Part 2: Multi-User MIMO: Precoding and Detection

Chapter 8: FDD Massive MIMO and CSI Feedback

Advanced~270 min

Learning Objectives

  • Quantify the downlink pilot overhead and uplink feedback overhead that make FDD massive MIMO fundamentally harder than TDD
  • Explain how channel sparsity in the angular domain enables compressed CSI feedback
  • Describe the 5G NR Type I and Type II CSI reporting frameworks and their beam-combination structure
  • Analyze the CsiNet deep learning architecture for CSI compression and its rate-distortion interpretation
  • Derive how JSDM reduces the FDD feedback dimension from NtN_t to the group rank rgr_g
  • Compare codebook-based, compressed, and learning-based CSI feedback in terms of overhead, accuracy, and complexity

Sections

Prerequisites & Notation
nextSpecial
The FDD Challenge
Compressed CSI Feedback
Codebook-Based Feedback
Deep Learning for CSI Compression
JSDM as a Structured FDD Solution
Chapter Summary
Special
Exercises
References & Further Reading
Special

Prerequisites

Massive MIMO fundamentals: channel hardening, favorable propagation (MIMO Ch. 1–4)TDD reciprocity and the FDD overhead challenge (MIMO Ch. 1, §5)JSDM framework: two-stage precoding, group-based dimensionality reduction (MIMO Ch. 7)Linear precoding: MRT, ZF, MMSE (MIMO Ch. 6)Angular-domain channel representation and spatial correlation (MIMO Ch. 2)Rate-distortion theory fundamentals (ITA Ch. 6)Linear algebra: SVD, subspace projection, principal components

💬 Discussion

Loading discussions...