Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MIMO in network simulators: Design, implementation and evaluation of single-user MIMO in ns-3 5G-LENA

Biljana Bojovic, Sandra Lagen

TL;DR

This paper addresses the absence of full SU-MIMO modeling in ns-3 5G-LENA by implementing a 3GPP-compliant SU-MIMO solution with Type-I codebooks, MMSE-IRC receivers, and exhaustive PMI/RI search. It introduces MatrixArray for efficient 3D frequency-domain channel representations and extends the ns-3 antenna and spectrum modules to accommodate multi-port, dual-polarized arrays, along with removal of the OFDMA downlink trick to enable clean MIMO signaling. The work details a comprehensive evaluation across multiple antenna configurations, PMI intervals, and interference scenarios, demonstrating both accuracy and the associated computational costs, and provides insights into fidelity versus complexity. The results affirm the viability of high-fidelity SU-MIMO in system-level ns-3 simulations and underscore the need for optimized PM search and interference modeling for large-scale studies, with future work spanning uplink and MU-MIMO extensions and algorithmic optimizations.

Abstract

MIMO technology has been studied in textbooks for several decades, and it has been adopted in 4G and 5G systems. Due to the recent evolution in 5G and beyond networks, designed to cover a wide range of use cases with every time more complex applications, it is essential to have network simulation tools (such as ns-3) to evaluate MIMO performance from the network perspective, before real implementation. Up to date, the well-known ns-3 simulator has been missing the inclusion of single-user MIMO (SU-MIMO) models for 5G. In this paper, we detail the implementation models and provide an exhaustive evaluation of SU-MIMO in the 5G-LENA module of ns-3. As per 3GPP 5G, we adopt a hybrid beamforming architecture and a closed-loop MIMO mechanism and follow all 3GPP specifications for MIMO implementation, including channel state information feedback with precoding matrix indicator and rank indicator reports, and codebook-based precoding following Precoding Type-I (used for SU-MIMO). The simulation models are released in open-source and currently support up to 32 antenna ports and 4 streams per user. The simulation results presented in this paper help in testing and verifying the simulated models, for different multi-antenna array and antenna ports configurations.

MIMO in network simulators: Design, implementation and evaluation of single-user MIMO in ns-3 5G-LENA

TL;DR

This paper addresses the absence of full SU-MIMO modeling in ns-3 5G-LENA by implementing a 3GPP-compliant SU-MIMO solution with Type-I codebooks, MMSE-IRC receivers, and exhaustive PMI/RI search. It introduces MatrixArray for efficient 3D frequency-domain channel representations and extends the ns-3 antenna and spectrum modules to accommodate multi-port, dual-polarized arrays, along with removal of the OFDMA downlink trick to enable clean MIMO signaling. The work details a comprehensive evaluation across multiple antenna configurations, PMI intervals, and interference scenarios, demonstrating both accuracy and the associated computational costs, and provides insights into fidelity versus complexity. The results affirm the viability of high-fidelity SU-MIMO in system-level ns-3 simulations and underscore the need for optimized PM search and interference modeling for large-scale studies, with future work spanning uplink and MU-MIMO extensions and algorithmic optimizations.

Abstract

MIMO technology has been studied in textbooks for several decades, and it has been adopted in 4G and 5G systems. Due to the recent evolution in 5G and beyond networks, designed to cover a wide range of use cases with every time more complex applications, it is essential to have network simulation tools (such as ns-3) to evaluate MIMO performance from the network perspective, before real implementation. Up to date, the well-known ns-3 simulator has been missing the inclusion of single-user MIMO (SU-MIMO) models for 5G. In this paper, we detail the implementation models and provide an exhaustive evaluation of SU-MIMO in the 5G-LENA module of ns-3. As per 3GPP 5G, we adopt a hybrid beamforming architecture and a closed-loop MIMO mechanism and follow all 3GPP specifications for MIMO implementation, including channel state information feedback with precoding matrix indicator and rank indicator reports, and codebook-based precoding following Precoding Type-I (used for SU-MIMO). The simulation models are released in open-source and currently support up to 32 antenna ports and 4 streams per user. The simulation results presented in this paper help in testing and verifying the simulated models, for different multi-antenna array and antenna ports configurations.
Paper Structure (23 sections, 6 equations, 8 figures, 3 tables, 1 algorithm)

This paper contains 23 sections, 6 equations, 8 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. MIMO background
  3. MIMO digital baseband processing preliminaries
  4. MIMO in 3GPP 5G NR
  5. New full SU-MIMO model for ns-3 and 5G-LENA
  6. Introducing MatrixArray into ns-3
  7. Extending ns-3 antenna and spectrum modules to support MIMO
  8. Removal of the "OFDMA downlink trick"
  9. The MIMO interference and SINR calculations
  10. Computation of TBLER based on the MIMO SINR
  11. Implementation of the 3GPP-compliant Type-I precoding codebooks
  12. Search for the optimal precoding matrix
  13. Rank parameter
  14. MIMO activation
  15. Evaluation and analysis
  16. ...and 8 more sections

Figures (8)

  • Figure 1: Hybrid precoding structure for MIMO systems in 3GPP 5G NR.
  • Figure 2: Comparison of the computational complexity of the ns-3 3D array data structure based on the std::vector of matrices and the proposed MatrixArray-based approach
  • Figure 3: Two-ports codebook (Table 5.2.2.2.1-1)
  • Figure 4: Single-Panel Codebook 3GPP TS 38.214, 16 AEs, 8 ports
  • Figure 5: Single-Panel Codebook 3GPP TS 38.214, 64 AEs, 16 ports
  • ...and 3 more figures