Limited Feedback MU Massive MISO Systems with Differential TCQ in Temporally Correlated Channels.

摘要: We propose a differential trellis coded quantization (TCQ) scheme for limited feedback massive multiple-input single-output (MISO) frequency division duplexed systems in temporally correlated channels. consider the multi-user (MU) scenario and derive an approximation to expected signal-to-interference-plus-noise ratio (SINR) with random vector (RVQ) based codebook channel direction information (CDI) matched-filter (MF) precoding at base station (BS). Using SINR approximation, we show that number of bits required by RVQ match even small fraction perfect CDI performance is very large. With such large numbers bits, exhaustive search conventional approaches makes them impractical MISO systems. This motivates proposed TCQ uses approach quantize each user QPSK as source constellation. Utilizing temporal correlation present channel, transform constellation stage using 2D translation scaling techniques, centers around previously selected point. parameter function BS antennas. Simulation results outperforms existing noncoherent (NTCQ) method, improving sum rate reducing overhead system