Blind Unique Channel Identification of Alamouti Space-Time Coded Channel via a Signalling Scheme

Abstract

<p> In this thesis, we present a novel signalling scheme for blind channel identification of Alamouti space-time coded (STBC) channel and a space-time coded multiple-input single-output (MISO) system under flat fading environment. By using p-ary and q-ary PSK signals (where p and q are co-prime integers), we prove that a) under a noise-free environment, only two distinct pairs of symbols are needed to uniquely decode the signal and identify the channel, and b) under complex Gaussian noise, if the pth and qth order statistics of the received signals are available, the channel coefficients can also be uniquely determined. In both cases, simple closed-form solutions are derived by exploiting specific properties of the Alamouti STBC code and linear Diophantine equation theory.</p> <p> When only a limited number of received data are available, under Gaussian noise environment, we suggest the use of the semi-definite relaxation method and/or the sphere decoding method to implement blind ML detection so that the joint estimation of the channel and the transmitted symbols can be efficiently facilitated. Simulation results show that blind ML detection methods with our signalling scheme provide superior normalized mean square error in channel estimation compared to the method using only one constellation and that the average symbol error rate is close to that of the coherent detector (which necessitates perfect channel knowledge at the receiver), particularly when the SNR is high.</p>