Real-valued modeling and channel estimation for transmissions based on orthogonal STBCs

The paper addresses two aspects related to orthogonal space-time block codes (STBCs). First, it is shown how to embed orthogonal STBCs into a truly linear matrix transmission model using real-valued notation. Thus, orthogonal STBCs may be interpreted as linear spreading in space and time. It simplifies analytical analysis considerably and puts them on a common basis with other multi subchannel schemes. Second, we consider channel estimation for orthogonal STBCs. We provide a brief analytical analysis on the degradation that is due to pilot based maximum likelihood channel estimation. Further, we exploit the fact that Alamouti's scheme does not only provide a simple maximum likelihood estimate of the data, but may also be deployed to obtain independent estimates of the MIMO channel impulse responses by a linear transformation simply using arbitrary detected data symbols. It is also shown how this fits into the framework of real-valued notation.