Transceiver design using generalized triangular decomposition for MIMO communications with QoS constraints

Assuming that the channel state information is available at both transmitter (CSIT) and receiver (CSIR), we propose a novel multiinput multioutput (MIMO) transceiver design that combines the generalized triangular decomposition (GTD) with either the zero-forcing VBLAST (ZF-VBLAST) detector or the zero-forcing dirty paper precoder (ZF-DP). Our approach decomposes a MIMO channel into multiple subchannels with prescribed signal-to-interference-and-noise ratios (SINR), useful in MIMO communications with QoS constraints. We show that the GTD scheme is asymptotically (for high SNR) optimal in terms of overall channel throughput for large range of QoS constraints. The numerical experiment verifies that the GTD scheme is much more power efficient than its linear counterpart.