POLARIZATION DIVERSITY IN INDOOR SCENARIOS: AN EXPERIMENTAL STUDY AT 1.8 AND 2.5 GHZ Susana Loredo 1 , Benito Manteca 2 , Rafael P. Torres 2 1 University of Oviedo, Campus de Viesques, Edif. departamental 4, 33204 Gijón, Spain, sloredo@tsc.uniovi.es 2 University of Cantabria, Avda. de Los Castros, 39005 Santander, Spain, torresrp@unican.es Abstract – This communication presents the results of a set of measurements carried out in order to research the behaviour of a scheme of polarization diversity in indoor environments in the 1.8 and 2.5 GHz bands. To this end, parameters such as the cross-correlation coefficient and the cross-polarization discrimination between the individual signals received in each of the combiner branches are analysed. By simulating various types of combiners, the improvement produced by the combination in the statistical behaviour of the signal, and the decrease of the bit error rate after the combination are also analysed. Keywords – indoor wireless communications, polarization diversity, indoor radio-channel measurements. I. INTRODUCTION Several techniques have been developed in order to reduce the fading effects to which the typical mobile communications radio signal is subjected, which can cause serious deterioration in the operability of both analogue and digital systems. Among these techniques, diversity reception is one of the most effective: when the signals that reach the different branches of the receiver are non-correlated and have similar average power levels, they can be suitably combined, making the signal available for higher percentages of time. In the case of digital communications, the bit error rate is substantially reduced for the same transmitted power and therefore, in applications such as IEEE 802.11, Bluetooth, etc. the transmitted power, for a given BER at the receiver, can be lower. Up to now, the most widely used diversity scheme has been spatial diversity. However, in indoor picocell environments, the space for the location of the antennas is normally limited and the implantation of a spatial diversity scheme may not always be easy or appropriate. In those circumstances, polarization diversity may prove to be a far more attractive option: on the one hand, the number of obstacles that usually exist in indoor scenarios contribute to the depolarisation and decorrelation of the signal, giving rise to a certain coupling of energy into the orthogonal polarization; on the other hand, it allows two independent multipath signals to be received without the need for any spatial separation between the two antennas. Furthermore, in certain situations, polarization diversity might be superimposed on spatial diversity. This communication presents some of the results obtained in the measurement campaign carried out in The Higher Technical School of Industrial Engineering and Telecommunications of The University of Cantabria in order to quantify the advantages of using polarization diversity in indoor environments in the bands of 1.8 and 2.5 GHz. The various parameters that characterize a diversity scheme are analysed, such as the cross-correlation coefficient and the cross-polarization discrimination, and a study is also made of the improvements produced through the combination in the statistics of the envelope amplitude, in the second order statistics and in the bit error rate or BER. II. MEASUREMENT SYSTEM AND SCENARIOS The measurement system [1] used to carry out the measurements is completely automated and governed by a computer, which exercises the functions of system controller and synchronizes the displacement of the receiving antenna with the acquisition of samples of the multipath signal. The vector signal analyser HP 89441A, provided with an internal RF source, was used both as transmitter and receiver. The CW signal generated by the source was radiated between a pair of identical omni-directional wideband antennas vertically polarized. The transmitting antenna was placed 2.10 meters above the floor and was located at a fixed position for every set of measurements. The height of the receiving antenna was 1.5 meters and it was mounted on an automatic displacement system, along which the antenna moved slowly at constant velocity covering a linear trajectory two-meter long. The vector signal analyser captured the received signal in the time domain, collecting samples every 0.032 mm., which were transferred to the computer via the bus HP-IB. A decimation was performed to select the samples separated by a distance equal to λ/8. The measurements were carried out in two scenarios in the Higher Technical School of Industrial Engineering and Telecommunications. The first scenario (Scenario 1) is a typical office environment (Fig. 1). The transmitting antenna (Tx) was located at the middle of a narrow U-shape corridor and three different trajectories were measured for the receiving antenna. In the first of these (to which we shall refer as trajectory T1), of six meters in length, the receiver moved away from the transmitter, there being at all times direct line of sight between them. Then the receiver turned a 0-7803-7589-0/02/$17.00 ©2002 IEEE PIMRC 2002