each has different efficiency characteristics that are tailored for different network setups. Convergence analysis is carried out and extensive numerical results are presented to confirm our findings.