Distributed Blind Adaptive Algorithms Based on Constant Modulus for Wireless Sensor Networks

Abstract

In this paper, we propose and study the distributed blind adaptive algorithms for wireless sensor network applications. Specifically, we derive a distributed forms of the blind least mean square (LMS) and recursive least square (RLS) algorithms based on the constant modulus (CM) criterion. We assume that the inter-sensor communication is single-hop with Hamiltonian cycle to save the power and communication resources. The distributed blind adaptive algorithm runs in the network with the collaboration of nodes in time and space to estimate the parameters of an unknown system or a physical phenomenon. Simulation results demonstrate the effectiveness of the proposed algorithms, and show their superior performance over the corresponding non-cooperative adaptive algorithms.Index Terms-Distributed adaptive algorithms; wireless sensor networks; incremental cooperative strategy; constant modulus criterion.

Diffusion Adaptation over Multi-Agent Networks with Wireless Link Impairments

“…Recall that ψ ℓk,i is related to ψ ℓ,i via (2). Moreover, since each node k uses data from nodes ℓ ∈ N k whose instantaneous SNR, ς ℓk (i), exceeds the threshold ς o ℓk , then we need to further adjust (9) and replace N k and a ℓ,k , respectively, with N k,i and a ℓ,k (i). This leads to:…”

“…Owing to their decentralized processing structure, simplicity of implementation, and adaptive learning capabilities, these algorithms are particularly well-suited for applications involving multi-agent wireless networks, where energy and radio resources are generally limited [? ], [8], [9]. Consensus strategies can also be used for distributed estimation purposes [10]- [15].…”

“…The problem of link imperfection was also investigated in other classes of distributed algorithms, such as consensus [28]- [32] and subgradient algorithms [9], [33]. In [28], [29] and [33], the authors have examined the performance of consensus algorithms over networks with link failures, where links are established according to some predefined probabilities.…”

Algoritmo Set-Membership CMA no modo Difusão Aplicado a Redes de Sensores

“…R. Abdolee e B. Champagne [7], foram os pioneiros neste contexto, adaptando o algoritmo do módulo constante (CMA) para o contexto de redes de sensores, mas ficando restrito ao caso de redes com topologia em anel. Em [8], tal técnica foi estendida para redes com qualquer topologia.…”

“…Assim, considerando os resultados já obtidos em [8] [7,8]. Consideraremos que a topologia não varia no tempo, todos os nós têm poder de processamento local e trocam informações com seus nós vizinhos adjacentes, ou seja, cada nó k só se comunica com os nós vizinhos diretamente conectados a ele.…”