Robust Cooperative Spectrum Sensing for Disaster Relief Networks in Correlated Environments

TL;DR

This paper introduces a cluster-based cooperative spectrum sensing scheme for disaster relief networks that adapts to environmental conditions and node correlations, enhancing spectrum detection accuracy in cognitive radio ad-hoc networks.

Contribution

It proposes both centralized and decentralized adaptive cooperative sensing schemes tailored for disaster relief networks, improving spectrum sensing performance by leveraging environmental and correlation information.

Findings

The proposed scheme outperforms blind Round Robin sensing in accuracy.

Centralized and decentralized approaches can be combined for robustness.

Adaptive sensing schemes improve spectrum estimation in correlated environments.

Abstract

Disaster relief networks are designed to be adaptable and resilient so to encompass the demands of the emergency service. Cognitive Radio enhanced ad-hoc architecture has been put forward as a candidate to enable such networks. Spectrum sensing, the cornerstone of the Cognitive Radio paradigm, has been the focus of intensive research, from which the main conclusion was that its performance can be greatly enhanced through the use of cooperative sensing schemes. To apply the Cognitive Radio paradigm to Ad-hoc disaster relief networks, the design of effective cooperative spectrum sensing schemes is essential. In this paper we propose a cluster based orchestration cooperative sensing scheme, which adapts to the cluster nodes surrounding radio environment state as well as to the degree of correlation observed between those nodes. The proposed scheme is given both in a centralized as well as in a decentralized approach. In the centralized approach, the cluster head controls and adapts the distribution of the cluster sensing nodes according to the monitored spectrum state. While in the decentralized approach, each of the cluster nodes decides which spectrum it should monitor, according to the past local sensing decisions of the cluster nodes. The centralized and decentralized schemes can be combined to achieve a more robust cooperative spectrum sensing scheme. The proposed scheme performance is evaluated through a framework, which allows measuring the accuracy of the spectrum sensing cooperative scheme by measuring the error in the estimation of the monitored spectrum state. Through this evaluation it is shown that the proposed scheme outperforms the case where the choice of which spectrum to sense is done without using the knowledge obtained in previous sensing iterations, i.e. a implementation of a blind Round Robin scheme.