Prerequisites & Notation
Before You Begin
This chapter generalises the Zheng-Tse diversity-multiplexing tradeoff of Chapter 12 to MIMO channels equipped with an ARQ feedback loop: the receiver can request up to retransmissions, and the system accumulates evidence across rounds until it decodes successfully or runs out of budget. The point of view is that of El Gamal, Caire, and Damen (2006): an ARQ round is a new channel realisation, each round incrementally lowers the effective rate, and both together reshape the DMT into a three-dimensional object . To follow the arguments the reader needs the static-DMT machinery of Ch. 12 (outage exponent, exponential equality, Gaussian random-coding achievability), the block-fading MIMO channel model of Ch. 10, and the pairwise-error / rank-determinant analysis of space-time codes from Ch. 11. On the practical side, the reader should know what a hybrid ARQ (HARQ) protocol looks like in LTE/NR β an LDPC-coded transport block, a circular-buffer rate matcher, and a small number of redundancy versions transmitted in sequence.
- Zheng-Tse DMT for block-fading MIMO(Review ch12)
Self-check: Can you write at integer corners for an i.i.d. Rayleigh channel with , and explain why the curve is piecewise-linear between corners?
- Outage probability of MIMO block-fading channels(Review ch10)
Self-check: Can you write and identify its high-SNR exponent with the DMT curve value?
- Rank and determinant criteria for space-time codes(Review ch11)
Self-check: Can you state the rank criterion for the codeword-difference matrix ?
- Exponential equality and large-deviations asymptotics(Review ch12)
Self-check: Can you apply the definition and manipulate sums, products, and integrals of the form ?
- Basic ARQ / HARQ protocols (CC-HARQ, IR-HARQ)(Review ch12)
Self-check: Can you explain what ACK/NACK feedback does, what "stop-and-wait" means, and how Chase combining differs operationally from incremental redundancy?
- Forward-error correction via LDPC and rate matching(Review ch08)
Self-check: Can you describe how a mother LDPC code is punctured into several redundancy versions via a circular buffer, and why RV_0 typically contains the systematic bits?
Notation for This Chapter
Symbols specific to the ARQ-DMT analysis. The Chapters 10β12 MIMO notation (channel matrix , codeword matrix , SNR , noise , capacity , mutual information ) continues to apply and is not repeated here.
| Symbol | Meaning | Introduced |
|---|---|---|
| Maximum number of ARQ rounds (delay budget). In Chapter 12 denoted the block length; here it denotes the ARQ round budget. | s01 | |
| Round index, | s01 | |
| Codeword matrix transmitted in round ; for IR, each round sends fresh parity | s01 | |
| MIMO channel realisation in round ; assumed i.i.d. across | s01 | |
| ARQ-DMT curve: optimal diversity at effective multiplexing with at most rounds | s02 | |
| Effective (long-term) multiplexing gain, | s01 | |
| -round outage probability β the probability that the accumulated -round mutual information is less than | s02 | |
| Effective throughput (bits per channel use) accounting for ACK/NACK and retransmissions | s04 | |
| HARQ round-trip time β the delay between transmission and the arrival of feedback | s05 | |
| Chase combining / incremental redundancy β the two canonical HARQ flavours | s01 | |
| Redundancy version β the LTE/NR identifier (RV_0, RV_1, RV_2, RV_3) specifying which portion of the circular buffer is transmitted | s05 | |
| Block error rate β the probability of decoding failure after all available HARQ rounds | s04 |