- The paper introduces a fixed mmWave multi-user MIMO system utilizing a 16x16 Butler matrix to generate orthogonal beams, significantly boosting summed spectral efficiency.
- It compares digital precoding techniques like MR, ZF, and RZF, demonstrating that advanced beamforming improves performance and spatial fairness under high user loads.
- Real-world tests closely match simulation results, confirming the architecture's viability as a low-cost, high-throughput solution for fixed wireless access.
This paper investigates a fixed millimeter-wave (mmWave) Multi-User Multiple-Input Multiple-Output (MIMO) system architecture, combining digital MIMO and an analog multi-beam antenna array utilizing a 16x16 Butler matrix to generate orthogonal beams. The study presents a performance comparison between the proposed system and traditional patch antenna arrays, offering a practical assessment through both simulations and real-world measurements.
Introduction
The evolution to 5G New Radio (NR) envisions significant improvements in spectral efficiency and data rates, particularly advantageous for high-density urban environments and underserved rural areas. Taking advantage of the mmWave frequency bands (26–300 GHz) opens avenues for high-throughput wireless broadband access, serving as a viable alternative to costly wired installations. In this context, the paper proposes a fixed wireless access architecture utilizing a multi-beam antenna array with a 16x16 Butler matrix. The architecture aims to bridge the connection gap between optical fiber termination points and Customer Premise Equipment (CPE) by supporting up to 16 simultaneous spatial data streams, each capable of providing several Gbps downlink to individual users.
Equipment and System Setup
The proposed architecture comprises a multi-beam antenna array and a MIMO testbed. The antenna array, expanded from previous work, employs a 16x16 Butler matrix to connect with a 1x16 linear antenna array, enabling the transmission of 16 orthogonal beams. The azimuth beam pattern of this setup is characterized by a half-power beam width of 7 degrees and a maximum gain of 16 dBi, covering a total spatial angle range of -68° to 68° at operating frequencies from 25 to 30 GHz. The MIMO testbed, designed on an LTE-TDD basis, operates with 16 RF chains at a 2.4 GHz intermediate frequency, using OFDM over a 20 MHz bandwidth for signal modulation.
System Model and Methodology
The study examines a fixed downlink multi-user MIMO configuration in a singular cell scenario, facilitating simultaneous transmission to a maximum of 16 users. The Base Station (BS) employs a digital precoder FBB in conjunction with the multi-beam array. The transmitted signals are precalculated based on an estimated channel impulse response (CIR) matrix acquired through standard channel estimation methods like the Minimum Mean Square Error (MMSE).
For delivery of spectral efficiency, the individual user and summed cell spectral efficiency (SE) are computed. These metrics are compared across a simulated classic linear patch antenna array and the multi-beam antenna configuration implemented in this research.
Digital Precoding Techniques
The authors explore three digital precoding strategies:
- Maximum Ratio (MR) Precoding: This simple technique aims to maximize the received signal for each user with a low computational overhead by directly leveraging the channel estimation, though at the cost of higher interference.
- Zero Forcing (ZF) Precoding: Primarily designed to minimize inter-user interference by inverting the channel matrix.
- Regularized Zero Forcing (RZF): A hybrid approach that offers a compromise between mitigating noise and interference.
The study computes these precoding matrices with estimated channel models to optimize transmission.
Results and Evaluation
The evaluated system displays significant competitive performance in terms of both individual and summed spectral efficiency when compared to traditional patch antenna arrays. Key differences are mainly observed under multi-user scenarios. For example, at higher user counts approaching the maximum capacity of the MIMO system, the novel beamforming antenna array demonstrates superior spectral efficiency and fairness compared to traditional patch antenna arrays.
The results from real-environment tests align closely with simulations, indicating the viability of the proposed fixed mmWave multi-beam architecture as a low-cost, spectrally efficient solution suitable for expanding wireless accessibility in both urban and rural locales.
Conclusion
The research showcases a novel millimeter-wave MIMO system architecture that effectively combines digital MIMO with an analog multi-beam antenna array, facilitated by an innovative high-dimension Butler matrix. The dual-mode beamforming approach proves efficacious in supporting high data rate connectivity in multi-user scenarios. While individual spectral efficiency is higher in traditional configurations, the proposed system demonstrates enhanced summed spectral efficiency and spatial fairness when multiple users are served simultaneously. The paper outlines the path for rigorous further validation and potential optimization to fully realize the capabilities of this architecture in practical fixed wireless access deployments.