Abstract
In this paper, we investigate the performance of point-to-point multiple-input-multiple-output (MIMO) systems with a large but finite number of antennas. By considering scenarios with and without hybrid automatic repeat request (HARQ) feedback, we determine the minimum number of transmit/receive antennas required to satisfy specific outage probability constraints. We analyze the effects of spatial correlation and various fading conditions. Our results demonstrate that stringent outage requirements can be met with relatively few antennas, particularly when leveraging HARQ.
Introduction: The Practical Limits of Massive MIMO
Massive MIMO is a foundational technology for 5G and future wireless standards. While theoretical research often assumes an infinite number of antennas to simplify mathematical models, real-world implementation is constrained by physical space, power consumption, and hardware costs. This study addresses a critical practical question: What is the minimum number of antennas ($N_t$ or $N_r$) needed to guarantee a specific level of reliability, such as a 99.99% success rate (0.01% outage probability)?
Methodology
The researchers modeled a point-to-point MIMO system using a combination of analytical derivations and numerical simulations.
Mathematical Framework
By utilizing the properties of large random matrices, the distribution of mutual information was approximated as a Gaussian random variable. This led to the derivation of Theorem 1, providing closed-form expressions for the required antenna count across various scenarios, including high/low Signal-to-Noise Ratios (SNR) and asymmetric link configurations.
HARQ and Fading Models
The study integrated Incremental Redundancy (INR) HARQ, which allows receivers to request additional bits upon failure, combining them with previous transmissions. Performance was evaluated across three temporal fading models: Fast-fading (channel changes during a packet), Slow-fading (changes between retransmissions), and Quasi-static (channel remains constant throughout the process).
Spatial Correlation
To account for physical antenna crowding, a first-order autoregressive model was used to simulate spatial correlation, measuring how performance degrades when antennas are placed too closely together.
Key Insights and Results
The Power of HARQ
A primary finding is that HARQ significantly reduces hardware complexity. By exploiting temporal diversity through retransmissions, the same reliability targets can be achieved with far fewer physical antennas.
As shown in Figure 4, in a quasi-static scenario with a $10^{-4}$ outage constraint, the system requires approximately 95 receive antennas without HARQ ($M=1$). However, with just one retransmission ($M=2$), the requirement drops dramatically to only 15 antennas.
Resilience to Spatial Correlation
While spatial correlation necessitates more antennas to compensate for lost diversity, the underlying scaling laws remain consistent.
Figure 6 illustrates that the impact of spatial correlation is negligible for coefficients up to $\beta \approx 0.4$. Beyond this threshold, the required antenna count begins to rise more sharply to maintain the same outage performance.
Conclusion
This research provides a bridge between theoretical massive MIMO and practical engineering. By demonstrating that high reliability is achievable with a "large-but-finite" number of antennas—often fewer than expected when using HARQ—it offers a roadmap for more efficient hardware design in next-generation wireless networks. Future work will extend these findings to multi-user scenarios and incorporate realistic hardware impairments.
Summary updated