On The Iterative Detection Algorithm For Turbo-BLAST

Space-time coding technique, which can achieve improved spectral efficiency by exploiting multiple antennas, has inspired many interests in communication area due to its great potential ability in realizing the wide-band wireless communication in the future.Layered space-time codes (LASTC) can increase the data rate without requiring any extra power and bandwidth. However, since no diversity gain is obtained by LASTC, it is commonly used in conjuction with other channel codes. Such a concatenation is denoted as Turbo-BLAST in many interatures.This dissertation first describes the background of MIMO systems as well as the development of concatenated space-time codes. Both the unconstrained capacity of MIMO system and constrained capacity of BLAST are studied.Then, the encoder structure and detection algorithm for BLAST are described in detail, followed by the turbo principle of Turbo-BLAST. Several soft-input soft-output iterative detections for Turbo-BLAST are also presented, with emphasis on the maximum a posterior detection and the soft sphere decoding.Finally, two low complexity iterative detection algorithms for Turbo-BLAST are proposed. The first, which is based on the Gaussian approximation principle, is a modified sequent Gaussian approximation algorithm(MGA). Avoiding the matrix inverse operation, the MGA algorithm has lower complexity than the available Gaussian approximation algorithm. Furthermore, the proposed algorithm can be employed in the system with more transmit antennas than receive antennas. The second algorithm is the group Gaussian approximation combined with ordering and signal interference cancellation (GGA-Ordering-SIC). Consisting of group Gaussian approximation, reduced dimensional MAP detection, ordering and signal interference cancellation, the GGA-Ordering-SIC can be used in the system with any modulation scheme. Its complexity is even lower than the previous MGA algorithm.