# Research Topic

Robust Nonparametric Change Detection in Cognitive Communication Systems

# Introduction

Traditionally, the radio frequency (RF) spectrum is divided into distinct frequency bands by some regulatory authority which subsequently proceeds to grant exclusive rights of use to licensees, usually upon payment of considerable royalties. In the face of today’s ever increasing use of wireless technologies, this rigid spectrum allocation approach may lead to signiﬁcant problems of spectrum scarcity as the task of assigning intrinsically limited radio frequency resources to an increasing number of users demanding llarge bandwidths eventually becomes futile. In order to avoid this future spectrum shortage, the concept of cognitive radio (CR) has lately sparked remarkable interest among researchers and led to a substantial number of published academic work.

The basic principle of cognitive radio centers around the fact that the electromagnetic spectrum turns out to actually be underutilized due to ﬁxed spectrum allocation. That is, the distribution of power over time iin numerous segments of the commercially used electromagnetic spectrum is actually all but uniform resulting in the occurrence of “spectrum holes” .

In conventional communication and broadcasting systems these spectral vacancies are neither monitored nor made available to opportunistic/secondary users or used in any other productive way. In other words, precious resources are wasted.

Cognitive radio stems from the consideration that, if there were to be devices with some degree of awareness of the status of the surrounding  electromagnetic environment they operate in, it may become feasible to replace ﬁxed, nonadaptive spectrum management with a dynamic spectrum allocation approach, thereby opening the possibility to end the wastage of resources by making use of the intermittent spectral vacancies. Simply put, a cognitive radio would sense for spectrum holes and make use of them until the actual licensee returns and demands back its frequency band.

# Research Goals

Essential to the feasibility of CR is the ability to quickly and reliably detect changes in a certain frequency band of interest, particularly the on- and offset of transmissions. This crucial ability has to be preserved under non-ideal, harsh conditions as well in order for the CR to work in real world scenarios.

The general aim of this research project is therefore to implement and assess existing and explore and develop novel statistical signal processing techniques to achieve the above mentioned goal. More specifically, the general aim of this research project is to develop novel statistical signal processing techniques to achieve both robust and agile sequential detection for local spectrum sensing in cognitive communication systems.

# Key Research Area

Communication Systems; Multi-Scale Modeling and Simulation

# Supervisors

Prof. Dr.-Ing. Abdelhak M. Zoubir, Signal ProcessingProf. Dr. rer. nat.  Frank Aurzada, Stochastics

# Contact

Michael Lang
M.Sc.

 Address: Dolivostraße 15 D-64293 Darmstadt Germany Phone: +49 6151 16 - 24393 Fax: +49 6151 16 - 24404 Office: S4|10-230 Email: mlang (at) gsc.tu...