Subspace-based detection for space-time block-coded co-operative diversity

A novel subspace-based detection scheme for space-time block-coded co-operative diversity is presented. By applying singular value decomposition to the received signal matrix, noise subspace orthogonal to a transmitted codeword is obtained. Detection without channel state information is achieved by searching for the codeword, including columns with the best orthogonal relationships to noise subspace. Such detection requires only the number of decoding relays and is robust to the fading channels varying from block to block, as verified by Monte-Carlo simulations.