Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Cooperative cognitive radios with diffusion networks

: Cavalcante, R.L.G.; Stanczak, S.; Yamada, I.


Alpcan, T.:
Mechanisms and games for dynamic spectrum allocation
Cambridge: Cambridge University Press, 2014
ISBN: 978-1-107-69036-3
ISBN: 1-107-69036-6
ISBN: 978-1-107-03412-9
ISBN: 1-107-03412-4
ISBN: 1-139-52442-9
ISBN: 978-1-139-52442-1
Book Article
Fraunhofer HHI ()

Introduction Reliable estimation of primary users is of paramount importance for the wide acceptance of cognitive radios, and recently a great deal of effort has been devoted to the development of cooperative spectrum sensing techniques [1, 11, 51, 59]. In cooperative methods, secondary users form a wireless network, and they cooperatively detect the presence of primary users by sensing or even probing the desired channels. Then, by using separate control channels, the secondary users exchange information with the intent to arrive at a reliable conclusion of whether primary users are active or not. One of the main advantages of cooperative approaches is the resilience gained by spatial diversity against small-scale deep fades of the signal of the primary users, which constitute one of the major limiting factors of traditional detection schemes [59, 61]. Unfortunately, spatial diversity comes at the expense of communication overhead, which can be a serious burden in larg e-scale systems because of coordination and reliability issues. In addition, current systems typically spent a great part of the energy budget on communication [50], so using the available wireless control channel parsimoniously while attaining good detection performance is one of the main objectives of spectrum sensing algorithms. With the above observations in mind, we show in this chapter recent algorithms for cooperative spectrum sensing that require simple communication protocols among secondary users. In particular, the focus is on distributed data fusion schemes where secondary users exchange information with only a few local neighbors in a scalable way. In more detail, we start by briefly showing that, in cooperative spectrum sensing, many solutions that are optimal in some statistical sense can be posed as the computation of functions, typically weighted averages, where the argument is dispersed throughout the network. In light of this observation, to compute functions in a network efficien