XIX IMEKO World Congress Fundamental and Applied Metrology September 6-11, 2009, Lisbon, Portugal AUTOMATIC SIGNAL RECOGNITION FOR A FLEXIBLE SPECTRUM MANAGEMENT Niclas Björsell 1 , Pasquale Daponte 2 , Luca De Vito 2 , Sergio Rapuano 2 1 ITB/Electronics, University of Gävle, Gävle, Sweden, niclas.bjorsell@hig.se 2 Department of Engineering, University of Sannio, Benevento, {daponte, devito, rapuano}@unisannio.it Abstract - The paper presents the prototype of an automatic digital modulation classifier, to be used for signal recognition in frequency bands managed in a flexible way. The prototype is based on a Data Acquisition System, consisting of an Analog-to-Digital converter embedded in an evaluation board, a frame grabber and a Personal Computer. The modulation classifier is able to recognize the most used digital modulations. An experimental validation of the realized prototype in a radio environment is also provided. Keywords: modulation classification, spectrum management. 1. INTRODUCTION The use of signal classification within radio communication will grove of importance since future usage of the spectrum for wireless communications will be available for a wider variety of services and applications. In today’s environment, electronic communication services are offered over a variety of electronic communications networks (e.g. different types of mobile, fixed and broadcasting networks), using a variety of terminals. The demand for certain services (such as mobile and Internet) has grown far beyond earlier predictions, and developments in radio technology have resulted in more efficient methods of sharing spectrum among different systems and users. Innovation requires rapid access to spectrum for individuals and service providers. This has resulted in a need for greater flexibility in the use of spectrum resources for wireless electronic communications. At the same time, the spectrum originally intended for distinct electronic communication services is now being used for services which compete against each other. This implies that a wide range of electronic communication networks and electronic communication services may be offered on a technology and service neutral basis. However, certain technical requirements to avoid interference should be met, to ensure that the spectrum is used in an effective and efficient way, and that the authorization conditions do not distort competition. The European Union is moving in the direction of a more flexible utilization of the spectrum [1] and, for this purpose, it created a specific work group, called Radio Spectrum Policy Group (RSPG), with the aim of come to a new Wireless Access Policy for European Communications Services (WAPECS). In its opinion [2], the RSPG indicated the technology neutrality and the service neutrality as some of the objectives to reach a more flexible spectrum utilization. In a flexible spectrum management scenario, in which several technologies and services can be present at the same time, an effective spectrum monitoring could be achieved by recognizing the signals which occupy the spectrum. For this purpose, the use of signal classification will play an important role. Moreover, some researchers propose a model of spectrum usage, based on an opportunistic access. It consists of finding a spatial and temporal white space in the spectrum, to dynamically allocate a secondary transmission, without affecting the primary one [3]. In this case, modulation classification could be very useful to estimate the parameters of the primary transmission and, thus, to set those of the secondary one, such that the interference produced on the primary transmission can be kept to a non-harmful level [4]. Moreover, the transmission equipment, able to work in a dynamic spectrum scenario, requires a receiver which has to adapt itself to the transmitted signal, without knowing a priori any physical layer parameter [3]. It is necessary, therefore, a proper module providing a complete description of the radio environment, where the equipment is placed. Such module should be capable of extracting the following information from the received signal: Modulation scheme; Symbol frequency; Carrier frequency; Pulse shaping. Modulation classification, thanks to its applicability to many fields, on both civil and military sector, has gained a relevant interest of the scientific community in the last years. In fact, several researches have been carried out to find reliable methods and algorithms for modulation classification [5, 6]. Most of them, however, support few modulation schemes or require some a priori information, such as the carrier frequency or the symbol rate. Moreover, no papers have been found presenting an instrument prototype and reporting its characterization in a RF (Radio Frequency) environment. In [7], a method for automatic digital modulation classification has been proposed, able to recognize the following modulation schemes: 2PSK (Phase Shift Keying), 4PSK, 8PSK; 568 ISBN 978-963-88410-0-1 © 2009 IMEKO