Session:  Session 2A01A  Smart and Signal Processing Antennas: MIMO (04a) 
Type:  Oral Antenna 
Date:  Tuesday, November 07, 2006 
Time:  08:30  12:20 
Room:  Athena 
Chair:  A. Alexiou & Winters 
Cochair:  
Remarks: 
Seq  Time  Title  Abs No  
1  08:30 
Reconfiguration Techniques to Enhance Cognitive Networks Efficiency
Alexiou, A.^{1}; Dimitrakopoulos, G.^{2}; Demestichas, P^{2} ^{1}Bell Labs, Lucent Technologies, UNITED KINGDOM; ^{2}University of Piraeus, GREECE The next generation of communication and information technologies invokes novel concepts, such as "cognition", aiming at supporting ubiquitous wireless connectivity in a cost efficient manner. Cognition renders the continuous adaptation to the environment claims compulsory, a necessity which can be importantly facilitated through the networks' reconfiguration of their infrastructure and/or operational parameters. In this respect, network elements (reconfigurable base station transceivers) could be properly reconfigured, in oredr to adapt to traffic variations and thus enhance cognitive networks efficiency. This paper discusses exactly such issues, by considering the possibility of evaluating versatile reconfigurations, in order to enhance their capacity and coverage capabilities and thus intelligently manage a cognitive network. The respective problem is mathematically analyzed and solved by means of an efficient algorithm. Finally, simulation results are indicated for validation reasons and potential future directions are outlined. 

2  08:50 
Frobenius Norm Based Receive Antenna Subarray Formation for MIMO Systems
Theofilakos, P.; Kanatas, A. G. University of Piraeus, GREECE
Introduction: Antenna Subarray Formation (ASF) has been recently proposed as an alternative method to Antenna Selection that reduces hardware complexity while offering increased data rates. With this method each RF chain is not allocated to a single element but instead to the combined and weighted output of a subarray of elements. Despite the need for as many vgLNAs and phase shifters as the number of antenna elements, the proposed method achieves decreased receiver hardware complexity, since less downconverters and analogtodigital converters (ADCs) are required with respect to the full system. System Model and Capacity with Receive Antenna Subarray Formation: We assume a flat fading, spatial multiplexing MIMO system with M_{r} receivers, M_{t} transmitters (let M_{t}<M_{r}) and channel matrix H, which is perfectly known at the receiver, but unknown at the transmitter. Furthermore, we assume uncorrelated transmitted signals and zeromean white additive Gaussian noise at the receiver. The combined effect of the propagation channel and the receiver antenna subarrays on the transmitted signal is described by the effective channel matrix The capacity of the resulted system can be lower bounded by Equality holds in the low SNR regime. Proposed Algorithm: Let S_{j} denote the set of antenna element indices that participate in the jth subarray and let H=U∑V^{H} be the singular value decomposition of channel matrix H. The entries a_{ij} of matrix A prior to normalization equal u_{ij} if iεS_{j} or 0 otherwise. where h_{i}=f_{i}^{H} and f_{i} denotes the ith row of H. The proposed algorithm appoints the receiver antenna elements to the appropriate subarray so that the upper bound is maximized. The algorithm is demonstrated in Table 1. Simulation Results and Conclusions:As demonstrated in Figures 12, the proposed algorithm offers increased outage capacity (with respect to the Gorokov decremental selection algorithm), while drastically reducing computation time.


3  09:10 
An Efficient Channel Estimation Scheme for Realistic SFBC MISOOFDM Systems
Marousis, A.D.; Constantinou, P. National Technical University of Athens, GREECE
This paper presents the design and evaluation of a channel estimation scheme that is efficient by means of both the mean square error (MSE) of channel estimation/tracking and its incorporation in a real MISO system. The evaluation has been performed over the spatial channel model developed for MIMO simulations according to 802.16e case of 3GPP.25.996, taking also into account all IF and RF stages in the communication chain. Orthogonality has been applied in spacefrequency dimension for both preamble and pilot symbols, as well as for the data symbols, with the application of Alamouti’s scheme. In 4G multicarrier systems that use spacetimefrequency coding, orthogonal design turns into a key factor for the performance of the system since the channel has to remain about constant during the transmission of one orthogonal block, something which becomes quite challenging in highly timevariant propagation channels. Furthermore, spacefrequency block coding (SFBC) becomes more efficient as the number of subcarriers increases. At the transmitter, the data are scrambled, encoded (ReedSolomon, convolutional), and mapped, giving real information data rates of 6.9Mbps (BPSK1/2), 13.8Mbps (QPSK1/2), 20.7Mbps (QPSK3/4), 27.7Mbps (16QAM1/2), 41.5Mbps (16QAM3/4), 55.3Mbps (64QAM2/3), and 62.2Mbps (64QAM3/4) in a 20MHz of bandwidth. The mapped data are multiplexed with the pilot symbols occupying 3.12% of an OFDM symbol arranged in blocks of 2 neighbouring subcarriers in predefined positions. The formed symbol stream enters the SFBC stage giving 2 output streams. Each symbol stream is splitted in packets of 200 symbols, forming the input to a 256FFTOFDM chain. The outputs are multiplexed with preambles for channel estimation and frequencytime synchronization. Finally, the two independent streams are converted to analog signals, filtered, and upconverted to a carrier frequency of 5.2GHz. On the receiver side, there had been only 1 activated antenna (2x1), wherein the opposite procedure has been followed applying finally Viterbi decoding. Each frame carries 2400 information bits and the evaluation is performed on the basis of achieving BER relative variance of 0.0001 with an upper limit of 1000 frames. Channel state information (CSI) is acquired by the receiver on a twostep procedure and no CSI is fed back to the transmitter, establishing an openloop system with equal transmission power per antenna. The 1st step in the channel estimation procedure is the use of a preamble OFDM symbol where the symbols are orthogonal on a SFBC subcarrier basis. The estimation has been implemented using a MMSE approach. In the 2nd step the pilot symbols are used only for the phase estimation (caused by Doppler effects) and using interpolation the correction factor for each subcarrier is taken into account in the preamble based estimation. These values are used for both channel compensation and soft decision stages. In fig.1 the performance of the whole system is depicted, indicating very good efficiencies at low Eb/No values for channel type A of 802.16e case. In fig.2 the MSE of channel estimation scheme is given making obvious the uniform performance of the estimator achieving a MSE of 5dB at Eb/No=10dB for all schemes. More results are going to be presented for various propagation channels and degrees of channel estimation depth, as well as a detailed description of all estimation and receiver stages is going to be addressed. 

4  09:30 
Reconfigurability to Antenna Correlation: System Level Performance Enhancements in HSDPA
Peppas, K.^{1}; Lazarakis, F.^{1}; Alexiou, A.^{2}; Alexandridis, A.^{1}; Dangakis, K.^{1} ^{1}Institute of Informatics & Telecommunications, NCSR “Demokritos”, Athens, GREECE; ^{2}Bell Labs, Wireless Research, Lucent Technologies, Swindon, UNITED KINGDOM In this paper the system level performance of an HSDPA MIMO network, which combats the effects of antenna correlation by applying novel reconfigurable techniques, is investigated. The impact of antenna correlation on adaptive array techniques has been thoroughly investigated and it has been shown that fading correlations reduce MIMO channel capacity and system performance. Nevertheless, considerable capacity and performance gains can be obtained by transmission on the eigenmodes of the transmit antenna correlation matrix. Reconfigurability to antenna correlation is achieved at the link level by applying a linear precoding scheme at the transmitter spacetime block encoder, which is based on the well known Alamouti scheme. The linear precoder is determined so as to minimise a given criterion, such as an upper bound on the pairwise error probability (PEP) of a codeword. The proposed reconfigurability scheme performs similarly to spacetime block coding for low antenna correlations and becomes equivalent to beamforming for high antenna correlation. For correlation values between the two extremes, the proposed approach outperforms both conventional schemes. In order to assess the performance of the wireless system enhanced with reconfigurable spacetime coding techniques in a realistic manner, a sufficient number of system level simulations should be carried out. The evaluation of the benefits of MIMO techniques at the system level introduces a number of challenges, such as the requirement for suitable spatiotemporal channel modelling and linktosystem level interface. A softwarebased HSDPA system level simulator has therefore been developed, whose input is specified according to a predefined number of test cases. Each test case consists of specific channel characteristics, system configuration, mobility and user requirements, antenna configurations and spacetime algorithms. Fast fading MIMO channels, HARQ and fast link adaptation are some of the features of the simulation platform. An appropriate linktosystem interface was applied, based on the instantaneous values of the channel capacity to achieve a onetoone mapping between a specific channel realization over frame duration and the probability of frame error. The objectives of the system level simulations are: i) To demonstrate the impact of the variation of critical link level parameters, such as the antenna correlation, on the performance of the HSDPA network and ii) to evaluate the enhancement in the system performance achieved by the application of reconfigurable MIMO techniques.
The adopted performance metrics of interest are the system throughput and the satisfied users. Web and FTP users are simulated. System level simulations were carried out for ideal, intermediate and high antenna correlation conditions with nonreconfigurable and reconfigurable techniques. The degradation of the system performance was demonstrated for increasing antenna correlation. In Figure 1, the number of satisfied users for the reconfigurable and the nonreconfigurable case is depicted for 5345 users. From Figure 1, one can see that the proposed scheme effectively mitigates the performance degradation due to antenna correlation. 

5  09:50 
Application of SDMA with Scheduling for MIMO MCMCDMA
Rodriguez, J.; Monteiro, V.; Gameiro, A. Instituto de Telecomunicaçoes, PORTUGAL
In order to utilise efficiently the physical spectral resources, we propose a Dynamic Resource Allocation (DRA) scheme that ensures the seamless transport of variable size IP packets over the airinterface. The DRA constitutes link adaptation, HARQ (Hybrid Automatic Repeat ReQuest) type II and priority scheduling with adaptive beamforming to provide an SDMA (Space Division Multiple Access) component. The combination of these radio resource management (RRM) entities ensures that variable size IP packets are mapped to a predefined set of radio block sizes, and are transported over the airinterface according to the HARQ protocol timing sequence. The use of adaptive beamforming with scheduling adds a spatial dimension to the resource space definition, so that users u_{i }and u_{j }separated by an angle δ_{i,j}≥δ_{min } w.r.t the Base Station (BS), have the opportunity to share the same time, and code slot. Furthermore, using crosslayer design techniques, we can ensure that each user is scheduled in a manner so as to ensure the highest average overtheair cell throughput whilst fulfilling the QoS (Quality of Service) requirements. The system level simulation results are obtained using a combineddynamic snapshot simulator, where the mobiles are created at the beginning of each run, and remain active for the complete run duration. Furthermore, the pathloss and shadowing values remain constant within a run, whilst the fast fading is updated on each TTI (transmission time interval). Simulations were conducted in indoor environment where each run corresponds to 300 seconds of real time. Preliminary results have shown that 42Mbps overtheair (OTA) throughput can be attained per spatial beam, with 37 Mbps service throughput; where the OTA represents the total number of transmitted bits during the simulation time, and the service throughput refers to the total number of correctly received bits. In summary, this paper evaluates the potential benefits of SDMA at the system level assuming an underlying 4x2 MIMO MCCDMA airinterface, and results have shown that adaptive beamforming with scheduling can provide a solution inline with the anticipated throughputs defined in the 4MORE/B3G cellular scenarios. 

6  10:40 
Comparison of MIMO Single and MultiPolarized Measured Channels in Indoor WLAN Scenarios
GarciaGarcia, L.; GomezCalero, C.; MoraCuevas, J.; Martinez, R.; de Haro, L. Universidad Politecnica de Madrid, SPAIN MultipleInput MultipleOutput (MIMO) systems have been introduced as a solution in order to increase the data rate in wireless communication systems. The MIMO channel capacity grows linearly with the number of the antenna at each side of the radio link. Dual polarized antennas are presented as a solution when spacing is limited in the user terminal. Few studies have been carried out taking into account multipolarized antennas in MIMO channels, despite the interesting advantages such as reduction in size that they can provide. Measurements with different single and dual polarization antennas are presented in this paper. The measurements have been performed with a 4 x 4 MIMO testbed for WLAN frequency. A photograph of the receiver modules of the used testbed can be seen in Figure 1, and the dualpolarized antennas designed for the measurement campaign are shown in Figure 2. The architecture of the testbed is based on softwareradio platforms where the signals are generated from a PC in an offline mode. Dualpolarized antennas are used in both transmit and receive modules, in order to account for multiple polarization in the system. As far as the MIMO channel characterization is concerned, some channel parameters such as DirectionofArrival (DoA) distribution, spatial correlation and crosspolarization discrimination (XPD) have been analysed for different measured environments. So far most works on MIMO channel characterization focuses on either indoor or outdoor environments. In this study we have also considered an outdoortoindoor environment and its propagation characteristics have been compared to the indoor scenario. DoA is estimated through powerfield measurements campaigns for a MIMO system. Measurements have been accomplished in a multipath propagation environment leading to consider multiple delayed copies of the signal. Therefore, angles as well as delays of the received signals are resolved from the MIMO channel response matrix instead of the input covariance matrix used by the subspace based methods. Also, the effect of using dual polarized antennas in DoA estimation performance is studied using classical algorithms. The well known MUSIC algorithm is extended to estimate DoA parameters. The spatial correlation and crosspolarization discrimination are measured for both nonline of sight and line of sight scenarios for the indoor environment. Since 4 elements are used simultaneously in the measurements, the effect of coupling among elements is also included in the study, which is interesting in order to consider a realistic array in the MIMO system. Finally, some MIMO algorithms will be evaluated for the case of multipolarized MIMO system and the improvement in performance will be presented in the final full paper, comparing the results when dualpolarized antennas are used with the singlepolarized antenna case.


7  11:00 
Evaluation of MIMO Arrays Using Antenna Patterns, Reverberation Chamber, and Channel Measurements
Lindmark, B.^{1}; GarciaGarcia, L.^{2}; Jalden, N.^{1}; Orlenius, C.^{3} ^{1}Royal Institute of Technology KTH, SWEDEN; ^{2}Universidad Politecnica de Madrid, SPAIN; ^{3}Bluetest AB, SWEDEN The performance of a MIMO system depends on a number of factors including the propagation conditions, the type of signaling used, and the amount of feedback from receiver to transmitter. However, it is also strongly dependent on the antenna arrays used and how well suited they are to the propagation conditions at hand. It has been found that mutual coupling may decrease antenna correlation and thus increase capacity in MIMO systems. However, if one also accounts for losses from impedance mismatch and mutual coupling, the overall effect of closely spaced elements with mutual coupling is to reduce the available capacity [1],[2]. The polarization and pattern of the antennas may also influence the MIMO performance if the radio propagation results in a incident field which is not isotropic and of random polarization [3].
Since the antenna arrays affect the theoretical MIMO capacity, it is of interest to be able to evaluate them in an efficient and accurate way. We have studied three ways to evaluate MIMO antenna arrays:
The first of the above is clearly the ideal way to evaluate any antenna, but it is very timeconsuming and there is also some question which environment should be used in the evaluation. The second approach is certainly less timeconsuming, but it calls for knowledge of the full 2D radiation pattern in azimuth and elevation as in e.g. [4]. The third approach circumvents the need for timeconsuming channel measurements or 2D radiation pattern data, but is less flexible in the type of radio channels which may be used in the evaluation. We have studied the MIMO capacity performance of a compact antenna array, antenna under test (AUT), for mobile communication or WLAN applications. The MIMO system performance of the antenna array has been compared to the performance of a reference array (REF) with 4 vertical monopoles. The figure shows channel measurement (1) above with the two arrays mounted on a cart. We find that all investigated methods give a similar result. The compact antennas array provides a slightly worse MIMO performance which may be attributed to its lower radiation efficiency. The difference is, however, quite small and corresponds to an SNR difference of approximately 1 dB.
[1] Waldschmidt, C.; Schulteis, S.; Wiesbeck, W., "Complete RF system model for analysis of compact MIMO arrays," IEEE Trans. Veh. Tech, May 2004


8  11:20 
The Influence of MIMO Terminal User's Hand on Channel Capacity
Zervos, T.^{1}; Alexandridis, A.^{1}; Peppas, K.^{1}; Lazarakis, F.^{1}; Dangakis, K.^{1}; Soras, C.^{2}; Lindmark, B.^{3} ^{1}Institute of Informatics & Telecommunications, NCSR “Demokritos”, Athens, GREECE; ^{2}Department of Electrical & Computer Engineering, University of Patras, Patras, GREECE; ^{3}Department of Signals, Sensors, and Systems, Royal Institute of Technology, Stockholm, SWEDEN During the last years a number of studies have proved that the presence of human body close to a mobile terminal strongly affects the characteristics of the terminal antennas which results in degradation of terminal's performance. In this paper we firstly investigate the effect of user's hand on the characteristics of a compact antenna array which is placed on a PDA. Then, we use this investigation to estimate the impact of user's hand on the channel capacity provided by the specific MIMO antenna system. The estimation of MIMO channel capacity is based on a methodology which uses the radiation patterns of the coupled antenna elements. According to that methodology, the antenna radiation patterns can lead to the estimation of the covariance between elements i and j of the antenna array. Then, covariance matrices are formed which are inserted into a Kronecker model to obtain the MIMO channel matrix H. Finally, H is applied to the well known formula which gives the capacity C of a MIMO channel with M transmit and N receive antennas. The above described methodology has been applied to a 4x4 MIMO system using a PDAlike mobile terminal. This terminal includes a specially designed compact array consisting of 4 patch elements. The elements are folded PIFA patches designed for operation at 1800 MHz and 2450 MHz. In order to study the effect of human body, the "multimedia viewing position" has been selected where it is assumed that the user is holding the terminal while the rest of his/her body does not influence the terminal's operation. Thus, the user's hand has been modeled by means of an electromagnetic simulator following the concept of a homogeneous model. Moreover, the PDA with the compact antenna array was also modeled with the same simulator (see Fig.1). Then, the antenna radiation patterns of the array have been simulated with and without the presence of the user's hand. The obtained Efields were used, according to the methodology described above, to calculate the covariance matrices, channel matrices and finally the channel capacity with and without the effect of the user's hand (see Fig.2). Our investigations proved that the presence of the user’s hand significantly increases the coupling losses especially for those elements that lie closer to user's hand. That kind of coupling has also a major impact on the radiation efficiency of those elements. As a result, channel capacity experiences significant degradation (as can be seen in Fig.2) which should not be neglected when designing a MIMO terminal or evaluating the advantages of MIMO enhancements.


9  11:40 
BER Variation in an UWB MIMO OFDM Communications System Based on Measurements Made in a Picocell Wireless Environment
Chambers, P.; Downing, C.; Baher, H. Dublin Institute of Technology, IRELAND A measurement based simulation of an UWB MIMO OFDM communications scheme which exploits the multipath environment as well as increasing spectral efficiency is presented. Further details, results and photographs will be presented in the full conference paper. 1. Introduction Physical measurements were made using a vector microwave analyser (VMNA) with UWB Antennas. A SIMULINK simulation of an OFDM system was created which incorporated these physical measurements. This simulation was then extended to incorporate the multipleinput/multipleoutput (MIMO) communications systems model described in section II below. 2. A MIMO Communications System Model. In an OFDM communications system, the bandwidth of a multipath communications system is used efficiently by converting a frequency selective communications channel into a series of orthogonal frequency flat subbands. MIMO communications systems exihibit a great degree of spectral efficiency by using multiple element arrays (MEAs) of transmitters and recievers. A standard approach is to apply weight factors to these MEAs, thus creating a series of orthogonal subchannels. These weight factors are derived from singular value decomposition (SVD) of measurements of the channel. 3. UWB Measurement based Analysis of MIMO OFDM. Specifically, the measurements described were made in a specially built enclosure which was rich in multipath signal components. This work focuses on the bit error rates (BERs) which are produced by a SIMULINK simulation analysis which is ultimately based on the physical measurements described in section I. In particular, the BER is examined with respect to the effect of obstructions within the multipath environment characteristic of ultrawide band (UWB) measurments (3.1 GHz  10.6 GHz). The multipath channel was investigated in a 5 x 5 MIMO communications system where the BER for each of its five orthogonal subchannels is plotted below for the cases where the line of sight (LOS) signal path and double reflection signal path were removed. It is clear that removal of the LOS signal path improves the BER. Figure (1): BER variation on each MIMO subchannel. BER=0 for removal of LOS on subchannels 1 to 4 

10  12:00 
On the Design of a Broadband Wireless MIMOOFDM Demonstrator
Nieto, X.; Ventura, L.M. Centre Tecnològic de Telecomunicacions de Catalunya (CTTC), SPAIN
GEDOMIS is a Generic hardware Demonstrator for MultipleInput MultipleOutput (MIMO) Systems which provides a new means for the implementation and test of wireless and cellular multiantenna communication systems. Therefore, it can be considered an heterogeneous and reconfigurable Software Defined Radio (SDR) platform. In this paper, we present a multiantenna wireless localarea network (WLAN) application. Hence, the main goal of this implementation is focused on the design of an experimental broadband wireless 4x4 MIMO orthogonal frequency division multiplexing (OFDM) system based on an extension of the IEEE 802.11a/g physical layer (PHY) to multiantenna scenarios. The paper is divided into two different sections: First, a general description of the hardware architecture of the demonstrator is shown, where we describe the main building blocks of the VMEbased transmit and receive cabinets. The architecture proposed for the RF section of the demonstrator is based on the integration in a singleslot VME board of multiple dualband RF transceivers, together with a multiple output, dualband frequency synthesizer. The multiband capabilities of the frontend are designed to operate on the frequency bands specified in the IEEE 802.11g/a standards: 2.4GHz and 5.2GHz. As for the design of the dualband RF transceivers that build the frontend, we present a new custom, areaefficient, reduced component count architecture achieving a highdegree of integration. In the DSP section of the testbed, we present a hardware solution combining multiFPGA and multiDSP VME boards with highspeed multichannel analogtodigital conversion (ADC) and digitaltoanalog conversion (DAC) modules. Second, we describe the signal processing algorithms implemented at the transmit and receive sides. First, we analyze the standard PHY specification for a singleantenna system and define the modifications introduced in order to extend it to scenarios using transmit diversity. The presented modifications are focused on the design of the preambles using cyclic delay diversity (CDD) techniques, and the rotation of the pilot symbols transmitted from each antenna. No feedback is considered for the design of the transmitter. Instead, orthogonal spacetime block coding (OSTBC) is implemented, and thus, no channelstate information at the transmitter (CSIT) is required. Different OSTBC schemes are supported in the presented implementation, such as the ones proposed by Alamouti, for two antennas, and Tarokh, for three and four antennas (G3, G4, H3 and H4). 