• Computer Science > Networking and Internet Architecture [Submitted on 12 Nov 2025 (v1), last revised 17 Feb 2026 (this version, v2)] Title:Refined Bayesian Optimization for Efficient Beam Alignment in Intelligent Indoor Wireless Environments View PDF HTML (experimental)Abstract:Future intelligent indoor wireless environments require fast and reliable beam alignment to sustain high-throughput links under mobility and blockage. • Exhaustive beam training achieves optimal performance but is prohibitively costly. • In indoor settings, dense scatterers and transceiver hardware imperfections introduce multipath and sidelobe leakage, producing measurable power across multiple angles and reducing the effectiveness of outdoor-oriented alignment algorithms. • This paper presents a Refined Bayesian Optimization (R-BO) framework that exploits the inherent structure of mmWave transceiver patterns, where received power gradually increases as the transmit and receive beams converge toward the optimum. • R-BO integrates a Gaussian Process (GP) surrogate with a Matern kernel and an Expected Improvement (EI) acquisition function, followed by a localized refinement around the predicted optimum. • The GP hyperparameters are re-optimized online to adapt to irregular variations in the measured angular power field caused by reflections and sidelobe leakage.

Article Summaries:

  • Refined Bayesian Optimization for Efficient Beam Alignment in Intelligent Indoor Wireless Environments

Researchers have introduced a Refined Bayesian Optimization (R‑BO) framework to accelerate beam alignment in indoor millimeter‑wave networks. The method models received power as a Gaussian Process with a Matern kernel and uses Expected Improvement to guide beam probing, followed by a local refinement around the predicted optimum. Online re‑optimization of GP hyperparameters adapts to irregular angular power variations caused by reflections and sidelobe leakage. In laboratory tests with 43 receiver positions, R‑BO achieved 97.7 % alignment accuracy within 10°, incurred less than 0.3 dB average loss, and cut probing overhead by 88 % compared to exhaustive search, demonstrating a practical, low‑overhead solution for real‑time indoor wireless systems.

Sources: