研究目的
To investigate beam training and allocation for multiuser millimeter wave massive MIMO systems, aiming to reduce training overhead, mitigate interference from beam conflicts, and improve spectral efficiency.
研究成果
The proposed QC-based beam allocation effectively mitigates interference from beam conflicts and improves spectral efficiency. The IS-based and SP-based schemes significantly reduce beam training overhead with minimal efficiency loss. Future work could explore other performance metrics and algorithms for beam allocation.
研究不足
The study assumes TDD systems with channel reciprocity and specific geometric channel models. The beam training overhead reduction comes at a cost of slight spectral efficiency degradation. The schemes may not account for all real-world impairments like mobility or hardware imperfections.
1：Experimental Design and Method Selection:
The study uses an orthogonal pilot (OP) based beam training scheme for simultaneous training of all users with the BS, followed by a QoS-constrained (QC) beam allocation scheme to handle beam conflicts. Two partial beam training schemes (IS-based and SP-based) are proposed to reduce overhead. Theoretical models include geometric channel models and optimization algorithms for beam selection and allocation.
2：Sample Selection and Data Sources:
Simulations are based on Monte Carlo methods with 2000 random channel implementations, assuming mmWave channels with 3-5 multipaths per user and complex Gaussian gains.
3：List of Experimental Equipment and Materials:
The system involves a BS with NBS ULA antennas and NRF RF chains, and users with NUE ULA antennas and a single RF chain. Phase shifters are used for analog precoding and combining.
4：Experimental Procedures and Operational Workflow:
Beam training is performed using OP-based, IS-based, or SP-based schemes, followed by channel estimation and digital precoder design (ZF or MMSE). Beam allocation is applied using Algorithm 1 to avoid conflicts. Spectral efficiency is evaluated under varying parameters (K, SNRdl, NUE, NBS).
5：Data Analysis Methods:
Performance is measured by spectral efficiency (sum-rate averaged over users), with comparisons between different schemes using simulation results. Statistical analysis includes overhead calculations and efficiency improvements.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
