Summary
Chapter 16 Summary: MIMO II β Space-Time Coding and Transceiver Architectures
Key Points
- 1.
Space-time code design is governed by the rank and determinant criteria: the rank of the codeword difference matrix determines the diversity order (), while controls the coding gain. Orthogonal STBCs (including Alamouti) enable single-symbol ML decoding but are limited to rate for with complex constellations.
- 2.
Spatial multiplexing (V-BLAST/D-BLAST) transmits independent data streams from each antenna, achieving throughput that scales linearly with . V-BLAST uses ordered successive interference cancellation at the receiver, detecting the strongest layer first to minimise error propagation.
- 3.
MIMO receivers span a performance-complexity hierarchy: ML detection (optimal, ) achieves diversity order ; linear ZF and MMSE () achieve diversity ; MMSE-OSIC () recovers diversity approaching through successive cancellation and optimal ordering.
- 4.
MMSE-SIC achieves MIMO capacity. With Gaussian codebooks and perfect cancellation, the sum of individual MMSE-SIC stream rates equals regardless of the decoding order. This is the MIMO analogue of successive decoding for the multiple-access channel.
- 5.
Precoding with CSIT enables the transmitter to shape signals before transmission: MRT maximises single-stream SNR, ZF precoding nulls inter-user interference, MMSE precoding balances interference suppression and power efficiency, and dirty paper coding achieves the broadcast channel capacity region.
- 6.
Limited feedback quantises the channel to bits using a codebook. Grassmannian packings optimise worst-case quantisation. The rate loss scales as , and maintaining a fixed rate gap as SNR grows requires adding feedback bits per 3 dB β a fundamental scaling law that motivates TDD reciprocity for massive MIMO.
Looking Ahead
Chapter 17 extends these ideas to multi-user MIMO and massive MIMO systems. We will see how the large number of base-station antennas () simplifies precoding (ZF becomes near-optimal due to channel hardening), how uplink pilot contamination limits performance, and how the spatial multiplexing and beamforming concepts from this chapter scale to serve dozens of users simultaneously.