Skip to Main Content
We consider the problem faced by a passive receiver in determining the symbol stream that has been output from a wireless transmitter array encoded as Orthogonal Space-Time Block Codes (OSTBC). An informed receiver will usually be able to perform maximum likelihood channel estimation using knowledge of a symbol training sequence embedded in the transmitted data. An uninformed receiver or eavesdropper may not have this prior knowledge and will therefore be unable to directly estimate the channel. In this paper we develop an approach based on the concept of Blind Source Separation (BSS) which is a technique that has application in a wide range of source separation problems where the sources and mixing matrix are unknown. We utilize a BSS optimization algorithm which allows interchangeability of the cost function that determines when the best estimate of the channel matrix has been found. Two cost functions are implemented in this algorithm: JADE and Mutual Information (MI) together with their associated complex gradients. The channel estimate obtained by BSS allows for the estimation of the sources which are assumed to be mutually independent in the BSS model. We show how BSS methods may be applied to estimating symbol streams from OSTBC which are inherently correlated. Further improvements to the estimated symbol streams may be obtained by exploiting knowledge of the encoding scheme and assuming that the receiver is able to correctly synchronize with the transmitted signals. A method for decoding the receiver observed data is presented and we show how this simplifies and improves the BSS symbol stream estimates.
Date of Conference: 13-15 Dec. 2010