Simple Antenna Indoor Technology (SAINT)

  • Getu, Beza Negash (Project Participant)
  • Andersen, Jørgen Bach (Project Participant)

Project Details


The purpose of the Simple Antenna Indoor Technology (SAINT) Ph.D. project is research and development of simple and low cost antenna technology for wireless communication systems operating at 2.4GHz and higher frequencies. The focus is on indoor wireless communication; however, the techniques should be flexible enough to support indoor and outdoor operation. The goal is to develop and implement promising techniques for increasing overall efficiency at modest complexity and cost. Different eigenmodes of transmission can be acquired for any number of transmitter (M) and receiver (N) antenna elements after appropriate orthogonal weight vectors, which are the results of the singular value decomposition (SVD) of the MIMO channel matrix, H, are applied on both sides. For instance, for a Rayleigh fading scenario, there are two completely independent physical channels with two antenna elements on both sides. Figure 1 shows the resulting pattern for two scatterers located at 58° and 115° in the radio channel between the transmitter and receiver. It shows that the beams have the optimum possible gain towards the scatterers keeping the orthogonality between them (i.e., having nulls pointing towards the interferer). We call these beams, which arise from the excitation of the channel with orthogonal eigenvectors, eigenbeams. This picture particularly shows the analogy with the SDMA. In fact, the MIMO system is another version of the SDMA except that in the former case the antenna elements are co-located in the same terminal at both the transmitter and the receiver. The existence of the two parallel channels can also be explained from this picture. In general, the overall MIMO system can be perfectly equalized with respect to the interference introduced by the wireless channel from the knowledge of the channel at both ends of the link. The only remaining impairment is the noise, which is in most cases white and Gaussian distributed (AWGN). Since the instantaneous SNR at the output varies due to the variation of the eigenvalues, the average error rate performance of the above system can be given as the average error performance in AWGN, appropriately weighted by the distribution of the instantaneous eigenvalues. Observations can tell us that the SNR at the receiver decreases as one proceeds from the first to the last available eigenchannel effectively causing an increase in BER. Therefore, it is important to distribute the total transmitter power among the available channels optimally so that the overall performance of the system is enhanced. Figure 2 shows the average BER for M=2 and N antennas at the transmitter and receiver, (2, N), respectively, when equal power is distributed among the two eigenchannels and BPSK modulation. The poor BER performance for the (2,2) case is due to the low eigenvalue of the second eigenchannel introducing high BER. However, the best possible BER performance can be achieved if only the maximum (largest eigenvalued) eigenchannel is used for transmission but then there will be only a spectral efficiency of 1 bit/s/Hz. For BER of 10-3, there is approximately 20, 18, 14 and 8.5dB gain for N=8, 6, 4 and 3 receiver antennas compared with only having one antenna on both sides, (1,1) case. The two available eigenchannels give also a spectral efficiency of 2bits/s/Hz since there are two possible transmission routes in contrast to the (1,1) case, if BPK modulation is used. However, QPSK modulation has to be used, if the (1,1) case is supposed to give the same spectral efficiency given by the two eigenchannels or if comparison between the two case is to be done for the same data rate/throughput. The project is supported by CSEM , Switzerland. For a more detailed description (Figures, Tables, references) please visit (Beza Negash Getu, Jørgen Bach Andersen)
Effective start/end date01/01/200131/12/2003


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.