ELEC5508 Chap.8 MIMO Receivers and Beamforming
MIMO Receivers and Beamforming
This chapter compares MIMO detection strategies through the interference-noise trade-off. Matched filtering preserves desired energy but leaves cross-stream interference. Zero forcing cancels interference but amplifies noise along weak singular directions. MMSE balances both, while successive interference cancellation decodes and subtracts layers in an order chosen to reduce error propagation.
Beamforming applies the same spatial logic to physical directions. No receiver is best without an operating point. A well-conditioned high-SNR channel can favour nulling, while a poorly conditioned moderate-SNR channel exposes the inverse's noise penalty. SIC improves later layers only when earlier decisions and channel estimates are accurate.
Transmit beamforming adds an information question because the transmitter must obtain channel direction through reciprocity, feedback or another acquisition path before placing a null or main lobe.
What this chapter covers
- 01
Matched filtering
- 02
Zero-forcing channel inversion
- 03
Linear MMSE filtering
- 04
Successive interference cancellation
- 05
V-BLAST and layer ordering
- 06
Transmit and receive beamforming
Worked example · free
Choose a SIC decoding order
- +1Select stream 2 because its post-processing SINR of 11 is largest.
- +1Decode the stream-2 constellation symbol.
- +1Reconstruct and subtract its channel contribution.
- +1Remove the decoded column from the active channel.
- +1Recompute the remaining filter or SINR before selecting the next layer.
Key terms
- Matched filter
- A combiner aligned with a desired channel to collect signal energy.
- Noise enhancement
- Amplification of receiver noise caused by inversion of weak channel directions.
- Linear MMSE
- A linear estimator balancing residual interference against amplified noise.
- V-BLAST
- A layered MIMO architecture using ordered detection and successive cancellation.
- Beamforming
- Complex spatial weighting that reinforces a desired direction or suppresses others.
MIMO Receivers and Beamforming FAQ
Why can zero forcing perform poorly?
When H is poorly conditioned, inversion divides by a weak singular value and greatly amplifies noise. Zero forcing removes inter-stream interference, but that benefit can be outweighed at low or moderate SNR by the amplified noise.
Why does MMSE depend on SNR?
MMSE includes the noise level in its optimisation. At low SNR it avoids aggressive inversion, while at high SNR interference becomes relatively more important and the solution moves closer to zero forcing. The preferred balance changes with both SNR and conditioning.
How does SIC improve a linear receiver?
After one reliable layer is decoded, its reconstructed contribution is subtracted. Later layers then face less interference. Ordering matters because an early mistake contaminates every later residual, so strongest-post-processing-SINR selection is commonly used.
How do I choose among matched filtering, ZF, MMSE and SIC?
Start from the channel conditioning, interference strength, SNR and available channel knowledge. A matched filter is simple but leaves cross-stream interference when channel vectors are not close to orthogonal. Zero forcing removes that interference under a suitable full-rank channel, yet weak singular values can amplify noise.
MMSE regularises the inversion and trades residual interference against noise enhancement, especially at low or moderate SNR. SIC can improve a linear front end by decoding, reconstructing and subtracting layers, but it adds ordering sensitivity and error propagation. Compare every receiver with the same transmitted power, channel estimate and noise model.
Report errors against both SNR and condition number so the selected method follows the observed channel.
Exam move
Compare filters on a plane of interference suppression versus noise robustness. Practise explaining every result through singular values, and write the residual update explicitly when using SIC. Start with a two-stream received-signal model and derive what remains after a matched filter. Identify the cross-correlation term rather than saying only that interference exists.
Move to zero forcing and connect the matrix inverse to weak singular values. Use a simple diagonalised example to show how dividing by a small gain increases noise variance. Add the MMSE regularisation term and describe how its behaviour shifts with SNR. At low or moderate SNR it avoids aggressive inversion; at high SNR it approaches an interference-nulling solution.
For SIC, write the strongest-first procedure as a sequence: compute post-processing SINRs, choose a layer, decide a symbol, reconstruct its channel contribution, subtract, reduce the channel and recompute. Never keep the old order automatically. Discuss error propagation by imagining one wrong early symbol and tracking the structured residual it leaves.
For beamforming, translate complex weights into coherent addition, destructive cancellation and null constraints. Count available spatial dimensions before promising several nulls and data streams simultaneously. Finally, compare algorithms under the same transmitted power, channel estimate and noise knowledge. Plot BER against SNR and condition number, not SNR alone.
If a result disagrees with the expected trade-off, inspect channel scaling, power normalisation, filter dimensions and whether CSI quality was held constant.
Working through MIMO Receivers and Beamforming in ELEC5508? Sia is AskSia’s AI Electrical Engineering tutor — ask any ELEC5508 MIMO Receivers and Beamforming question and get a clear, step-by-step explanation grounded in how ELEC5508 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.