基于现实网络数据的通信感知一体化网络覆盖预测与优化

A Real-World Data-Based Network Coverage Prediction and Optimization for Integrated Sensing and Communications

传统基于人工经验、试错迭代的低效方法已无法适应未来多功能网络的参数优化。为了提升通信感知一体化(ISAC)网络的优化效率以及服务质量,亟需建立系统级的全局网络性能建模与优化新范式。为此,提出了一种面向6G的ISAC网络覆盖预测与优化框架。该框架利用离线的现实网络数据与多波束信道建模的方法,实现任意天线参数下的ISAC网络性能预测,并且利用零阶块坐标下降等数学优化工具求解黑盒优化问题,实现网络参数的精准寻优。基于所提框架,我们对现实世界中的空地协同ISAC网络进行了系统级性能仿真,并对低空用户进行感知性能优化,对地面用户进行通信性能优化。基于真实现网数据的实验结果表明,所提方法在性能上显著超越了传统的建模与优化方法。

The conventional network optimization methods,based on manual experience and trial-and-error iterations are no longer suitable for parameter optimization of future multifunctional networks.To enhance the optimization efficiency and service quality of integrated sens⁃ing and communication(ISAC)networks,there is an urgent need to establish a new paradigm for system-level network performance model⁃ing and optimization.To this end,a 6G-oriented ISAC network coverage prediction and optimization framework is proposed.This framework utilizes offline real network data and multi-beam channel modeling methods to predict ISAC network performance under various parameters and employs advanced mathematical optimization tools such as zero-order block coordinate descent to solve black-box optimization prob⁃lems,achieving precise network parameter optimization.Based on the proposed framework,we conducted system-level performance simu⁃lations for real-world aerial-ground cooperative ISAC networks,optimizing sensing performance for low-altitude users and communication performance for ground users.Experimental results demonstrate that our method significantly outperforms traditional modeling and optimi⁃zation methods.