Automatic Network Planning with Digital Radio Twin

TL;DR

This work tackles automatic cellular network planning by introducing AutoPlan, a framework that combines a digitally twinned radio environment with data-driven calibration and Bayesian optimization. The Digital Radio Twin (DRT) is derived by crowdsourcing real-user measurements to tune building-material parameters within a ray-tracing simulator, mitigating the sim-to-real gap while enabling fast queries via GPU acceleration. AutoPlan then performs incremental Bayesian optimization over base-station placements, using a GP surrogate and Expected Improvement to sequentially select new BS locations in a feasible region, guided by a composite objective that blends coverage and capacity. On Husker-Net campus data, AutoPlan achieves near-exhaustive search performance with orders-of-magnitude less computation, demonstrating robust adaptation to diverse deployment scenarios and highlighting the value of calibrated DRTs for data-driven network planning. The approach offers practical impact for rapid, data-driven deployment planning in real-world, complex urban environments, where precise propagation modeling and efficient search are critical.

Abstract

Network planning seeks to determine base station parameters that maximize coverage and capacity in cellular networks. However, achieving optimal planning remains challenging due to the diversity of deployment scenarios and the significant simulation-to-reality discrepancy. In this paper, we propose \emph{AutoPlan}, a new automatic network planning framework by leveraging digital radio twin (DRT) techniques. We derive the DRT by finetuning the parameters of building materials to reduce the sim-to-real discrepancy based on crowdsource real-world user data. Leveraging the DRT, we design a Bayesian optimization based algorithm to optimize the deployment parameters of base stations efficiently. Using the field measurement from Husker-Net, we extensively evaluate \emph{AutoPlan} under various deployment scenarios, in terms of both coverage and capacity. The evaluation results show that \emph{AutoPlan} flexibly adapts to different scenarios and achieves performance comparable to exhaustive search, while requiring less than 2\% of its computation time.

Figures (11)

• Figure 1: Illustration of BSs deployment in feasible region $\mathcal{F}$.
• Figure 2: The trajectory of measured RSRP.
• Figure 3: The MSE loss during DRT calibration.
• Figure 4: Conductivity of building materials, during DRT calibration.
• Figure 5: Relative permittivity of building materials, during DRT calibration.