Multiple Sources Discrimination by Array Processing G. Di Massa * , S.Costanzo + , A. Borgia * , I. Venneri * , G. Galati + , M. Leonardi + , E. Piracci + * Dipartimento di Elettronica, Informatica e Sistemistica, Università della Calabria 87036 Rende (CS), Italy {dimassa, costanzo, aborgia, ivenneri}@deis.unical.it + Dipartimento di Informatica, Sistemi e Produzione e Centro Volterra, Università degli Studi di Roma “Tor Vergata” Via del Politecnico 1, 00133 Roma, Italy [galati,leonardi, piracci]@disp.uniroma2.it Abstract— An array system solution is adopted in the paper to face the problem of separation of superimposed signals by performing blind source separation. A five channels receiving station has been implemented to provide a large dynamic range, wide bandwidth and fast sampling, with a proper array of six independent antenna elements properly designed to give a sufficiently wide covering pattern. The source separation algorithm is fully described and the experimental results concerning the characterization of the array elements are discussed. I. INTRODUCTION To fully exploit the information in the air-ground 1090 MHz channel, used in Air Traffic Control for surveillance and data link purposes, array processing is an interesting option, whose capabilities in the separation of superimposed signals has been demonstrated [1], [2] by analysis, simulation and evaluation on a limited set of live data. In order to provide a vaster data base with higher quality signals, a five channels (four linear in array configuration with other two, at both sides, on a matched load, i.e. one dummy and one feeding a logarithmic channel) receiving station has been designed at Tor Vergata University, with the fundamental contribution by the Università della Calabria which performed the design, implementation and testing of the six-elements array antenna, whose main results are described. II. THEORY Today, the air traffic control systems make a large usage of the 1090 MHz channel [3], [4]: secondary surveillance radar (SSR), Multilateration and its Wide Area version (MLAT/WAM), Automatic Dependent Surveillance ( with its Broadcast and Retransmit functions : ADS-B and ADS-R), Traffic Information Service-Broadcast (TIS-B), and others. The Multilateration and the Wide Area MLAT (WAM) implement a cooperative independent surveillance of civil airborne and vehicles targets. They measure the Time of Arrival (TOA) at many stations and uses the time difference of arrival (TDOA) of the signals of the 1090 MHz transponders at the various receiving stations. Therefore, they track and identify aircraft and ground vehicles equipped with a 1090 MHz Mode S transponder. Multilateration systems are basically distributed surveillance and identification systems, made up by a network of receiving stations. The distributed sensors with omni-directional (or, anyway, wide beam) antennas may face with a dramatic increase of received replies per unit time, causing overlapping between replies in time domain. When replies overlap, very often the message transmitted by the aircraft is corrupted and cannot be recovered by conventional receivers/decoders within the present-day SSR Mode S stations. In order to extend the surveillance also to the airport service vehicles (and to the Apron), it is necessary to equip the vehicles with a 1090 MHz signals (unsolicited replies or squitter) transmitter with a 1-2 s -1 squitter rate according to ICAO/EUROCAE standards and recommendations. This surveillance extension will strongly increases the 1090 MHz channel traffic when a significant number of vehicles will be equipped. In Multilateration applications, the use of omni directional (or, anyway, wide beam) antennas and the presence of both SSR Mode S replies and squitters makes the superposition of 1090 MHz signals rather probable, not to mention the presence of Mode A/C signals on the same band as well as the above-mentioned emerging surveillance applications. As a consequence of the superposition, the signals can be garbled and the related replies cannot be detected nor localized; their information is basically lost. A possible and novel solution can be obtained by adopting array systems, i.e. a group of several antennas, each of them having an independent receiver which allows us to combine their output digitally. They have the power to perform several tasks, such as the estimation of the sources number, adaptive beam forming or blind source separation. Two main advantages can be recognized to Blind Source Separation over beam forming, namely: 1) no need of antenna calibration; 2) the ability to work without the knowledge of the direction of arrival, i.e. blind. We consider the reception of δ independent source signals on an µ-element antenna array (of arbitrary form). The baseband antenna signals are sampled at a frequency greater than the signal bandwidth and stacked in vectors ν (size µ). We assume d<m. After collecting Ν samples, the observation model is given as: Proceedings of the 5th European Conference on Antennas and Propagation (EUCAP) 620