25 Years of Signal Processing Advances for Multiantenna Communications

Emil Björnson, Yonina C. Eldar, Erik G. Larsson, Angel Lozano, H. Vincent Poor

TL;DR

This paper provides a comprehensive retrospective on the 25-year evolution of multiantenna communications, tracing the journey from the simple sectorized antennas of 2G to the massive MIMO arrays that define 5G. It highlights how signal processing has shifted the industry's focus from basic link reliability (spatial diversity) to massive capacity gains (spatial multiplexing), enabling the modern era of data-hungry mobile applications.

What Problem Does This Paper Address?

In 1998, mobile communication was primarily about connecting people via voice and text, with roughly 300 million subscribers globally. Fast forward to today, and over six billion subscribers generate massive amounts of data through video streaming, gaming, and IoT devices.

The core challenge addressed by the signal processing community over these decades was how to accommodate this exponential growth in traffic (doubling every 18 months, per Cooper’s Law) within a finite and expensive radio spectrum. The paper explores how the industry moved beyond simply making links more reliable to fundamentally multiplying network capacity by using multiple antennas at both the base station (BS) and user devices.

Methods

The paper reviews the progression of several key technological paradigms:

  • Array Evolution: It categorizes the shift from fixed array antennas (2G) to phased arrays (capable of steering beams) and finally to digital antenna arrays, where every element is individually controlled by a baseband unit (BBU).
  • MIMO Paradigms: It details the transition from Single-User MIMO (increasing the rate for one device) to Multi-User MIMO (serving many devices simultaneously on the same frequency).
  • Massive MIMO: A pivotal shift where BSs use significantly more antennas ($M$) than the number of users ($K$). Key techniques discussed include Time-Division Duplexing (TDD) to exploit channel reciprocity for efficient channel state information (CSI) acquisition and linear processing (like regularized zero-forcing) to manage interference.
  • Hardware-Constrained Processing: To handle the high power consumption of millimeter-wave (mmWave) frequencies, the authors discuss hybrid analog-digital architectures and low-resolution ADCs (down to 1-bit quantization) paired with sophisticated compensation algorithms.
  • Model-Aided Machine Learning: The paper highlights algorithm unrolling (or deep unfolding), which transforms classical iterative signal processing algorithms into neural network layers, allowing for data-driven fine-tuning while maintaining theoretical interpretability.
Figure 2: There are three classical categories of arrays: 1) array antennas that generate fixed beams;

Datasets

As a review and tutorial paper, it does not introduce or utilize a specific new experimental dataset. Instead, it references standardized channel models and theoretical frameworks established in the literature (e.g., [3]–[5], [14], [22]) to validate the discussed signal processing advances.

Key Insights and Ideas

  • Multipath as an Asset: Perhaps the most radical "aha moment" was realizing that multipath propagation—once considered a nuisance causing fading—is actually the enabler for spatial multiplexing. The existence of multiple paths allows a receiver to resolve distinct linear combinations of signals, effectively creating parallel "spatial pipes" for data.
  • The Diversity-Multiplexing Tradeoff: The authors illustrate a fundamental shift in priorities. Early systems focused on spatial diversity (reliability), but as bandwidths increased, diversity was increasingly reaped in the frequency domain. This allowed spatial resources to be shifted toward spatial multiplexing (capacity), as shown in the transition from the "green plane" to the "blue plane" in Figure 3.
  • Channel Hardening: In Massive MIMO, as the number of antennas grows, the "effective" channel becomes remarkably stable and deterministic, subject to only minimal fading fluctuations. This simplifies resource allocation and power control.
  • TDD is the Scalability King: Massive MIMO's success relies heavily on TDD. Because the uplink and downlink share the same frequency, the BS can learn the channel for both directions using only uplink pilots. This makes the pilot overhead independent of the number of BS antennas, solving a major scaling bottleneck of FDD systems.
Figure 3: There are three benefit dimensions of multiantenna communication that have developed

Key Results

  • Capacity Explosion: The transition from 2G to 5G has seen bandwidths grow from ~20 MHz at 1 GHz to 100–500+ MHz in the 3.5 GHz and mmWave bands.
  • Antenna Scaling: Modern 5G base stations commonly integrate $M=32$ or $M=64$ digital antenna elements, a massive leap from the 2-4 elements common in early digital arrays.
  • Efficiency Gains: Massive MIMO has turned Multi-User MIMO into a practical reality, allowing a single BS to spatially multiplex $K=8$ or more users simultaneously, providing a direct "traffic multiplier" for the network.
Figure 1: (a) 2G deployment in 1998, consisting of fixed directive antennas that broadcast each

Limitations and Future Work

  • Complexity Bottlenecks: As $M$ and $K$ scale, the computational complexity of linear processing grows cubically ($MK^2$). Future systems will need even more efficient dimensionality reduction and compressed sensing techniques.
  • Beyond the Tower: The "boxes-in-a-tower" paradigm is reaching its physical limits. The authors point toward Cell-Free Massive MIMO, where antennas are distributed across many locations but processed coherently.
  • New Frontiers: Emerging research is looking at Integrated Communication and Sensing (ISAC), where arrays perform radar-like sensing alongside data transfer, and Reconfigurable Intelligent Surfaces (RIS), which use metamaterials to programmatically control signal reflections in the environment.

Summary updated

Abstract

Wireless communication technology has progressed dramatically over the past 25 years, in terms of societal adoption as well as technical sophistication. In 1998, mobile phones were still in the process of becoming compact and affordable devices that could be widely utilized in both developed and developing countries. There were "only" 300 million mobile subscribers in the world [1]. Cellular networks were among the first privatized telecommunication markets, and competition turned the devices into fashion accessories with attractive designs that could be individualized. The service was circumscribed to telephony and text messaging, but it was groundbreaking in that, for the first time, telecommunication was between people rather than locations. Wireless networks have changed dramatically over the past few decades, enabling this revolution in service provisioning and making it possible to accommodate the ensuing dramatic growth in traffic. There are many contributing components, including new air interfaces for faster transmission, channel coding for enhanced reliability, improved source compression to remove redundancies, and leaner protocols to reduce overheads. Signal processing is at the core of these improvements, but nowhere has it played a bigger role than in the development of multiantenna communication. This article tells the story of how major signal processing advances have transformed the early multiantenna concepts into mainstream technology over the past 25 years. The story therefore begins somewhat arbitrarily in 1998. A broad account of the state-of-the-art signal processing techniques for wireless systems by 1998 can be found in [2], and its contrast with recent textbooks such as [3]-[5] reveals the dramatic leap forward that has taken place in the interim.
BibTeX
@article{bjornson2023years,
  author = {Emil Björnson and Yonina C. Eldar and Erik G. Larsson and Angel Lozano and H. Vincent Poor},
  title = {{25 Years of Signal Processing Advances for Multiantenna Communications}},
  journal = {IEEE Signal Process. Mag.},
  year = {2023},
  doi = {10.1109/msp.2023.3261505},
}
Figure