Reduced ML-Decoding Complexity, Full-Rate STBCs for $2^a$ Transmit Antenna Systems

For an $n_t$ transmit, $n_r$ receive antenna system ($n_t \times n_r$ system), a {\it{full-rate}} space time block code (STBC) transmits $n_{min} = min(n_t,n_r)$ complex symbols per channel use and in general, has an ML-decoding complexity of the order of $M^{n_tn_{min}}$ (considering square designs), where $M$ is the constellation size. In this paper, a scheme to obtain a full-rate STBC for $2^a$ transmit antennas and any $n_r$, with reduced ML-decoding complexity of the order of $M^{n_t(n_{min}-3/4)}$, is presented. The weight matrices of the proposed STBC are obtained from the unitary matrix representations of a Clifford Algebra. For any value of $n_r$, the proposed design offers a reduction from the full ML-decoding complexity by a factor of M^{3n_t/4}}. The well known Silver code for 2 transmit antennas is a special case of the proposed scheme. Further, it is shown that the codes constructed using the scheme have higher ergodic capacity than the well known punctured Perfect codes for $n_r < n_t$. Simulation results of the symbol error rates are shown for $8 \times 2$ systems, where the comparison of the proposed code is with the punctured Perfect code for 8 transmit antennas. The proposed code matches the punctured perfect code in error performance, while having reduced ML-decoding complexity and higher ergodic capacity.